Where it stands: No. The year-round turf-to-AW returner rule after 150+ days scored 1.14 at discovery but 0.979 on 3155 in the sealed vault (p 0.64), so there is no robust year-round AW switch effect.
The notebook
In order: what the lab saw, what it asked, what it proposed and why, what the tests said, and what an investigator concluded. Numbers are aggregates on the discovery data unless a step says otherwise.
Are Irish splits of under 100 bets in this thread informative at all, given they flip from 1.48 to 0.73 across overlapping cuts? open
Within the H126 handicap base, do returners priced 16/1 or bigger (sp_decimal > 15) score at or below 0.90, or as well as the capped group? testing
Do lightly raced three-year-old handicap returners in spring show the same lift, or is the mispricing specific to four-year-olds? open
Is the early-spring lift really confined to returners at 14/1 or shorter, given that 16/1+ went 0.79 on 114 in March–April but 1.19 on 245 in the earlier unfancied cut? open
Is the broad H126 lift (1.13 on 1904 at discovery) decaying over time, given that the March–April sharpen fell to 1.02 on 1548 in the vault? open
Is the returner lift specific to horses whose last win came in a handicap (a real mark rise), rather than to maiden or novice winners making a handicap debut? open
Is the 1.07 lift on 971 lightly raced three- and four-year-old non-winner returners a stable feature, or a property of the 2026 spring alone? open
Does the spring returner lift need a handicap (a mark that lags the horse's progress), or does a rust discount alone produce it in non-handicaps too? testing
Within spring Flat handicap returners after 150+ days, does the lift sit in the second tier of the market (roughly 3/1 to 7/1) rather than at 3/1 or shorter, at any age? testing
Is the fade after 27 April in the fancied-returner rule (1.18 to 0.97) a May effect that repeats each spring, or just noise in one season? open
Is the repeated Irish null for spring returners structural (big yards' returners in Ireland priced sharply) or just small samples of 50 to 80 bets? open
Within the non-four-year-old second-tier lift, is it the three-year-olds or the exposed older horses (prior_runs > 10) that beat their price? open
Is the late-April window (28-30 April, 1.41 on 101) a real step before May (0.986 on 792), or noise from a few wins? open
Do Irish all-weather (Dundalk) returners really fail, or is 0.45 on 40 just noise in a thread that is GB-led everywhere? open
Within turf-to-AW Flat handicap switchers, does A/E rise steadily with the length of the layoff (14-60, 61-149, 150+ days), as a turf-discount-plus-readiness mechanism predicts? testing
Is the long-gap turf-to-AW lift carried by a handful of 16/1+ winners rather than spread across the price bands? open
Does the turf-to-AW lift exist in Ireland at all, or is it a GB AW-programme effect? open
Is the turf-to-AW lift really a class drop, with turf horses dropping into weaker AW handicaps, rather than a surface discount? testing
Is the gap between discovery and vault in this thread a time effect (the market has priced turf-to-AW switchers since about 2024) or regression from selecting the best slice? open
Does the 1.08 for non-class-dropping turf-to-AW switchers survive in data outside the discovery window, or is it the same artefact as H-98ae1db9? open
Did 5 of 5 survived attacks fail to predict a vault contradiction because the attacks all use the discovery window and so cannot detect in-window overfitting? open
because A horse that won its last start at three and has had the winter to strengthen. Lightly raced four-year-olds are still improving, and a handicap mark set on three-year-old form can lag behind.
race_type = 'Flat' AND is_handicap = 1 AND month IN (3, 4, 5) AND age = 4 AND prior_runs <= 10 AND days_since_last_run >= 150 AND last_position = 1
H126-lightly-raced-four-year-old-returns (H126: lightly raced four year old returns) was GREEN in the earlier research: A/E 1.12 on 2958 bets (periods 1.21 / 1.01 / 1.20); now PROMISING.
Why was the middle period flat (1.01 on 1329 bets) when the outer periods were 1.21 and 1.20? And is the edge real only at Betfair SP (BSP ROI +1.6%, SP ROI -19.6%)?
An unstable middle period plus a profit that disappears at SP means the published system needs a check of its stability and of the price it can be bet at. A test splits by period, then by region, sp_rank and book_pct.
openpriority 0.45raised by O-7b05ae3b
New question16 Oct, 12:09
If the horses are really improving, does the edge carry on to their second run after the winter break?
Maturation that the mark hasn't caught up with should still be there on the next start, and the first-run result may have been missed by the market. Freshness or yard readiness would fade by then. A test reads last_days_since_last_run, days_since_last_run, age, prior_runs and last_position.
answeredpriority 0.55raised by O-7b05ae3b
New question16 Oct, 12:09
Is the H126 lift caused by four-year-old physical maturation that the handicap mark hasn't caught up with, or just by lightly raced horses returning fresh at any age?
O-7b05ae3b shows A/E 1.12 on 2958 bets, but the rule mixes age, light racing and a long break. If lightly raced five-year-olds returning after 150+ days score the same, the age story is wrong. A test reads age, prior_runs, days_since_last_run, race_type, is_handicap and month.
To answer: If the horses are really improving, does the edge carry on to their second run after the winter break?
because A four-year-old that strengthened over the winter is still ahead of its mark on its second start. The handicapper reacts slowly, and a rusty return run may hide the progress.
and the market misses it because A return-run defeat after a long break reads as poor form, and the market discounts it rather than putting it down to needing the run.
predicting positive, A/E 1.05 to 1.12 on 800+ bets
and it is wrong if A/E below 1.00 on 800+ bets. That would mean the edge is first-time-out readiness or freshness, not lasting improvement.
race_type = 'Flat' AND is_handicap = 1 AND month IN (4, 5, 6) AND age = 4 AND prior_runs <= 11 AND last_days_since_last_run >= 150 AND days_since_last_run <= 60
To answer: Is the H126 lift caused by four-year-old physical maturation that the handicap mark hasn't caught up with, or just by lightly raced horses returning fresh at any age?
because A horse with five or fewer runs was handicapped on three runs or so, often as a backward juvenile or three-year-old. It has the most room left to improve, and its mark is the most likely to understate its ability at four. This is stated before seeing any split.
and the market misses it because The form book shows only a few modest runs, so punters price it off a thin record and an unimpressive profile.
predicting positive, A/E 1.15 to 1.25 on 700+ bets
and it is wrong if A/E at or below the full H126 figure (1.12), or any period below 1.0 on 700+ bets
race_type = 'Flat' AND is_handicap = 1 AND month IN (3, 4, 5) AND age = 4 AND prior_runs <= 5 AND days_since_last_run >= 150
To answer: Is the H126 lift caused by four-year-old physical maturation that the handicap mark hasn't caught up with, or just by lightly raced horses returning fresh at any age?
because If the cause is growing stronger between three and four, five-year-olds with the same light record and long break should show little or no lift. If they lift as much, the cause is freshness or a light record, not age.
and the market misses it because Under the maturation story, the market reads the old three-year-old form and mark; a five-year-old has had another season for the handicapper to catch up.
predicting negative, A/E 0.95 to 1.03 on 400+ bets (no edge), well below H126's 1.12
and it is wrong if Five-year-olds score A/E 1.10 or more on 400+ bets with every period above 1.0. That would mean light racing plus a break, not age, carries the edge.
race_type = 'Flat' AND is_handicap = 1 AND month IN (3, 4, 5) AND age = 5 AND prior_runs <= 10 AND days_since_last_run >= 150
The H126 rule (Flat handicap, March to May, a four-year-old with ten or fewer runs, first run for 150+ days) holds at A/E 1.12 on 2958 bets. The three periods were 1.21, 1.01 and 1.20, and discovery gave 1.13 on 1904 bets. The 'won last time out' slice reached 1.27 on 371 bets and is now the live system SYS-four-year-old-returns-after-a-win. Two warnings come with it. The middle period was flat on the largest sample (1.01 on 1329 bets), so the edge is not uniform over time. And the money only shows at Betfair SP: BSP ROI is +1.6%, while SP ROI is -19.6%. This is a selection edge that needs a better price than SP.
The thread has shown that the rule finds runners who beat their price. It has not shown why. The stated cause is that a four-year-old has grown stronger over the winter while its mark still reflects three-year-old form. That story is mixed up with two others: light racing, and freshness or yard readiness first time out. The seasonal finding already says the seasonal edges that work are about who is ready first time out.
Three tests would separate these causes, so the next step is to test the cause rather than slice further:
- **Five-year-olds (rival explanation):** the maturation story predicts no edge.
- **Second run after the break (consequence of the same cause):** lasting improvement predicts the edge carries on; readiness predicts it fades.
- **Five or fewer prior runs (sharpen):** chosen because the cause itself says the mark is least accurate there, not by slicing the near result.
No open questions existed to close.
investigator v1
Observed16 Oct, 12:11
H-3cdb61f5 (Second run after the winter break for lightly raced four-year-olds) pointed the predicted way but failed the bar: A/E 1.06 on 459 bets (p 0.353), control 0.99.
near miss0.56reasoned about 16 Oct, 12:13from H-3cdb61f5
Observed16 Oct, 12:11
H-3cdb61f5 (Second run after the winter break for lightly raced four-year-olds): the period 'after 2026-04-27' carries the effect (A/E 1.16) while the rest score 2025-10-01 to 2026-04-27 0.61.
concentration0.55reasoned about 16 Oct, 12:13from H-3cdb61f5
Observed16 Oct, 12:11
H-ef8fb404 (Lightly raced five-year-olds back from a break: is age the cause?) pointed the predicted way but failed the bar: A/E 0.76 on 89 bets (p 0.795), control 1.03.
near miss0.60reasoned about 16 Oct, 12:13from H-ef8fb404
Tested16 Oct, 12:11
broad rule: A/E 1.06 on 459 bets, p = 0.35 (95%: 0.80 to 1.38), by period 0.61 · 1.16. fail
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Observed16 Oct, 12:12
H-625f1f2d (Fewest prior runs: the mark has the least information) passed discovery as a niche: A/E 1.18 on 218 bets (p 0.236), same-race control 0.98.
result0.50reasoned about 16 Oct, 12:13from H-625f1f2d
Tested16 Oct, 12:12
control: Two comparisons. (1) Same-race other runners: expected near 1.00 (about 0.97 to 1.02), showing the race itself : A/E 1.21 on 301 bets, p = 0.14 (95%: 0.86 to 1.66), by period 1.30 · 1.09. info
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
New question16 Oct, 12:13
Is the edge strongest when the handicapper has left the mark unchanged since the three-year-old evidence, as a lagging-mark mechanism requires?
The fewest-runs cut (1.18 on 218) points to information scarcity as the cause. If the mark lag is real, horses whose rating has not been raised should carry the lift, and horses already raised should not. A test reads rating_change, official_rating, last_official_rating and prior_runs.
answeredpriority 0.50raised by O-e2338df0, O-cee491b3
New question16 Oct, 12:13
Do lightly raced three-year-olds on their first handicap run after the winter show the same lift, given their marks rest on even thinner juvenile form?
If the cause is a young horse maturing faster than a mark set on early form, three-year-olds coming back off juvenile marks should show it at least as strongly. The five-year-old test was too thin (89 bets) to settle the age question; a younger-age comparison has far more runners. A test reads age, prior_runs, days_since_last_run, is_handicap and month.
answeredpriority 0.45raised by O-cee491b3, O-e2338df0
New question16 Oct, 12:13
Is the H126 lift and its second-run follow-up a late-spring turf effect, weak or absent on the all-weather in March and early April?
The second-run rule scored 0.61 before 27 April 2026 and 1.16 after. With only one spring of data, the period split is also a surface and calendar split. If the edge is turf-only, the published system should be restricted. A test reads surface, month, days_since_last_run, prior_runs and age.
answeredpriority 0.55raised by O-c392fc29, O-2c8b0b64
To answer: Is the H126 lift and its second-run follow-up a late-spring turf effect, weak or absent on the all-weather in March and early April?
because Four-year-olds returning on turf in spring meet handicappers who are still rated on last season's form. Horses that spent the winter on the all-weather have already been re-assessed. So the mark lag should show on turf and fade on the all-weather.
and the market misses it because The form book shows a three-year-old turf mark and a long absence. Punters cannot see how much the horse has strengthened, and the all-weather form some rivals bring makes those rivals look more current.
predicting positive, A/E 1.12 to 1.20 on 400+ bets on turf, with the all-weather part near 1.00
and it is wrong if Turf scores A/E below 1.05 on 400+ bets, or the all-weather part scores as high as turf
race_type = 'Flat' AND is_handicap = 1 AND month IN (3, 4, 5) AND age = 4 AND prior_runs <= 10 AND days_since_last_run >= 150 AND surface = 'Turf'
To answer: Do lightly raced three-year-olds on their first handicap run after the winter show the same lift, given their marks rest on even thinner juvenile form?
because Three-year-olds grow and strengthen most over their first winter. A handicap mark set on a few two-year-old runs lags that growth even more than a three-year-old mark lags a four-year-old.
and the market misses it because The form shows a few modest juvenile runs and a long absence. Punters and the market anchor on that form, while trainers know how much the horse has strengthened.
predicting positive, A/E 1.08 to 1.15 on 600+ bets
and it is wrong if A/E below 1.03 on 600+ bets, or any period below 1.0
race_type = 'Flat' AND is_handicap = 1 AND month IN (3, 4, 5) AND age = 3 AND prior_runs <= 6 AND days_since_last_run >= 150
To answer: Is the edge strongest when the handicapper has left the mark unchanged since the three-year-old evidence, as a lagging-mark mechanism requires?
because If the edge comes from improvement the mark has not caught up with, it should be largest where the mark is unchanged or lower than last time. It should be smaller where the handicapper has already raised the horse over the winter.
and the market misses it because Punters treat an unchanged mark after a long break as neutral, and the long absence makes the horse look rusty. Neither shows the physical improvement the horse has made over the winter.
predicting positive, A/E 1.15 to 1.25 on 250+ bets
and it is wrong if A/E at or below 1.10 on 250+ bets, or raised-mark returners score as high
race_type = 'Flat' AND is_handicap = 1 AND month IN (3, 4, 5) AND age = 4 AND prior_runs <= 10 AND days_since_last_run >= 150 AND rating_change <= 0
H126 is still a published system: Flat handicaps, March to May, four-year-olds with 10 or fewer runs, first run after 150+ days. It scored A/E 1.13 on 1904 discovery bets, and the won-last-time subset scored 1.27 on 371. This round's three children all point the way the maturation story predicts, but none is strong. (1) Cutting to 5 or fewer runs gave 1.18 on 218 bets (p 0.24) against a control of 0.98, with 6/6 stable neighbours. That is above the parent's 1.13 but well short of the 700+ bets predicted, so it stays a niche. (2) Five-year-olds on the same rule gave 0.76 on 89 bets. That is consistent with age mattering more than freshness, but the confidence interval runs from 0.28 to 1.66, IRE scored 1.74 on 28 bets, and 89 bets is under a third of the 300 needed. The rival is not ruled out; there are simply too few such horses. (3) The second run after the break gave 1.06 on 459 bets (p 0.35), so any carry-over is weak at best. It also split sharply by period: 0.61 before 27 April 2026 against 1.16 after. That split matters across the thread. Since the post-October-2025 data covers only one spring, 'period' here really means 'early season vs late season'. Early spring is mostly all-weather racing, and turf takes over from April. So the lab's question is shifting from 'is it age?' to 'where in the season does the mark lag show?'. The mechanism that best fits everything so far is a mark set on thin three-year-old evidence. Under that mechanism, the edge should be strongest where the mark has not moved since that evidence and where the horse is least exposed. Pure freshness gets little support. None of the open questions can be closed: each relevant test either missed its sample size or has an interval that spans 1.0.
investigator v1
Tested16 Oct, 12:20
control: Control: H126 four-year-olds with 6-10 prior runs: A/E 1.21 on 301 bets, p = 0.14 (95%: 0.86 to 1.66), by period 1.30 · 1.09. pass
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Tested16 Oct, 12:20
partition: Partition: 3-5 prior runs only (data-window artefact check): A/E 1.20 on 193 bets, p = 0.22 (95%: 0.76 to 1.80), by period 1.32 · 1.02. pass
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Observed16 Oct, 12:20
H-625f1f2d (Fewest prior runs: the mark has the least information) passed discovery but was falsified: driven by one price band (up to 14/1): the rest score 1.00.
near miss0.50reasoned about 16 Oct, 12:23from H-625f1f2d
Tested16 Oct, 12:20
price_band: 14/1 or shorter, excluding the 8/1-14/1 hot band: A/E 0.99 on 66 bets, p = 0.55 (95%: 0.56 to 1.64). fail
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Tested16 Oct, 12:20
partition: Partition: already handicapped last time (not handicap debut): A/E 1.35 on 124 bets, p = 0.14 (95%: 0.77 to 2.20), by period 1.41 · 1.21. pass
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Tested16 Oct, 12:20
refinement: 14/1 or shorter (edges die at big prices; known 16/1+ longshot artefact): A/E 1.19 on 245 bets, p = 0.16 (95%: 0.85 to 1.62), by period 1.34 · 1.07. info
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Tested16 Oct, 12:20
refinement: Most lightly raced (5 or fewer runs: mark carries least information): A/E 1.11 on 156 bets, p = 0.37 (95%: 0.64 to 1.81), by period 1.07 · 1.15. info
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Observed16 Oct, 12:20
H-29eb96fe (Lightly raced four-year-olds back from the winter on turf only) passed discovery as a niche: A/E 1.2 on 379 bets (p 0.127), same-race control 0.97.
result0.50reasoned about 16 Oct, 12:23from H-29eb96fe
Tested16 Oct, 12:20
control: Two comparisons. (1) Same-race other runners on turf: expected near A/E 1.00 (H126's control was 0.98), showing: A/E 1.18 on 140 bets, p = 0.27 (95%: 0.71 to 1.84), by period 1.30. info
H-3449e2e4 (Lightly raced three-year-olds on their first run after the winter in h) pointed the predicted way but failed the bar: A/E 1.04 on 568 bets (p 0.405), control 1.0.
near miss0.52reasoned about 16 Oct, 12:23from H-3449e2e4
Observed16 Oct, 12:20
H-d1543740 (Lightly raced four-year-old returns whose mark has not been raised) was predicted positive but came out the other way: A/E 0.95 on 151 bets (p 0.498 for the opposite direction).
surprise0.80reasoned about 16 Oct, 12:23from H-d1543740
Tested16 Oct, 12:20
broad rule: A/E 1.04 on 568 bets, p = 0.41 (95%: 0.78 to 1.36), by period 1.02 · 1.08. fail
Lightly raced four-year-old returns whose mark has not been raisedDiscovery → KILLEDfails: A/E 0.953, needs 1.1; p 0.606, needs 0.05 or lower; periods [1.04, 0.86], each needs 1.0; A/E 0.953 against same-race control 1.01, needs 0.03 above (not a niche pass: A/E 0.953 below 1.1; p 0.6059551900052518 above 0.5; does not beat the same-race control 1.01; a period points the other way [1.04, 0.86])
Tested16 Oct, 12:22
control: Surface flip: identical selection on the all-weather: A/E 1.18 on 140 bets, p = 0.27 (95%: 0.71 to 1.84), by period 1.30. pass
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Tested16 Oct, 12:22
control: Not fresh: lightly raced 4yo spring turf handicappers who ran within 60 days: A/E 0.99 on 647 bets, p = 0.55 (95%: 0.77 to 1.25), by period 0.84 · 1.04. fail
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Tested16 Oct, 12:22
partition: April-May turf at 14/1 or shorter (calendar-matched, no longshots): A/E 1.14 on 221 bets, p = 0.23 (95%: 0.80 to 1.58), by period 1.25 · 1.07. pass
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Observed16 Oct, 12:22
H-29eb96fe (Lightly raced four-year-olds back from the winter on turf only) survived falsification: 5 of 6 attacks left it standing; it is frozen for its one vault query.
result0.55reasoned about 16 Oct, 12:23from H-29eb96fe
Tested16 Oct, 12:22
price_band: Shorter-priced half only (8/1 or shorter): A/E 1.11 on 143 bets, p = 0.31 (95%: 0.75 to 1.58), by period 1.15 · 1.07. pass
H-29eb96fe (Lightly raced four-year-olds back from the winter on turf only) met the sealed vault supportive: A/E 1.1 on 1741 bets (p 0.102, threshold 0.0033).
result0.55reasoned about 16 Oct, 12:23from H-29eb96fe
Tested16 Oct, 12:22
vault query (supportive): A/E 1.10 on 1,741 bets, p = 0.10 (95%: 0.95 to 1.27), by period 1.19 · 1.02. pass
Is the Irish part of H126 weaker for unraised horses because the Irish handicapper treats light three-year-old form differently?
Unraised Irish returners won 0 of 42, while Irish turf returners overall scored 1.47 on 121. Region interacts with rating_change. A test reads region, rating_change and surface. The samples are small, so information gain is limited.
openpriority 0.25raised by O-fa53ff56, O-94a1c4ce
New question16 Oct, 12:23
Do lightly raced four-year-old returners whose mark was raised during their three-year-old season carry the H126 edge, as an improver-trajectory mechanism would predict?
The unraised-mark cut came out the opposite way (0.95 on 151), while the won-last-time system scores 1.27. If raised horses carry the lift, the mechanism is the market doubting continued progression, not a stale mark. A test reads rating_change, last_official_rating, last_position, prior_runs, days_since_last_run and age.
testingpriority 0.60raised by O-fa53ff56, O-8d9fdbf6
To answer: Do lightly raced four-year-old returners whose mark was raised during their three-year-old season carry the H126 edge, as an improver-trajectory mechanism would predict?
because A three-year-old that earned rises is on an upward curve. With few runs and a winter to strengthen, it often keeps improving faster than the handicapper raises it. Horses whose mark stood still or fell had shown they were not progressing.
and the market misses it because Punters see a career-high mark after a long absence and expect the horse to need the run or to be held by the handicapper. The form book shows a raised horse coming back off a break, which looks like a negative.
predicting positive, A/E 1.15 to 1.25 on 600+ bets
and it is wrong if A/E below 1.08 on 600+ bets, or no better than the unraised group's 0.95 plus 0.10
race_type = 'Flat' AND is_handicap = 1 AND month IN (3, 4, 5) AND age = 4 AND prior_runs <= 10 AND days_since_last_run >= 150 AND rating_change > 0
To answer: Is the H126 lift and its second-run follow-up a late-spring turf effect, weak or absent on the all-weather in March and early April?
because If the cause is maturation plus the mark, it should not depend on surface. If it is going or late-spring turf readiness, all-weather returners should score about 1.0.
and the market misses it because Early-season all-weather handicaps are dominated by fit, race-hardened horses. The market prices fresh returners against that sharpness and may over-discount the winter's physical gain.
predicting positive, A/E 1.05 to 1.12 on 400+ bets
and it is wrong if A/E below 1.00 on 400+ bets, which would favour the turf-only explanation and argue for restricting the system
race_type = 'Flat' AND is_handicap = 1 AND month IN (3, 4, 5) AND age = 4 AND prior_runs <= 10 AND days_since_last_run >= 150 AND surface = 'AW'
The thread's core holds: lightly raced four-year-olds returning in spring Flat handicaps beat their price (H126 1.13 on 1904 in discovery). Narrowing it to turf (H-29eb96fe) gave 1.20 on 379 in discovery and survived 5 of 6 attacks. In the sealed vault it was supportive but not significant: 1.10 on 1741 bets, p 0.10. Every way we tried to locate the cause in a 'lagging mark' failed. The cut to 5 or fewer runs scored 1.18 on 218 but was killed because one price band (up to 14/1) carried it. The second run after the break was weak (1.06 on 459). Three-year-olds did not replicate it (1.04 on 568, 16/1+ at 0.58). The sharpest test came out backwards: four-year-olds whose mark had NOT been raised scored 0.95 on 151, against a same-race control of 1.01, with Ireland 0 wins from 42. This is a discovery. The lagging-mark story says unraised horses should carry the edge, and they do not. The other evidence points to trajectory instead. The published won-last-time cut scores 1.27 on 371, and the up-to-14/1 concentration fits too. The edge seems to sit in horses that were already improving at three and were raised for it, and the market seems to doubt that the improvement carries on. Five-year-olds (89 bets) are still too few to settle age against freshness. Turf on its own is supportive, but the all-weather part has not been shown to be weaker. A rough subtraction from the H126 totals puts the non-turf remainder near 1.1. That figure is approximate and should not be cited.
investigator v1
Answered16 Oct, 12:23
If the horses are really improving, does the edge carry on to their second run after the winter break?
Only weakly. The second run after the break scored A/E 1.06 on 459 bets (p 0.35), and it all came from the late-season period (0.61 before, 1.16 after). That is not the persistent lift that maturation would predict.
Is the edge strongest when the handicapper has left the mark unchanged since the three-year-old evidence, as a lagging-mark mechanism requires?
No. Returners whose mark has not been raised scored A/E 0.95 on 151 bets, against a same-race control of 1.01 (Ireland 0 wins from 42). The edge does not sit with unchanged marks, so a simple lagging-mark mechanism is not supported.
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Observed16 Oct, 12:28
H-568b7703 (Lightly raced four-year-old returners whose mark was raised: improvers) passed discovery as a niche: A/E 1.38 on 41 bets (p 0.231), same-race control 0.94.
result0.50reasoned about 16 Oct, 12:33from H-568b7703
Tested16 Oct, 12:28
refinement: 14/1 or shorter: A/E 1.24 on 35 bets, p = 0.34 (95%: 0.50 to 2.54). info
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Tested16 Oct, 12:28
control: Same-race other runners should score about 1.00 (H126 control ran 0.98 to 1.01), so any lift is not a property : A/E 0.95 on 151 bets, p = 0.61 (95%: 0.52 to 1.60), by period 1.04 · 0.86. info
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Observed16 Oct, 12:28
H-b05a33c3 (Lightly raced four-year-old returners on the all-weather: is the edge ) passed discovery as a niche: A/E 1.18 on 140 bets (p 0.268), same-race control 0.97.
result0.50reasoned about 16 Oct, 12:33from H-b05a33c3
Tested16 Oct, 12:28
control: Same-race other runners: expected about 1.00, since early-season AW handicaps carry no general mispricing. If t: A/E 0.99 on 198 bets, p = 0.55 (95%: 0.53 to 1.70), by period 0.70 · 1.98. info
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Observed16 Oct, 12:29
H-568b7703 (Lightly raced four-year-old returners whose mark was raised: improvers) survived falsification: 1 of 1 attacks left it standing; it is frozen for its one vault query.
result0.55reasoned about 16 Oct, 12:33from H-568b7703
Tested16 Oct, 12:29
perturbation: Thresholds tightened: prior_runs <= 8, days off >= 180: A/E 0.98 on 13 bets, p = 0.61 (95%: 0.12 to 3.53). info
H-b05a33c3 (Lightly raced four-year-old returners on the all-weather: is the edge ) survived falsification: 4 of 5 attacks left it standing; it is frozen for its one vault query.
result0.55reasoned about 16 Oct, 12:33from H-b05a33c3
Tested16 Oct, 12:29
partition: April-May only: removes the March calendar confound: A/E 0.85 on 88 bets, p = 0.73 (95%: 0.39 to 1.62), by period 0.93. fail
H-b05a33c3 (Lightly raced four-year-old returners on the all-weather: is the edge ) met the sealed vault supportive: A/E 1.1 on 698 bets (p 0.2014, threshold 0.0032).
result0.55reasoned about 16 Oct, 12:33from H-b05a33c3
Tested16 Oct, 12:29
vault query (supportive): A/E 1.10 on 698 bets, p = 0.20 (95%: 0.88 to 1.36), by period 1.26 · 0.98. pass
Does the lift sit mainly in the market's first four or five in the betting, with returners at 16/1 or bigger simply unready?
The all-weather version had 0 wins from 48 bets at 16/1 or bigger (O-a51c66ea), and the fewest-runs cut lived in one price band. This repeated pattern should be described before anyone slices on price. Columns: sp_rank, sp_decimal, days_since_last_run.
openpriority 0.30raised by O-a51c66ea
New question16 Oct, 12:33
Is the returner lift concentrated in a few big yards whose four-year-olds return ready, rather than spread across trainers?
Discovery samples are small (140 all-weather bets, 41 raised-mark bets), so a handful of yards could carry them. A trainer concentration scan on the H126 base would rule a yard effect in or out. Columns: trainer, prior_runs, days_since_last_run, sp_rank.
openpriority 0.35raised by O-a51c66ea, O-f62c7b90
New question16 Oct, 12:33
Is the lift a property of the handicap mark, or does it also show in Flat non-handicaps, where there is no mark to lag?
The raised-mark cut (O-f62c7b90) points to the market doubting continued progress, not to the mark. If that is right, lightly raced four-year-old returners in conditions and maiden races should also beat their prices. Columns: is_handicap, is_maiden, age, prior_runs, days_since_last_run, race_class.
testingpriority 0.45raised by O-f62c7b90, O-6bb82233
New question16 Oct, 12:33
Does the four-year-old returner lift depend on being lightly raced, or do exposed four-year-olds (more than 10 prior runs) back after 150+ days in spring handicaps score the same?
Both surfaces met the vault at about 1.10 (O-3b4db8ee, O-bbb9c77c), and the stale-mark story failed. If exposed four-year-olds also score about 1.10, the cause is a spring-freshness mispricing for this age, not unexposed improvement. Columns: age, prior_runs, days_since_last_run, is_handicap, month, race_type.
testingpriority 0.50raised by O-3b4db8ee, O-a51c66ea
To answer: Is the lift a property of the handicap mark, or does it also show in Flat non-handicaps, where there is no mark to lag?
because A lightly raced four-year-old can keep strengthening over the winter. The market anchors on its last visible three-year-old form whatever the race conditions, so the progress should be underpriced in conditions races too.
and the market misses it because The form book shows only three-year-old runs that are months old, and there is no fresh evidence of the winter's development. In non-handicaps, the weight-for-age and penalty structure hides this further.
predicting positive, A/E 1.05 to 1.12 on 300+ bets
and it is wrong if A/E below 1.00 on 300+ bets. That would tie the effect to handicaps and the mark, not to underrated progress.
race_type = 'Flat' AND is_handicap = 0 AND month IN (3, 4, 5) AND age = 4 AND prior_runs <= 10 AND days_since_last_run >= 150
To answer: Does the four-year-old returner lift depend on being lightly raced, or do exposed four-year-olds (more than 10 prior runs) back after 150+ days in spring handicaps score the same?
because If the cause is unexposed improvement over the winter, horses that already have plenty of runs have little left to show and should be priced correctly. If they score just as well, the cause is a general spring-freshness mispricing for four-year-olds.
and the market misses it because For an exposed horse the form book is full and reliable. The market should price it well unless it applies a blanket discount to anything returning from a long break.
predicting negative, A/E 0.95 to 1.03 on 300+ bets (no edge), well below the lightly raced 1.10
and it is wrong if Exposed four-year-old returners score A/E 1.08 or more on 300+ bets. That would mean 'lightly raced' is not the cause.
race_type = 'Flat' AND is_handicap = 1 AND month IN (3, 4, 5) AND age = 4 AND prior_runs > 10 AND days_since_last_run >= 150
The thread has settled into one steady but modest picture. In the sealed vault, lightly raced four-year-olds back from the winter scored A/E 1.10 on 1741 turf bets (p 0.10) and A/E 1.10 on 698 all-weather bets (p 0.20). Neither is significant, but the two surfaces now agree. So the surface split I suspected from the second-run period swing (0.61 then 1.16) is not what divides the effect. Discovery figures were higher on small samples (turf 1.20 on 379, all-weather 1.18 on 140) and shrank toward about 1.10 on the larger vault samples, which is what a real but small edge looks like.
The mechanism chain is clearer than the size. The stale-mark story failed: unraised marks scored 0.95 on 151. Three-year-olds (1.04 on 568) and the second run back (1.06 on 459) both failed. The raised-mark improver cut (1.38 on 41) survived its one attack and is frozen, but 41 bets is far too few to judge. Across cuts, the lift keeps sitting in the shorter, credible price bands: the all-weather version had 0 wins from 48 bets at 16/1 or bigger, and the fewest-runs cut was killed for depending on one price band. The best current reading is that the market underrates a lightly raced four-year-old's progress over the winter, mainly when the horse is fancied enough to be ready. A mark set too low is not the cause.
Two things are still open. First, age versus freshness: the five-year-old control had only 89 bets. Second, whether 'lightly raced' matters at all. We have only one spring of data, so the forward record in spring 2027 will judge the live system.
investigator v1
Answered16 Oct, 12:33
Do lightly raced three-year-olds on their first handicap run after the winter show the same lift, given their marks rest on even thinner juvenile form?
No. Lightly raced three-year-olds back from the winter scored A/E 1.04 on 568 bets (16/1 or bigger at 0.58) and failed the bar, so the effect does not extend one year younger.
Is the H126 lift and its second-run follow-up a late-spring turf effect, weak or absent on the all-weather in March and early April?
No. The all-weather version met the vault at A/E 1.10 on 698 bets, the same as turf at 1.10 on 1741, so the lift is not a late-spring turf effect and the published system need not be restricted to turf.
H-ee2c7a96 (Exposed four-year-olds back from the winter in spring handicaps: is li) pointed the predicted way but failed the bar: A/E 0.8 on 191 bets (p 0.814), control 1.02.
near miss0.70reasoned about 16 Oct, 12:36from H-ee2c7a96
Tested16 Oct, 12:35
broad rule (niche): A/E 1.07 on 182 bets, p = 0.38 (95%: 0.70 to 1.57), by period 1.00 · 1.17. pass
H-1a3e129f (Lightly raced four-year-old returners in Flat non-handicaps: progress ): the country 'GB' carries the effect (A/E 1.21) while the rest score IRE 0.86.
concentration0.55reasoned about 16 Oct, 12:36from H-1a3e129f
Observed16 Oct, 12:35
H-1a3e129f (Lightly raced four-year-old returners in Flat non-handicaps: progress ) passed discovery as a niche: A/E 1.07 on 182 bets (p 0.384), same-race control 0.96.
result0.50reasoned about 16 Oct, 12:36from H-1a3e129f
Tested16 Oct, 12:35
refinement: 14/1 or shorter: A/E 1.04 on 118 bets, p = 0.46 (95%: 0.67 to 1.54), by period 0.97 · 1.13. info
Is the four-year-old spring returner edge a British phenomenon: do Irish runners in the H126 base score at or below their prices while British runners carry the whole lift?
Ireland is weak in every cut that reports a country split: non-handicaps 0.86 on 75 against GB 1.21 on 107 (O-b0af52cb), exposed returners 0 from 21 (O-2b083477), and earlier the unraised-mark and all-weather cuts. If Ireland scores 1.0 or less on the full base, the live system and the shadow candidates should be restricted to GB. A test reads region, trainer, sp_decimal, days_since_last_run, prior_runs and age.
answeredpriority 0.55raised by O-b0af52cb, O-2b083477
New question16 Oct, 12:36
With more data, do lightly raced four-year-old returners in Flat non-handicaps beat their prices at 16/1 or shorter, or is the 1.07 just two longshot winners?
The non-handicap niche (O-c3a86a60) is the only test of whether the effect exists without a mark. Its BSP ROI rests on 2 winners from 64 at 16/1 or bigger. This needs forward data, not slicing now. Columns: is_handicap, is_maiden, is_novice, race_class, sp_decimal, region.
openpriority 0.30raised by O-c3a86a60, O-b0af52cb
New question16 Oct, 12:36
Does A/E fall steadily as prior_runs rises among four-year-old spring returners (0-5, 6-10, 11-15, 16+), or is there a step at about 10 runs?
Exposed returners scored 0.80 on 191 against the base's 1.13, but the fewest-runs cut failed on price-band concentration. A descriptive scan of A/E by prior_runs band would show whether light racing works as a gradient (supporting an unexposed-improver story) or whether the 10-run cut-off is arbitrary. Columns: prior_runs, flat_runs, age, days_since_last_run, month, is_handicap, sp_decimal.
To answer: Is the four-year-old spring returner edge a British phenomenon: do Irish runners in the H126 base score at or below their prices while British runners carry the whole lift?
because Irish Flat racing is concentrated in a few big yards. Their lightly raced four-year-olds are closely followed, and their winter progress is well signalled by stable form and market moves. British returners come from a wider spread of yards, where winter improvement is less visible to punters.
and the market misses it because In Britain the form book shows only three-year-old runs and a long absence, so the market leans on old form. In Ireland the same information is priced in through stable reputation and a well-informed market, so there is nothing left to miss.
predicting negative, A/E 0.80 to 1.00 on 300+ Irish bets, against about 1.15 for GB in the same base
and it is wrong if Irish runners score A/E 1.08 or more on 300+ bets, or match the GB figure within 0.05
race_type = 'Flat' AND is_handicap = 1 AND month IN (3, 4, 5) AND age = 4 AND prior_runs <= 10 AND days_since_last_run >= 150 AND region = 'IRE'
The thread still has one robust core. The H126 base (Flat handicaps, March to May, four-year-olds with 10 or fewer runs, back after 150+ days) scored 1.13 on 1904 bets in discovery. In the vault it shrank to about 1.10 on both surfaces. Every attempt to explain it through the handicap mark has failed: the unraised-mark cut scored 0.95 and the fewest-runs cut sat in one price band. Three-year-olds and the second run back did not carry it either.
Two new tests checked the boundaries.
- **Exposed four-year-olds** (more than 10 prior runs) returning in spring handicaps scored A/E 0.80 on 191 bets against a control of 1.02 (O-2b083477). Those at 16/1 or bigger went 0 from 86. This leans towards being lightly raced mattering, so the effect is not simply spring freshness at this age. But the confidence interval (0.40 to 1.42) still includes 1.10, so it is a lean, not proof.
- **Non-handicaps** (O-c3a86a60): lightly raced four-year-old returners scored 1.07 on 182 bets (p 0.38, control 0.96). That fits a "market doubts continued progress" story that does not depend on a mark. But the +54.5% BSP ROI rests on 2 winners from 64 at 16/1 or bigger, so it is weak evidence.
A pattern is repeating across separate cuts: Ireland underperforms. In non-handicaps GB scored 1.21 on 107 and Ireland 0.86 on 75 (O-b0af52cb). Among exposed horses Ireland went 0 from 21. Earlier, the unraised-mark cut had Ireland at 0 from 42 and the all-weather cut had Ireland at 0 from 14. Whether the edge is a British phenomenon is now the most useful structural question. It should be tested on the Irish side of the existing base, where the mechanism makes a prediction, not by cutting the near miss into slices.
investigator v1
Tested16 Oct, 12:43
partition: 14/1 or shorter only: strip the 2 long-price winners: A/E 1.04 on 118 bets, p = 0.46 (95%: 0.67 to 1.54), by period 0.97 · 1.13. pass
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Observed16 Oct, 12:43
H-52d3c28e (Irish lightly raced four-year-old returners in spring handicaps are fu) was predicted negative but came out the other way: A/E 1.35 on 135 bets (p 0.212 for the opposite direction).
surprise0.90reasoned about 16 Oct, 12:45from H-52d3c28e
Observed16 Oct, 12:43
H-1a3e129f (Lightly raced four-year-old returners in Flat non-handicaps: progress ) passed discovery but was falsified: IRE A/E 0.861; driven by one price band (up to 3/1): the rest score 1.00.
near miss0.50reasoned about 16 Oct, 12:45from H-1a3e129f
Tested16 Oct, 12:43
control: Control: same four-year-olds that are not returners (14 to 60 days): A/E 0.95 on 218 bets, p = 0.62 (95%: 0.54 to 1.54), by period 0.87 · 0.98. fail
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
New question16 Oct, 12:45
Within the H126 base, is the lift confined to returners priced 14/1 or shorter, with 16/1+ returners at or below their prices in both countries and on both surfaces?
Across separate cuts the long-priced returners blank: Irish 16/1+ 1 win from 63 (A/E 0.77) against up-to-14/1 2.28 on 49; all-weather 16/1+ 0 from 48; exposed 16/1+ 0 from 86. The H-625f1f2d kill was also driven by the up-to-14/1 band. A test reads sp_decimal and sp_rank within the H126 predicate (age, prior_runs, days_since_last_run, is_handicap, month), split by region and surface. If the effect is entirely a price-band effect, the lab should publish it with a price cap or drop it.
answeredpriority 0.60raised by O-6b28e152, O-d5ec2505
New question16 Oct, 12:45
Do short-priced lightly raced four-year-old returners in non-handicaps (up to 3/1) beat their prices because the yard's confidence shows in the market, or is that band one of many segments?
O-d5ec2505 killed the non-handicap version because only the up-to-3/1 band carried it (rest 1.00). If short-priced returners also do best in handicaps, the shared cause is market-signalled readiness, not the mark. A test reads sp_decimal, sp_rank, is_handicap, race_type, days_since_last_run, prior_runs and age.
To answer: Within the H126 base, is the lift confined to returners priced 14/1 or shorter, with 16/1+ returners at or below their prices in both countries and on both surfaces?
because A lightly raced four-year-old back from a winter break may have strengthened physically. Only some are fully wound up first time out, and the stable's confidence (gallops reports, money) pushes the ready ones to credible prices. The ones that need the run drift to 16/1 or bigger and rarely win.
and the market misses it because Punters treat a 150+ day absence and a thin three-year-old record as uncertainty and price for some regression. They do not credit enough the improvement of a fit, maturing horse whose form figures come from when it was weaker. The market gets the direction right (shorter means readier) but not the size.
predicting positive, A/E 1.15 to 1.30 on 900+ bets, with 16/1+ returners at or below 0.90
and it is wrong if A/E below 1.08 on 900+ bets, or either country or either surface below 1.0, or the 16/1+ complement scores as high as the capped group
race_type = 'Flat' AND is_handicap = 1 AND month IN (3, 4, 5) AND age = 4 AND prior_runs <= 10 AND days_since_last_run >= 150 AND sp_decimal <= 15
To answer: Do short-priced lightly raced four-year-old returners in non-handicaps (up to 3/1) beat their prices because the yard's confidence shows in the market, or is that band one of many segments?
because If market-signalled readiness is the cause, it needs no handicap mark. Fancied lightly raced four-year-olds back from a break should beat their prices in handicaps and non-handicaps alike.
and the market misses it because The absence and the thin record make the price drift against strong stable confidence. The headline 'first run for 150 days' weighs heavier with the public than the yard's money does.
predicting positive, A/E 1.10 to 1.20 on 500+ bets across both race kinds
and it is wrong if A/E below 1.05 on 500+ bets, or the non-handicap part below 1.0 while handicaps carry it all, which would put the mark back as the cause
race_type = 'Flat' AND month IN (3, 4, 5) AND age = 4 AND prior_runs <= 10 AND days_since_last_run >= 150 AND sp_rank <= 3
The thread still holds one thing: lightly raced four-year-olds on their first spring handicap run after 150+ days score A/E 1.13 on 1904. The vault puts it at about 1.10 on both surfaces. Most attempts to explain why have failed. It did not carry to three-year-olds (1.04), to the second run back (1.06), to five-year-olds (0.76 on 89), to exposed four-year-olds (0.80 on 191) or to unraised marks (0.95 on 151). So the simple stale-mark story is weak.
Two new results:
- **Ireland is not a boundary.** We predicted Irish handicap returners would be over-bet. They came out the other way: A/E 1.35 on 135 (10 wins, CI 0.65 to 2.49). The Irish weakness seen before belongs to the side cuts (non-handicaps, unraised marks, exposed horses), not to the core handicap base.
- **Non-handicaps carry no general edge.** That version was killed on Irish 0.86 and because only the up-to-3/1 band carried it; the rest scored 1.00.
The pattern across the thread is about price. The lift keeps sitting at credible prices: Irish up-to-14/1 at 2.28 on 49 (7 wins) against 16/1+ at 0.77 on 63 (1 win); all-weather 16/1+ 0 from 48; exposed 16/1+ 0 from 86. The H-625f1f2d kill was also 'driven by up to 14/1'. That points to a readiness story: the market price on return tells you which yards have their horse fit. Long-priced returners are the ones that need the run, and they are rightly dismissed. This is a lead built from several cuts, not a finding.
investigator v1
Answered16 Oct, 12:45
Is the four-year-old spring returner edge a British phenomenon: do Irish runners in the H126 base score at or below their prices while British runners carry the whole lift?
No: Irish runners in the H126 handicap base scored A/E 1.35 on 135 bets (p 0.21 against the over-bet prediction). Irish weakness is confined to the non-handicap, unraised-mark and exposed cuts, not the core effect.
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Observed16 Oct, 12:49
H-1a6f83c9 (Fancied lightly raced four-year-old returners: the market flags which ) passed discovery as a niche: A/E 1.21 on 337 bets (p 0.083), same-race control 0.95.
result0.50reasoned about 16 Oct, 13:02from H-1a6f83c9
Tested16 Oct, 12:49
control: Two comparisons. (1) Inside the H126 base, the complement priced 16/1 or bigger (same rule but sp_decimal > 15): A/E 1.00 on 7,052 bets, p = 0.53 (95%: 0.94 to 1.06), by period 0.99 · 1.01. info
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Tested16 Oct, 12:49
refinement: non-handicap part only (the rival test: readiness without a mark): A/E 1.14 on 83 bets, p = 0.29 (95%: 0.72 to 1.72), by period 1.11 · 1.19. info
H-3c53e905 (Short-priced first-time-out four-year-olds: rival to the handicap-mark): the country 'GB' carries the effect (A/E 1.24) while the rest score IRE 0.89.
concentration0.55reasoned about 16 Oct, 13:02from H-3c53e905
Observed16 Oct, 12:49
H-3c53e905 (Short-priced first-time-out four-year-olds: rival to the handicap-mark) passed discovery as a niche: A/E 1.18 on 251 bets (p 0.108), same-race control 0.93.
result0.50reasoned about 16 Oct, 13:02from H-3c53e905
Tested16 Oct, 12:49
refinement: handicap part only (comparison to the non-handicap part): A/E 1.19 on 168 bets, p = 0.14 (95%: 0.86 to 1.61), by period 1.28 · 1.09. info
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Tested16 Oct, 12:49
refinement: GB only (Ireland has repeatedly blanked in the non-handicap and side cuts): A/E 1.24 on 201 bets, p = 0.065 (95%: 0.94 to 1.60), by period 1.39 · 1.07. info
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Tested16 Oct, 12:49
control: Same-race other runners should score about 1.0 or a little below, because the field absorbs the overround and t: A/E 1.11 on 450 bets, p = 0.34 (95%: 0.70 to 1.66), by period 1.13 · 1.07. info
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Tested16 Oct, 12:51
control: Control: same fancied lightly raced 4yo, NOT returning from a break: A/E 1.00 on 786 bets, p = 0.51 (95%: 0.83 to 1.19), by period 0.85 · 1.15. pass
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Observed16 Oct, 12:51
H-1a6f83c9 (Fancied lightly raced four-year-old returners: the market flags which ) survived falsification: 4 of 5 attacks left it standing; it is frozen for its one vault query.
result0.55reasoned about 16 Oct, 13:02from H-1a6f83c9
Tested16 Oct, 12:51
price_band: Strongest readiness signal only: 7/1 or shorter: A/E 1.11 on 180 bets, p = 0.27 (95%: 0.80 to 1.50), by period 1.21 · 0.99. pass
H-1a6f83c9 (Fancied lightly raced four-year-old returners: the market flags which ) met the sealed vault supportive: A/E 1.11 on 1567 bets (p 0.0484, threshold 0.0031).
result0.55reasoned about 16 Oct, 13:02from H-1a6f83c9
Tested16 Oct, 12:51
vault query (supportive): A/E 1.11 on 1,567 bets, p = 0.048 (95%: 0.98 to 1.26), by period 1.23 · 1.01. pass
H-3c53e905 (Short-priced first-time-out four-year-olds: rival to the handicap-mark) passed discovery but was falsified: IRE A/E 0.887.
near miss0.50reasoned about 16 Oct, 13:02from H-3c53e905
Tested16 Oct, 12:51
control: Control: fancied lightly raced 4yo in spring that are NOT returning from a break: A/E 0.97 on 463 bets, p = 0.62 (95%: 0.79 to 1.19), by period 0.74 · 1.15. fail
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
New question16 Oct, 13:02
Is the four-year-old returner lift concentrated in March and April, fading by May when late returners are more likely to be horses that had setbacks and the market has seen how the crop came out of the winter?
Every cut decays from period 1 (to 27 April) to period 2. The capped rule went 1.36 to 1.02; at 180+ days off, 1.46 to 0.97; top three in the betting, 1.22 to 1.12. A test reads month, days_since_last_run, age, prior_runs and is_handicap. If March and April carry it, the published rule should narrow. If May matches, the decay is noise from one spring.
answeredpriority 0.55raised by O-b22a8ae7, O-da190682, O-c4322e77
New question16 Oct, 13:02
Are Irish splits of under 100 bets in this thread informative at all, given they flip from 1.48 to 0.73 across overlapping cuts?
H-3c53e905 was killed on Ireland at 0.887 on 50 bets. The sibling capped rule scored Ireland 1.48 on 72 (1.60 on 50 when tightened to 11/1). A test would compare, by region, the Irish A/E of the H126 base against its top three in the betting (sp_rank) and its 14/1-or-shorter subset (sp_decimal), with bootstrap intervals. This is a method question for the falsification gate more than for the edge.
openpriority 0.30raised by O-9dd3148f, O-f503c681, O-b22a8ae7
New question16 Oct, 13:02
Within the H126 handicap base, do returners priced 16/1 or bigger (sp_decimal > 15) score at or below 0.90, or as well as the capped group?
The capped rule's planned complement control was never run. Unfancied returners across both race kinds scored 1.11 on 450, while earlier side cuts blanked at 16/1+. A test reads sp_decimal, sp_rank and is_handicap within the H126 base. The answer decides whether a price filter belongs in the published system at all.
testingpriority 0.50raised by O-c4322e77, O-da190682
To answer: Within the H126 handicap base, do returners priced 16/1 or bigger (sp_decimal > 15) score at or below 0.90, or as well as the capped group?
because If stable confidence moves the ready horses' prices, the drifters are the horses needing the run and should win about as often as their prices say, or less. If instead the cause is winter improvement that the mark misses, the improvement is spread across all returners and the long-priced ones should beat their prices too.
and the market misses it because Under the readiness story, a long price correctly reflects an unfit horse. Under the improvement story, the long price rests on old three-year-old figures, and the form book shows weak runs that understate a stronger horse.
predicting negative, A/E 0.80 to 0.95 on 500+ bets
and it is wrong if A/E of 1.05 or higher on 500+ bets, which would mean the price cap is not where the lift lives and the whole base should be kept
race_type = 'Flat' AND is_handicap = 1 AND month IN (3, 4, 5) AND age = 4 AND prior_runs <= 10 AND days_since_last_run >= 150 AND sp_decimal > 15
To answer: Is the four-year-old returner lift concentrated in March and April, fading by May when late returners are more likely to be horses that had setbacks and the market has seen how the crop came out of the winter?
because A horse that has strengthened over the winter is still unknown to the market when it first reappears in March or April. By May the four-year-old crop has been seen, trainers' form with returners is visible, and the remaining first-time-out returners include more horses whose return was delayed by a setback.
and the market misses it because Early in the season the price for a returner rests on its three-year-old form and its mark. Little current evidence exists on how the crop has developed. Later the market updates from the other returners' results, and the horses still returning are partly those that had problems.
predicting positive, A/E 1.12 to 1.22 on 1000+ bets across all data, with May returners clearly lower (about 1.00)
and it is wrong if A/E below 1.08 on the March-April rule, or May returners scoring within 0.05 of it
race_type = 'Flat' AND is_handicap = 1 AND month IN (3, 4) AND age = 4 AND prior_runs <= 10 AND days_since_last_run >= 150
To answer: Is the H126 lift caused by four-year-old physical maturation that the handicap mark hasn't caught up with, or just by lightly raced horses returning fresh at any age?
because Physical maturation between three and four is the proposed cause. Horses aged five and up have finished maturing, so a long break should give them freshness but not a mark that lags behind improvement.
and the market misses it because If older lightly raced returners also beat their prices, the miss is a general under-pricing of first runs back, not something specific to four-year-olds. The five-year-old test (0.76 on 89) was too thin to tell.
predicting negative, A/E 0.88 to 1.02 on 300+ bets
and it is wrong if A/E of 1.08 or higher on 300+ bets, which would make age irrelevant and point to a general freshness mispricing
race_type = 'Flat' AND is_handicap = 1 AND month IN (3, 4, 5) AND age >= 5 AND prior_runs <= 10 AND days_since_last_run >= 150
The thread keeps coming back to the same figure: lightly raced four-year-old spring handicap returners beat their prices by about 10%, and slicing the group does not concentrate that edge reliably. The price-capped version (H-1a6f83c9, SP 14/1 or shorter) scored 1.21 on 337 in discovery and held 4 of 5 attacks. On the sealed vault it fell to 1.11 on 1567 bets (p 0.048 against a threshold of 0.003), which is about the parent H126 level of 1.13. The cap shows the same shrinkage as every earlier cut, so it adds little.
The strongest evidence now is the set of non-returner controls, and they point to the break itself. The same profile of fancied, lightly raced spring four-year-olds that were not returning scored 1.00 on 786 (14/1 or shorter) and 0.97 on 463 (top three in the betting, 7 to 60 days off). Price-matched spring handicappers not returning scored 1.00 on 7052. So the lift is not favourite-longshot bias, and it is not an age-in-spring effect. It belongs to the first run back.
The idea that the market's price marks which returners are ready is weaker than it looked. In H-3c53e905's own control, unfancied returners (outside the top three in the betting) scored 1.11 on 450, and 1.19 on 245 at 16/1 or bigger. Widening the cap to 17/1 added 3 winners from 18 bets. The earlier long-price blanks came from side cuts.
H-3c53e905 (top three in the betting, handicaps and non-handicaps) was killed on Ireland at 0.89 on 50 bets. Ireland flips sign across overlapping cuts: 1.48 on 72 in the capped rule, 0.73 on 21 among top-three handicappers. Irish splits of 20 to 80 bets are noise, not a mechanism.
One pattern repeats across every cut: period 1 (to 27 April 2026) scores 1.3 to 1.46 and period 2 (late April to May) scores 0.97 to 1.10. The 180+ day version decays the most (1.46 to 0.97). That is a lead about timing within the spring, not yet a finding.
investigator v1
Answered16 Oct, 13:02
Within the H126 base, is the lift confined to returners priced 14/1 or shorter, with 16/1+ returners at or below their prices in both countries and on both surfaces?
No. The 14/1 cap met the vault at only 1.11 on 1567, about the uncapped base level. Unfancied returners scored 1.11 on 450 (1.19 on 245 at 16/1+), so the lift is not confined to the shorter prices.
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Tested16 Oct, 13:07
refinement: Turf only (tests the surface-mix confounder: is the March-April lift really the turf-season opening?): A/E 1.27 on 204 bets, p = 0.15 (95%: 0.81 to 1.91), by period 1.27. info
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Tested16 Oct, 13:07
refinement: All-weather only (if the lift holds here too, the timing story is not a turf-opening artefact): A/E 1.30 on 118 bets, p = 0.16 (95%: 0.77 to 2.06), by period 1.30. info
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Observed16 Oct, 13:07
H-2fa8d5cc (Lightly raced older handicappers back from a break: a powered age cont) pointed the predicted way but failed the bar: A/E 0.64 on 119 bets (p 0.908), control 1.04.
near miss0.60reasoned about 16 Oct, 13:09from H-2fa8d5cc
Observed16 Oct, 13:07
H-2d115a58 (Early-spring four-year-old returners: the lift sits in March and April) passed discovery as a niche: A/E 1.29 on 322 bets (p 0.068), same-race control 0.95.
result0.50reasoned about 16 Oct, 13:09from H-2d115a58
Tested16 Oct, 13:07
broad rule: A/E 0.99 on 182 bets, p = 0.57 (95%: 0.27 to 2.55), by period 0.41 · 1.86. fail
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Tested16 Oct, 13:07
control: The key comparison is the identical profile returning in May: lightly raced four-year-old handicap returners af: A/E 1.06 on 197 bets, p = 0.43 (95%: 0.66 to 1.60), by period 1.06. info
Do lightly raced three-year-old handicap returners in spring show the same lift, or is the mispricing specific to four-year-olds?
The older-horse control (O-6ea233f1, 0.64 on 119) points toward an age-specific effect but is underpowered. Three-year-olds are the stronger contrast: they mature over the winter even more, yet the market expects them to improve. If they score about 1.0 while four-year-olds score 1.13, the edge lies in what the market expects of four-year-olds rather than in maturation alone. A test reads age, prior_runs, days_since_last_run, month and is_handicap.
openpriority 0.50raised by O-6ea233f1, O-ed858a4b
New question16 Oct, 13:09
Do four-year-old returners whose first run back falls in May score at or below their prices, as a setback-selection reading of the early-spring concentration predicts?
O-ed858a4b puts March–April at 1.29 on 322, while the full March–May H126 base sits near 1.13. That implies May is weak, but May has not been measured as its own population. If May returners score below 1.0, the edge is an early-calendar effect and H126 should be restricted. If May scores near 1.13, the March–April figure is just noise. A test reads month, days_since_last_run, age, prior_runs, is_handicap and sp_decimal.
answeredpriority 0.55raised by O-ed858a4b
New question16 Oct, 13:09
Is the early-spring lift really confined to returners at 14/1 or shorter, given that 16/1+ went 0.79 on 114 in March–April but 1.19 on 245 in the earlier unfancied cut?
The longshot band contradicts itself across cuts in this thread, and the vault already showed that the 14/1 cap does not sharpen H126. This should be read by the descriptive scanner on sp_decimal and sp_rank by month before anyone proposes a rule, so that a near miss is not sliced until it passes.
To answer: Is the H126 lift caused by four-year-old physical maturation that the handicap mark hasn't caught up with, or just by lightly raced horses returning fresh at any age?
because If the cause is physical development over the winter that the mark has not caught up with, three-year-olds coming back from their juvenile season should show it as strongly as four-year-olds or more. If the cause is the market doubting that a four-year-old still has improvement in it, three-year-olds, whose improvement everyone expects, should be fully priced.
and the market misses it because Everyone expects three-year-olds to improve from two to three, so the market already prices that in. A lightly raced four-year-old is often written off as having had its chance at three, which is a doubt about further progress that the market does not shade for.
predicting negative, A/E 0.92 to 1.03 on 1000+ bets, clearly below the four-year-old base of 1.13
and it is wrong if three-year-old returners score A/E 1.10 or more on 1000+ bets, which would favour general young-horse maturation over a misjudgement specific to four-year-olds
race_type = 'Flat' AND is_handicap = 1 AND month IN (3, 4, 5) AND age = 3 AND prior_runs <= 10 AND days_since_last_run >= 150
To answer: Do four-year-old returners whose first run back falls in May score at or below their prices, as a setback-selection reading of the early-spring concentration predicts?
because A lightly raced four-year-old that reappears in March or April has been brought back on schedule after a winter of growing. One that first appears in May has more often had a setback. By May the market has also seen how that crop's four-year-olds are running, so the pre-season doubt has been priced away.
and the market misses it because In March and April, the form book shows only three-year-old runs that are five or more months old, so the market prices the horse on a stale profile. By May there are trends across the crop and stable form to price, and the late reappearance itself is a warning sign that bettors underweight less than the early unknown.
predicting negative, A/E 0.85 to 1.02 on 400+ bets
and it is wrong if May returners score A/E 1.10 or more on 400+ bets, meaning the March–April concentration was noise and the full March–May window should stand
race_type = 'Flat' AND is_handicap = 1 AND month = 5 AND age = 4 AND prior_runs <= 10 AND days_since_last_run >= 150
Two new results make the thread's shape clearer, but neither settles anything on its own. First, the early-spring cut (H-2d115a58: four-year-old returners in March and April) passed discovery as a niche at A/E 1.29 on 322 bets (p 0.068), against a same-race control of 0.95. That is above H126's March–May level of 1.13 on 1904. It fits the pattern of every earlier cut fading from period 1 to period 2: the lift seems to sit early in the season. Inside that cut, 16/1+ runners scored only 0.79 on 114 (2 wins), so the lift comes from the shorter-priced runners. GB scored 1.33 on 225 and Ireland 1.08 on 97. Second, the powered age control (H-2fa8d5cc: lightly raced older handicappers back from a break) went the predicted way, scoring 0.64 on 119 against a control of 1.04. But with p 0.908 and only 6 winners it cannot rule anything out, and it splits GB 0.42 on 77 against IRE 1.30 on 42. Read as a chain, the evidence says: the lift belongs to the first run back, not to spring four-year-olds in general (1.00 on 786 for the same profile without a break). It needs light racing (exposed returners scored 0.80 on 191). It is not tied to an unchanged mark (unraised marks scored 0.95 on 151). It now leans toward being specific to four-year-olds and to the early part of the season. All of these leans rest on underpowered controls. The vault showed that the 14/1 cap does not sharpen the edge (1.11 on 1567). So the honest base remains H126 at about 1.11 to 1.13. The open causal question is whether the season calendar (first weeks of turf, before the crop has been seen) or age-specific maturation is doing the work. The two tests below separate those causes rather than slicing H126 further.
investigator v1
Tested16 Oct, 13:15
perturbation: Loosen thresholds: 120+ days off, up to 12 prior runs: A/E 1.23 on 460 bets, p = 0.083 (95%: 0.92 to 1.61), by period 1.19. pass
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Tested16 Oct, 13:15
partition: March only: returners before the turf season and big April handicaps: A/E 1.77 on 92 bets, p = 0.031 (95%: 0.97 to 2.97), by period 1.77. pass
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Observed16 Oct, 13:15
H-2d115a58 (Early-spring four-year-old returners: the lift sits in March and April) survived falsification: 4 of 5 attacks left it standing; it is frozen for its one vault query.
result0.55reasoned about 16 Oct, 13:17from H-2d115a58
Tested16 Oct, 13:15
partition: GB class 4-6 handicaps only: leaves out the big April showcase handicaps: A/E 1.24 on 171 bets, p = 0.17 (95%: 0.79 to 1.87), by period 1.19. pass
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Tested16 Oct, 13:15
control: Control: same March-April four-year-old profile, recently raced (no winter break): A/E 0.94 on 424 bets, p = 0.68 (95%: 0.69 to 1.26), by period 0.92. fail
H-2d115a58 (Early-spring four-year-old returners: the lift sits in March and April) met the sealed vault supportive: A/E 1.02 on 1548 bets (p 0.3968, threshold 0.0029).
result0.55reasoned about 16 Oct, 13:17from H-2d115a58
Tested16 Oct, 13:15
vault query (supportive): A/E 1.02 on 1,548 bets, p = 0.40 (95%: 0.87 to 1.19), by period 1.14 · 0.90. pass
H-cbe10008 (May four-year-old returners: the late end of the window carries no edg) was predicted negative but came out the other way: A/E 1.06 on 197 bets (p 0.427 for the opposite direction).
surprise0.81reasoned about 16 Oct, 13:17from H-cbe10008
Tested16 Oct, 13:15
broad rule: A/E 1.06 on 197 bets, p = 0.43 (95%: 0.66 to 1.60), by period 1.06. fail
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Killed16 Oct, 13:15
May four-year-old returners: the late end of the window carries no edgeDiscovery → KILLEDfails: 197 bets, needs 300; A/E 1.057, needs 0.95 or lower; p 0.6551 (lower tail), needs 0.05; periods [1.06], each needs 0.99 or lower; A/E 1.057 against same-race control 0.993, needs 0.03 below
Observed16 Oct, 13:15
H-b2362b70 (Lightly raced three-year-old spring returners: is the mispricing only ) was predicted negative but came out the other way: A/E 1.03 on 640 bets (p 0.414 for the opposite direction).
surprise0.79reasoned about 16 Oct, 13:17from H-b2362b70
Tested16 Oct, 13:15
broad rule: A/E 1.03 on 640 bets, p = 0.41 (95%: 0.79 to 1.34), by period 1.04 · 1.02. fail
Does the won-last-time returner effect (1.27 on 371 at age four) also appear in lightly raced three-year-old spring returners, whose overall group scored only 1.03 on 640?
If a returner coming off a win is underrated whatever its age, the published system's mechanism is the market discounting a winner's mark rise after a long break, not four-year-old maturity. Columns: age, last_position, prior_runs, days_since_last_run, month, is_handicap, last_official_rating.
answeredpriority 0.55raised by O-4c05b7fc, O-73ee471f
New question16 Oct, 13:17
Is the broad H126 lift (1.13 on 1904 at discovery) decaying over time, given that the March–April sharpen fell to 1.02 on 1548 in the vault?
The vault shrinkage could be regression from a lucky discovery slice, or a real fade as the market learns. A season-by-season split of the base rule would tell us which and decide whether to keep investing in the thread. Columns: race_type, is_handicap, month, age, prior_runs, days_since_last_run, plus the race-date year.
openpriority 0.65raised by O-73ee471f, O-0751fe6e, O-18cd4b74
To answer: Does the won-last-time returner effect (1.27 on 371 at age four) also appear in lightly raced three-year-old spring returners, whose overall group scored only 1.03 on 640?
because A horse that won its last run gets a mark rise. After 150+ days off, punters assume the rise plus the rust will beat it. But a lightly raced young winner is often still improving over the winter, so the rise understates its current ability. That logic does not depend on being exactly four.
and the market misses it because The form book shows a win followed by a hike and a long absence, which reads as 'handicapped out and needs the run'. The market cannot see how much the horse has developed over the winter.
predicting positive, A/E 1.10 to 1.20 on 600+ bets, with the age-three part at 1.05 or more
and it is wrong if A/E below 1.03 on 600+ bets, or the age-three part below 1.00 on 250+ bets. That would tie the published system to age four only.
race_type = 'Flat' AND is_handicap = 1 AND month IN (3, 4, 5) AND age IN (3, 4) AND prior_runs <= 10 AND days_since_last_run >= 150 AND last_position = 1
The thread is weaker now. The early-spring sharpen (H-2d115a58) survived 4 of 5 attacks, then met the sealed vault at A/E 1.02 on 1548 bets (p 0.40). Its discovery figure was 1.29 on 322. So the March–April concentration mostly shrank back to the market price once a larger sample was measured. The May-only cut (H-cbe10008) was predicted weak but came out at 1.06 on 197. Its interval (0.66 to 1.60) is wide, and GB was 0.95 on 159 while IRE was 2.16 on 38. So May is not shown to be weak, and the setback-selection story has no support. Taken together: March–April 1.02 in the vault, May 1.06, and a full-window discovery of 1.13 on 1904. That reads as a modest, possibly decaying broad effect, not an early-spring one. Lightly raced three-year-old spring returners scored 1.03 on 640 (16/1+ at 0.81 on 242). That is close to the market price and weaker than the four-year-old base. Earlier, older horses scored 0.64 on 119 and exposed four-year-olds 0.80 on 191. So the four-year-old, lightly raced combination remains the best group, but its size is now uncertain. The piece with the strongest evidence is the won-last-time subset, 1.27 on 371 (published). Splitting by month or price band has not paid. The next useful tests are whether the base holds over time and whether the won-last-time mechanism holds outside age four. Ireland stays unreadable: 2.16 on 38 here.
investigator v1
Answered16 Oct, 13:17
Is the four-year-old returner lift concentrated in March and April, fading by May when late returners are more likely to be horses that had setbacks and the market has seen how the crop came out of the winter?
No concentration is shown: March–April scored only 1.02 on 1548 in the sealed vault, against 1.29 on 322 at discovery, and May scored 1.06 on 197. The early-spring lift did not replicate.
Do four-year-old returners whose first run back falls in May score at or below their prices, as a setback-selection reading of the early-spring concentration predicts?
No: May four-year-old returners scored 1.06 on 197 (CI 0.66–1.60; GB 0.95 on 159), so May is not shown to be below its prices, and the setback-selection reading is unsupported.
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Tested16 Oct, 13:22
refinement: age three only (the generalisation the question asks about): A/E 1.05 on 110 bets, p = 0.47 (95%: 0.57 to 1.76), by period 1.27 · 0.52. info
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Observed16 Oct, 13:22
H-5998bc20 (Lightly raced spring handicap returners that won last time, at age thr) passed discovery as a niche: A/E 1.27 on 185 bets (p 0.11), same-race control 0.95.
result0.50reasoned about 16 Oct, 13:28from H-5998bc20
Tested16 Oct, 13:22
refinement: age four only (overlap with the published system, to show how much of the pooled figure it carries): A/E 1.51 on 75 bets, p = 0.059 (95%: 0.90 to 2.39), by period 1.71. info
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Tested16 Oct, 13:22
control: Two comparisons. (1) The other runners in the same races should be at or slightly below A/E 1.00 (around 0.95-1: A/E 1.07 on 971 bets, p = 0.28 (95%: 0.86 to 1.31), by period 1.07 · 1.07. info
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Observed16 Oct, 13:23
H-5998bc20 (Lightly raced spring handicap returners that won last time, at age thr) passed discovery but was falsified: driven by one price band (up to 3/1): the rest score 0.90.
near miss0.50reasoned about 16 Oct, 13:28from H-5998bc20
Tested16 Oct, 13:23
price_band: Remove the 3/1-or-shorter band, which has 17 of the 32 wins: A/E 0.96 on 151 bets, p = 0.59 (95%: 0.51 to 1.64), by period 1.27 · 0.41. fail
Is the returner lift specific to horses whose last win came in a handicap (a real mark rise), rather than to maiden or novice winners making a handicap debut?
The last_is_handicap = 1 partition scored 1.50 on 84, against 1.27 for the whole won-last-time group. This is the case the mark-rise mechanism names, but 84 bets cannot be read alone. A test would pool ages 3 to 5 and read last_is_handicap, last_position, last_official_rating, days_since_last_run and prior_runs.
openpriority 0.40raised by O-46d14312
New question16 Oct, 13:28
Do short-priced (3/1 or shorter) Flat handicap returners after 150+ days in spring beat their prices at any age and experience level, which would make the H126 short-band lift a general returner-confidence effect?
Three separate rules in this thread were each carried by the up-to-3/1 band (H-5998bc20 1.97 on 28; H-1a3e129f; sp_rank<=3 at 1.175). If all ages and all experience levels show the same lift, then age four and light racing are not the mechanism. Columns: sp_decimal, sp_rank, days_since_last_run, age, prior_runs, is_handicap, month, race_type.
answeredpriority 0.60raised by O-c2bb350a, O-46d14312
New question16 Oct, 13:28
Is the 1.07 lift on 971 lightly raced three- and four-year-old non-winner returners a stable feature, or a property of the 2026 spring alone?
The non-winner control matched the base H126 level and was flat across both periods (1.07, 1.07), with the best band at up to 14/1 (1.33 on 290). Only one spring is in discovery, so the vault and the forward period must decide. Columns: last_position, age, prior_runs, days_since_last_run, sp_decimal.
To answer: Do short-priced (3/1 or shorter) Flat handicap returners after 150+ days in spring beat their prices at any age and experience level, which would make the H126 short-band lift a general returner-confidence effect?
because When a horse returns from a long break, the trainer decides whether it runs and the money shows whether the yard thinks it is fit. A short price on a returner after 150+ days reflects stable confidence, but the public and the compilers still shade it for rust. That discount would apply at any age.
and the market misses it because The form book shows a long absence. Odds compilers apply a standard 'needs the run' discount even when market support says the yard has the horse ready, so fancied returners go off slightly too long.
predicting positive, A/E 1.05 to 1.15 on 300+ bets across all ages
and it is wrong if A/E below 1.02 on 300+ bets, or ages 5 and over below 0.97 while ages 3 and 4 carry the lift. Either would point the short-band effect back to young, lightly raced horses.
race_type = 'Flat' AND is_handicap = 1 AND month IN (3, 4, 5) AND days_since_last_run >= 150 AND sp_decimal <= 4
Pooling three- and four-year-olds coming off a win did not widen the published system. H-5998bc20 scored 1.27 on 185 bets (p 0.11, same-race control 0.95). The falsification step then killed it: the up-to-3/1 band produced 1.97 on 28 bets with 17 of the 32 wins, and without that band the rule scored 0.96 on 151. Most of the pooled figure came from the age-four part, 1.51 on 75. The age-three part was 1.05 on 110, and its second period fell to 0.52. Two cuts were more interesting. Horses whose last win came in a handicap (the mark-rise case the mechanism describes) scored 1.50 on 84. The non-winner control scored 1.07 on 971 (GB 1.09), so the 'won last time' feature adds less than the raw gap suggests. Reading the whole chain, the same pattern keeps returning. The non-handicap version (H-1a3e129f), the sp_rank 3-or-better rule (1.175 on 251) and now this rule all depended on the shortest price band. Every 16/1+ or long-tail cut sits near or below 1.0, and the March-April sharpen fell from 1.29 to 1.02 in the vault. What survives is a mild returner lift (about 1.07 to 1.13) that concentrates in horses the market already makes fancied. The specific stories (age four, early spring, a lagging mark, a winner's mark rise) remain unseparated from that. The discovery window covers only one spring (2026), so every split in this thread comes from a single crop of horses and a single set of conditions.
investigator v1
Answered16 Oct, 13:28
Does the won-last-time returner effect (1.27 on 371 at age four) also appear in lightly raced three-year-old spring returners, whose overall group scored only 1.03 on 640?
Mostly no. Lightly raced three-year-old spring returners that won last time scored 1.05 on 110 (periods 1.27 then 0.52, RED), so the pooled 1.27 rested on the age-four part (1.51 on 75) and the up-to-3/1 band.
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Tested16 Oct, 13:28
refinement: Ages other than four (the direct rival test: lift without H126's four-year-olds): A/E 1.08 on 70 bets, p = 0.39 (95%: 0.68 to 1.63), by period 1.14. info
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Observed16 Oct, 13:28
H-71a94a07 (Rival: short-priced spring handicap returners pay at any age, so H126') passed discovery as a niche: A/E 1.1 on 141 bets (p 0.271), same-race control 0.95.
result0.50reasoned about 16 Oct, 13:30from H-71a94a07
Observed16 Oct, 13:28
H-71a94a07 (Rival: short-priced spring handicap returners pay at any age, so H126'): the period '2025-10-01 to 2026-04-27' carries the effect (A/E 1.18) while the rest score after 2026-04-27 0.97.
concentration0.55reasoned about 16 Oct, 13:30from H-71a94a07
Tested16 Oct, 13:28
refinement: Exposed returners (more than 10 prior runs): is the lift free of light racing?: A/E 1.15 on 25 bets, p = 0.40 (95%: 0.49 to 2.26). info
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Tested16 Oct, 13:28
control: Same-race other runners (lab default) should score at or below 1.0, since they absorb the overround and the lon: A/E 0.99 on 1,239 bets, p = 0.59 (95%: 0.89 to 1.09), by period 0.98 · 1.00. info
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
New question16 Oct, 13:30
Does the spring returner lift need a handicap (a mark that lags the horse's progress), or does a rust discount alone produce it in non-handicaps too?
Dropping is_handicap from H-71a94a07 cut A/E from 1.10 on 141 to 1.02 on 421. That suggests non-handicap fancied returners are priced about right. If confirmed, the mechanism is about the handicap mark, not market rust, and the thread should look at last_official_rating and days_since_win rather than price. A test reads is_handicap, is_maiden, is_novice, days_since_last_run and sp_decimal.
testingpriority 0.50raised by O-21cb941f
New question16 Oct, 13:30
Within spring Flat handicap returners after 150+ days, does the lift sit in the second tier of the market (roughly 3/1 to 7/1) rather than at 3/1 or shorter, at any age?
The sp_decimal <= 5 neighbour of H-71a94a07 scored 1.214 on 273 with periods 1.22 and 1.20. Its 4/1-to-7/1 band scored 1.42 on 132. The up-to-3/1 rule scored 1.10, and the evens-or-shorter band 0.71 on 11. H-1a6f83c9 (14/1 or shorter) also scored 1.21. If the lift lives in the mid-price band, the thread's 'short band carries it' reading was mislocated. A test reads sp_decimal, sp_rank, days_since_last_run, is_handicap, month and age.
testingpriority 0.55raised by O-21cb941f, O-7ebaf3d7
New question16 Oct, 13:30
Is the fade after 27 April in the fancied-returner rule (1.18 to 0.97) a May effect that repeats each spring, or just noise in one season?
The second period is almost all May 2026. That lines up with the March-April sharpen (1.29) against May (1.06), but the 4/1-cap neighbour held 1.20 after 27 April. A month-by-month read of month, sp_decimal and days_since_last_run would show whether timing or price band drives the split.
To answer: Does the spring returner lift need a handicap (a mark that lags the horse's progress), or does a rust discount alone produce it in non-handicaps too?
because In a handicap, a horse that has strengthened over the winter runs off a mark set on its old form, so it is better than its weight says. In a non-handicap there is no mark to lag, and the only possible mispricing is rust. If the lift is about the mark, non-handicap fancied returners should score about 1.0.
and the market misses it because The handicap mark is visible but is read as current ability. In a non-handicap, the market prices the horse on its form and the yard's signal, with no stale weight to mislead it.
predicting negative, A/E 0.90 to 1.02 on 300+ bets, well below the handicap version
and it is wrong if A/E of 1.08 or higher on 300+ bets, which would make rust, not the mark, the driver
race_type = 'Flat' AND is_handicap = 0 AND month IN (3, 4, 5) AND days_since_last_run >= 150 AND sp_decimal <= 8
To answer: Within spring Flat handicap returners after 150+ days, does the lift sit in the second tier of the market (roughly 3/1 to 7/1) rather than at 3/1 or shorter, at any age?
because Some money for a returner shows the yard thinks it is ready. When a returner is clear favourite at short odds, that confidence is plain and heavily scrutinised, so the rust discount is mostly removed. A returner at 4/1 to 7/1 has partial support, so the standard 'needs the run' discount, plus a mark that has not caught up with winter progress, still holds the price up. This is a lead from one neighbour (the 4/1 cap) and is stated as such before testing the band on its own.
and the market misses it because The form book shows only a long absence and an old handicap mark. Compilers price the absence as a uniform penalty. The odds shorten fully only for the obvious stable favourite, so moderately backed returners stay slightly too big.
predicting positive, A/E 1.10 to 1.25 on 250+ bets, with ages other than four at or above about 1.05
and it is wrong if A/E below 1.03 on 250+ bets; or ages other than four below 1.0 while four-year-olds carry it; or either period below 1.0
race_type = 'Flat' AND is_handicap = 1 AND month IN (3, 4, 5) AND days_since_last_run >= 150 AND sp_decimal > 4 AND sp_decimal <= 8
The thread keeps finding a mild spring-returner lift, and every attempt to pin it to one cause (age four, light racing, early season, the up-to-3/1 band) has blurred on contact. H-71a94a07 tested the rival idea that fancied returners pay at any age. It passed only as a niche: A/E 1.10 on 141 bets (p 0.27), against 0.95 for same-race runners and 0.99 on 1,239 for short-priced spring handicappers that are not returning. The discovery window holds a single spring, and the split looks like a fade: 1.18 on 91 up to 27 April, then 0.97 on 50 after it. But the later period is mostly May, so this repeats the May-versus-March/April question more than it shows decay. The refinements are too thin to separate age: ages other than four scored 1.08 on 70, ages 5+ 0.98 on 22, and exposed horses 1.15 on 25. Ireland is unreadable again (0.81 on 9). Two neighbour results matter more than the headline. First, widening the cap to 4/1 gave 1.214 on 273 (p 0.048, both periods about 1.2), and the new 4/1-to-7/1 band scored 1.42 on 132. The evens-or-shorter band was 0.71 on 11, so the lift is not in the shortest prices. Second, dropping the handicap condition fell to 1.02 on 421, so non-handicap returners dilute the effect. Taken together, the lift looks like a property of moderately fancied handicap returners, not the most strongly backed ones. That fits a handicap-mark or rust discount better than an 'obvious stable confidence' story. The 4/1-to-7/1 band is still a lead from one neighbour on one spring, not a finding.
investigator v1
Tested16 Oct, 13:34
partition: Rival test: ages other than four: A/E 1.08 on 70 bets, p = 0.39 (95%: 0.68 to 1.63), by period 1.14. pass
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Observed16 Oct, 13:34
H-71a94a07 (Rival: short-priced spring handicap returners pay at any age, so H126') passed discovery but was falsified: driven by one price band (up to 3/1): the rest score 0.71; IRE: too few bets to judge.
near miss0.50reasoned about 16 Oct, 13:38from H-71a94a07
Tested16 Oct, 13:34
perturbation: Tighter thresholds: 180+ days off, up to 5/2: A/E 1.13 on 76 bets, p = 0.28 (95%: 0.76 to 1.62), by period 1.10. pass
refinement: ages other than four (tests 'at any age', separates from the H126 four-year-old thread): A/E 1.14 on 315 bets, p = 0.20 (95%: 0.84 to 1.50), by period 1.17 · 1.06. info
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Tested16 Oct, 13:34
refinement: GB only (Irish sample tiny and Irish splits have repeatedly killed this thread): A/E 1.12 on 393 bets, p = 0.20 (95%: 0.86 to 1.43), by period 1.17 · 1.04. info
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Observed16 Oct, 13:34
H-12ef9cb4 (Second-tier fancied spring handicap returners (3/1 to 7/1) beat their ) passed discovery as a niche: A/E 1.08 on 473 bets (p 0.263), same-race control 0.99.
result0.50reasoned about 16 Oct, 13:38from H-12ef9cb4
Observed16 Oct, 13:34
H-12ef9cb4 (Second-tier fancied spring handicap returners (3/1 to 7/1) beat their ): the country 'GB' carries the effect (A/E 1.12) while the rest score IRE 0.88.
concentration0.55reasoned about 16 Oct, 13:38from H-12ef9cb4
Tested16 Oct, 13:34
broad rule: A/E 1.00 on 536 bets, p = 0.52 (95%: 0.84 to 1.17), by period 0.99 · 1.01. fail
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Tested16 Oct, 13:34
control: Same-race other runners should sit near 0.95-1.00 (normal overround drag); if they also score above 1.05 the li: A/E 1.10 on 141 bets, p = 0.27 (95%: 0.81 to 1.46), by period 1.18 · 0.97. info
partition: Ages other than four (separate from H126 four-year-old thread): A/E 1.14 on 315 bets, p = 0.20 (95%: 0.84 to 1.50), by period 1.17 · 1.06. pass
Is the repeated Irish null for spring returners structural (big yards' returners in Ireland priced sharply) or just small samples of 50 to 80 bets?
Ireland scored 0.88 on 80 here and 0.887 on 50 in the sp_rank rule, while GB carried 1.12. A descriptive scan of returners by region and trainer volume (region, trainer, sp_decimal, days_since_last_run, age) would show whether the Irish shortfall sits in a few big yards or is spread evenly.
openpriority 0.30raised by O-48b34371, O-f8eac365
New question16 Oct, 13:38
Within the same 3/1–7/1 price band, do spring handicap returners that are not four-year-olds score at or below their prices, so that age four carries the all-age 1.08?
H-12ef9cb4 mixes ages. If age-four maturation is the cause, the other ages in the same band should score about 1.0. If they score 1.08 or more, the age story collapses into a mild general returner lift. Columns: age, sp_decimal, days_since_last_run, is_handicap, month, region.
answeredpriority 0.50raised by O-3011d3a2, O-48b34371, O-bf6f33c9
To answer: Within the same 3/1–7/1 price band, do spring handicap returners that are not four-year-olds score at or below their prices, so that age four carries the all-age 1.08?
because Four-year-olds fill out physically over their first winter as older horses, but their mark still reflects three-year-old form. Three-year-olds are still progressing on a mark that keeps moving, and older horses are fully developed, so neither group should have a hidden gain after a break.
and the market misses it because The form book shows a long absence and a mark set the previous season. Punters price a returner by its fitness and connections, not by how much its age class matures over a winter, so a hidden gain should exist only at age four.
predicting negative, A/E 0.92 to 1.02 on 300+ bets
and it is wrong if A/E of 1.08 or higher on 300+ bets, which would mean the second-tier returner lift does not need a four-year-old
race_type = 'Flat' AND is_handicap = 1 AND month IN (3, 4, 5) AND age <> 4 AND days_since_last_run >= 150 AND sp_decimal >= 4 AND sp_decimal <= 8
The chain has now tested whether the H126 lift belongs to four-year-olds or to any fancied spring handicap returner, and the any-age version does not hold up. H-71a94a07 (short-priced returners, any age) depended entirely on the up-to-3/1 band: take that band away and the rule scores 0.71. H-12ef9cb4 (the 3/1 to 7/1 band, any age) scored A/E 1.08 on 473 bets with p 0.26, against a same-race control of 0.99. Nearly all of that came from GB (1.12 on 393); Ireland scored 0.88 on 80, and the rule was killed there. The rest of the thread fits the same picture. Three-year-olds scored 1.035 on 640. Age 5+ scored 0.64, but on only 119 bets. The base four-year-old rule scored 1.13 on 1904, and its two shadow sharpens (March–April 1.29 on 322; 14/1 or shorter 1.21 on 337) are still alive. So the any-age "returner confidence" story is no better than a mild, fragile GB lift that rests on one price band. Age four is still the most specific carrier, but it has not been cleanly separated from a general mild returner lift. Every slice of this thread has died on Ireland or on one price band, and the shortened-price rules also fade after late April 2026. Slicing further looks unlikely to pay. What the thread needs is the forward test of the two shadows and one clean age contrast inside a fixed price band.
investigator v1
Answered16 Oct, 13:38
Do short-priced (3/1 or shorter) Flat handicap returners after 150+ days in spring beat their prices at any age and experience level, which would make the H126 short-band lift a general returner-confidence effect?
No. Short-priced returners at any age were carried only by the up-to-3/1 band (the rest scored 0.71), and the 3/1–7/1 any-age band scored 1.08 on 473 (p 0.26, IRE 0.88). That is not a general returner-confidence effect.
H-1369adb4 (Non-four-year-old second-tier spring returners: age control inside the) was predicted negative but came out the other way: A/E 1.17 on 333 bets (p 0.141 for the opposite direction).
surprise0.90reasoned about 16 Oct, 13:41from H-1369adb4
Tested16 Oct, 13:40
broad rule: A/E 1.17 on 333 bets, p = 0.14 (95%: 0.88 to 1.51), by period 1.24 · 1.01. fail
Is the thread's lift a calendar effect of one early spring: returners of any age and any price beating their SP in March to 27 April 2026, then fading?
Every rule shows the same pre/post 27 April split (here 1.24 vs 1.01, H-71a94a07 1.18 vs 0.97), and the age control lifted as much as the four-year-olds. If all March-April handicap returners score about 1.10, the age, price-band and won-last-time slices are just cuts of one seasonal effect. Columns: month, days_since_last_run, age, sp_decimal, region, is_handicap, last_surface (all-weather prep vs turf).
answeredpriority 0.65raised by O-54e07df5, O-7ebaf3d7
New question16 Oct, 13:41
Within the non-four-year-old second-tier lift, is it the three-year-olds or the exposed older horses (prior_runs > 10) that beat their price?
The rule had no prior_runs cap, while lightly raced age 5+ returners scored 0.64 on 119. If exposed older horses carry the lift, the idea of lightly raced horses improving behind their mark is irrelevant, and the effect looks more like fitness or the stable's intent after a break. Columns: age, prior_runs, sp_decimal, days_since_last_run.
To answer: Is the thread's lift a calendar effect of one early spring: returners of any age and any price beating their SP in March to 27 April 2026, then fading?
because Early in the turf season, a horse that comes back after a winter break was usually brought back on purpose by its trainer, fit and pointed at a race. Its form figures, though, are months old, often from different ground or a different surface. That holds at every age. Once late April arrives, the horses have had a run and current form takes over, so the effect fades.
and the market misses it because The form book shows a long absence and old figures, and punters mark returners down for rust. In March and April almost everything has fresh-looking form from the all-weather except these horses, so the gap makes them look worse than their readiness justifies. The visible layoff is overweighted.
predicting positive, A/E 1.05 to 1.12 on 1500+ bets, with March-April clearly above May
and it is wrong if A/E below 1.02 on 1500+ bets, or no gap between non-four-year-olds and four-year-olds such that four-year-olds score 1.10+ while the rest sit at 1.0 or below, which would restore the age story
race_type = 'Flat' AND is_handicap = 1 AND month IN (3, 4) AND days_since_last_run >= 150
The age control came out the wrong way for the age story. Spring Flat handicap returners (150+ days) at 3/1 to 7/1 that were NOT four-year-olds were predicted to score about 1.0. They scored A/E 1.17 on 333 bets (56 wins against 48 expected, CI 0.88 to 1.51, p 0.14 for a positive lift), with a same-race control of 0.967. GB was 1.20 on 275 and IRE 1.01 on 58. This rule had no prior_runs cap, so it mixes exposed horses with lightly raced ones.
The any-age second-tier rule scored 1.08 on 473. Taking 333 of those bets out at 1.17 leaves the four-year-olds in that band near or below 1.0. So inside the second tier, age four does not carry the lift. If anything it is the weaker part.
The chain now looks like this:
- the four-year-old lift in H126 and its March to April sharpen (1.286 on 322) sits in the 14/1 band and at up to 3/1;
- the 3/1 to 7/1 lift is shared across ages;
- every rule shows the same split around 27 April: here 1.24 before and 1.01 after, and earlier 1.18 against 0.97.
Every rule in this thread draws on a single spring (2026). The simplest reading is that early spring returners in general beat their prices in that one spring, and the age, price-band and won-last-time slices are cuts of one calendar effect. Maturation that the handicap mark has not caught up with is still not established. Nothing here is significant on its own.
investigator v1
Answered16 Oct, 13:41
Within the same 3/1–7/1 price band, do spring handicap returners that are not four-year-olds score at or below their prices, so that age four carries the all-age 1.08?
No. Within the 3/1 to 7/1 band, non-four-year-old spring returners scored 1.17 on 333 against a 0.967 control, so the second-tier lift does not need a four-year-old. The four-year-olds in that band are, if anything, the weaker part.
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Tested16 Oct, 13:46
refinement: 14/1 or shorter: edges in this thread die in the 16/1+ tail: A/E 1.14 on 861 bets, p = 0.093 (95%: 0.94 to 1.36), by period 1.12 · 1.32. info
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Observed16 Oct, 13:46
H-5167b372 (All early-spring Flat handicap returners: a calendar effect rival to t): the country 'GB' carries the effect (A/E 1.11) while the rest score IRE 1.0.
concentration0.55reasoned about 16 Oct, 13:53from H-5167b372
Observed16 Oct, 13:46
H-5167b372 (All early-spring Flat handicap returners: a calendar effect rival to t) passed discovery as a niche: A/E 1.09 on 1424 bets (p 0.173), same-race control 0.98.
result0.50reasoned about 16 Oct, 13:53from H-5167b372
Tested16 Oct, 13:46
refinement: Turf only: the mechanism is about the start of the turf season against AW-fresh rivals: A/E 1.03 on 944 bets, p = 0.43 (95%: 0.81 to 1.28), by period 1.01 · 1.21. info
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Tested16 Oct, 13:46
control: Same-race other runners should score about 1.0 (0.97 to 1.0). If they also beat their prices, the lift belongs : A/E 0.99 on 792 bets, p = 0.56 (95%: 0.75 to 1.27), by period 0.99. info
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Observed16 Oct, 13:51
H-5167b372 (All early-spring Flat handicap returners: a calendar effect rival to t) survived falsification: 5 of 6 attacks left it standing; it is frozen for its one vault query.
result0.55reasoned about 16 Oct, 13:53from H-5167b372
Tested16 Oct, 13:51
partition: Turf only: start of the turf season: A/E 1.03 on 944 bets, p = 0.43 (95%: 0.81 to 1.28), by period 1.01 · 1.21. pass
Are long-priced (16/1+) early-spring Flat handicap returners over-bet beyond the general longshot shortfall, while 16/1+ returners later in the year are not?
The 16/1+ band of the calendar rule scored 0.742 on 563 (0.612 on 337 at 180+ days). All March-April 16/1+ handicap runners scored 0.91 on 2211, and year-round 16/1+ returners scored 1.118 on 1821. If the shortfall is specific to early spring, it is a negative-direction edge (lay or avoid). It also explains why every positive rule in this thread needed a price cap. Columns: sp_decimal, days_since_last_run, month, race_type, is_handicap, field_size, region.
answeredpriority 0.50raised by O-972b7090, O-c14824ce
New question16 Oct, 13:53
Does the March-April returner lift sit in all-weather races rather than turf, contrary to the turf-season mechanism?
Turf-only scored 1.027 on 944 against 1.09 pooled on 1424, so roughly 480 non-turf returners must score about 1.2. The mechanism predicted turf. An opposite result is a discovery: returners from the turf season may be under-rated on the all-weather against rivals that have run all winter. Columns: surface, last_surface, days_since_last_run, month, sp_decimal.
answeredpriority 0.30raised by O-c14824ce, O-352505b3
New question16 Oct, 13:53
Is the late-April window (28-30 April, 1.41 on 101) a real step before May (0.986 on 792), or noise from a few wins?
The thread's fade story has moved twice: first to 27 April, now to the month boundary. Before anyone builds rules on dates, the lab should know whether returner A/E changes smoothly by day of season or only in one window of a single spring. Columns: month, days_since_last_run, race date, course.
To answer: Are long-priced (16/1+) early-spring Flat handicap returners over-bet beyond the general longshot shortfall, while 16/1+ returners later in the year are not?
because Early in the season, trainers use a first run back for two purposes. Returners they have primed are fancied and go off short. Returners at long prices are mostly getting a pipe-opener before their real targets, short of peak fitness and not aimed to win. The same trainer-intent story that would lift the short end of the returner pool predicts a shortfall at the long end.
and the market misses it because Big-field spring handicaps draw heavy each-way and novelty money. A fresh horse at 20/1 with a decent old mark looks like value, because the form book shows only the mark and the layoff, not whether the trainer intends this run as preparation. Long-price SPs are therefore shorter than these horses' real chances.
predicting negative, A/E 0.72 to 0.88 on 500+ bets
and it is wrong if A/E of 0.95 or higher on 500+ bets, or a same-race control of 16/1+ non-returners scoring as low (within 0.05), which would make it plain favourite-longshot bias
race_type = 'Flat' AND is_handicap = 1 AND month IN (3, 4) AND days_since_last_run >= 150 AND sp_decimal > 15
To answer: Does the March-April returner lift sit in all-weather races rather than turf, contrary to the turf-season mechanism?
because In March and early April, all-weather handicaps are full of horses that have raced all winter: exposed, with marks that are well known and often stuck. A horse returning from a summer and autumn turf campaign meets them fresh, with a turf mark that may flatter or understate it on the surface. Trainers who bring such a horse to the all-weather in spring usually do so because it is ready and the race suits.
and the market misses it because On the all-weather in spring, punters lean on recent all-weather figures. A returner has none, only old turf form months back, so the market discounts it as rusty and unproven on the surface.
predicting positive, A/E 1.08 to 1.20 on 400+ bets
and it is wrong if A/E below 1.03 on 400+ bets, or not above the turf-only partition (1.027)
race_type = 'Flat' AND is_handicap = 1 AND month IN (3, 4) AND days_since_last_run >= 150 AND surface = 'AW'
The calendar rival (H-5167b372) was the right test, and it has mostly emptied the thread. All March-April Flat handicap returners after 150+ days scored 1.09 on 1424 at discovery (p 0.17, control 0.98). In the sealed vault they scored 1.03 on 6934 (p 0.235). The four-year-old March-April sharpen fell to 1.02 on 1548 in the vault. Out of sample, then, a lightly raced four-year-old scores no better than any spring returner, and spring returners as a group are only slightly better than their price. Most of the four-year-old story (1.286 on 322, one spring) looks like a lucky season plus a small general returner effect. The discovery detail points the same way. Non-four-year-olds scored 1.048 on 965 (1.075 on 576 at 14/1 or shorter). The calendar mechanism's own test went against it: it predicted turf, but turf-only scored 1.027 on 944, which leaves the roughly 480 non-turf bets carrying the pooled lift. The predicted fade around 27 April also failed. The 28-30 April bets inside March-April scored 1.41 on 101, and May scored 0.986 on 792. Any fade is at the month boundary and rests on very few bets. The one strong, repeated pattern is the long-price tail. March-April returners at 16/1+ scored 0.742 on 563 (10 wins), and 0.612 on 337 at 180+ days. All March-April 16/1+ Flat handicap runners scored 0.91 on 2211, and 16/1+ returners across the whole year scored 1.118 on 1821. That is a lead, not a finding. Only the published won-last-time system and the H-2d115a58 shadow remain live, and their forward records should decide them, not more slicing.
investigator v1
Answered16 Oct, 13:53
Is the H126 lift caused by four-year-old physical maturation that the handicap mark hasn't caught up with, or just by lightly raced horses returning fresh at any age?
Not maturation. In the vault the four-year-old March-April rule scored 1.02 on 1548 and all spring returners 1.03 on 6934. Non-four-year-olds scored 1.048 on 965 at discovery. Age four adds nothing measurable once tested out of sample.
Is the thread's lift a calendar effect of one early spring: returners of any age and any price beating their SP in March to 27 April 2026, then fading?
Only weakly. All March-April handicap returners scored 1.09 on 1424 at discovery (mostly spring 2026) but 1.03 on 6934 in the vault (p 0.235). Any repeatable calendar lift is about 3% at most, and spring 2026 looks unusual. Inside March-April, the 28-30 April bets scored 1.41 on 101, so the fade does not start on 27 April.
H-5ad83d6a (Long-priced early-spring handicap returners are over-bet: the 16/1+ ta) pointed the predicted way but failed the bar: A/E 0.74 on 563 bets (p 0.863), control 1.01.
near miss0.70reasoned about 16 Oct, 14:03from H-5ad83d6a
Tested16 Oct, 13:58
broad rule: A/E 0.74 on 563 bets, p = 0.86 (95%: 0.36 to 1.36), by period 0.64. fail
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Tested16 Oct, 13:58
refinement: GB only (Irish AW is Dundalk alone and the thread is weak in Ireland): A/E 1.23 on 440 bets, p = 0.076 (95%: 0.92 to 1.61), by period 1.19 · 1.83. info
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Observed16 Oct, 13:58
H-ce91b64d (All-weather spring returners carry the calendar lift, not turf returne) passed discovery as a niche: A/E 1.2 on 480 bets (p 0.11), same-race control 0.96.
result0.50reasoned about 16 Oct, 14:03from H-ce91b64d
Observed16 Oct, 13:58
H-ce91b64d (All-weather spring returners carry the calendar lift, not turf returne): the country 'GB' carries the effect (A/E 1.23) while the rest score IRE 0.45.
concentration0.55reasoned about 16 Oct, 14:03from H-ce91b64d
Tested16 Oct, 13:58
control: Same-race other runners (mostly horses that raced through the winter) should score at or below 1.0. If they als: A/E 1.03 on 944 bets, p = 0.43 (95%: 0.81 to 1.28), by period 1.01 · 1.21. info
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
New question16 Oct, 14:03
Do Irish all-weather (Dundalk) returners really fail, or is 0.45 on 40 just noise in a thread that is GB-led everywhere?
Ireland weakness has ended several hypotheses in this thread. Dundalk returners year-round (128 bets, 0.82) give the only larger sample. Columns: region, course, surface, days_since_last_run.
openpriority 0.30raised by O-389ee4f8
New question16 Oct, 14:03
Is it the turf-to-AW switch that is mispriced at any layoff, or does it need a 150+ day absence?
This separates two causes. One is a market that discounts old turf form on the AW. The other is a market that discounts rust, and the AW is just where ready horses return. A short-gap switch control decides it. Columns: surface, last_surface, days_since_last_run, prior_surface_runs.
answeredpriority 0.55raised by O-0ab1a07a
New question16 Oct, 14:03
Is the returner lift a year-round all-weather effect, strongest when a horse switches from a turf campaign to the AW after 150+ days, rather than anything seasonal?
Dropping month from H-ce91b64d gave 1.138 on 1213 (periods 1.12 and 1.20). The last_surface='Turf' refinement scored 1.32 on 283, and removing the layoff gave 1.00 on 3925. If this is true, the thread should move out of the seasonal family. Columns: surface, last_surface, days_since_last_run, month, prior_surface_runs.
answeredpriority 0.70raised by O-0ab1a07a, O-389ee4f8
To answer: Is the returner lift a year-round all-weather effect, strongest when a horse switches from a turf campaign to the AW after 150+ days, rather than anything seasonal?
because A horse that last ran on turf and returns to an AW handicap after 150+ days does so by its trainer's choice. AW programmes run all year, so the trainer can pick the day the horse is ready and the race suits. It meets exposed AW regulars on a mark set on turf.
and the market misses it because Punters price AW handicaps on recent AW figures. A turf horse back from a long break has none, only old turf form, so the market discounts it twice: once for rust and once for the surface.
predicting positive, A/E 1.10 to 1.25 on 600+ bets
and it is wrong if A/E below 1.03 on 600+ bets, or the months outside March-April scoring below 1.0, which would put the lift back in the calendar
race_type = 'Flat' AND is_handicap = 1 AND surface = 'AW' AND last_surface = 'Turf' AND days_since_last_run >= 150
To answer: Is it the turf-to-AW switch that is mispriced at any layoff, or does it need a 150+ day absence?
because If the market only discounts turf form on the AW, a horse switching surface after a normal 2-8 week gap should be underpriced too. If the cause is trainer-chosen readiness after a break, short-gap switchers should be fairly priced.
and the market misses it because Under the rival, punters distrust turf form on the AW whatever the gap. Under the main idea, the switch alone is well understood and priced.
predicting negative, A/E 0.93 to 1.03 on 1500+ bets
and it is wrong if A/E of 1.08 or higher on 1500+ bets, which would make the surface switch, not the layoff, the driver
race_type = 'Flat' AND is_handicap = 1 AND surface = 'AW' AND last_surface = 'Turf' AND days_since_last_run BETWEEN 14 AND 60
The thread began as a four-year-old story (H126, 1.13 on 1904) and has since shrunk. Age four adds nothing out of sample. All March-April handicap returners met the vault at only 1.03 on 6934.
This round split that pooled lift by surface:
- **All-weather returners (H-ce91b64d):** 1.195 on 480 (p 0.11), same-race control 0.96. Promising niche.
- **Turf returners (the predeclared partition):** 1.027 on 944.
- **AW returners whose last run was on turf:** 1.32 on 283 (p 0.058).
The most telling neighbour drops the month condition entirely. AW handicap returners after 150+ days in any month scored 1.138 on 1213, with periods 1.12 and 1.20. So the lift looks attached to returning on the all-weather, not to spring. The "seasonal" framing may be the wrong frame. Removing the 150-day condition kills it (1.00 on 3925), so the layoff matters.
Two cautions:
- **Ireland:** Dundalk returners scored 0.45 on 40. Too few to judge, but it repeats the thread's GB-only pattern.
- **Same data:** the AW figure was inferred from the pooled rule before it was proposed, so this discovery is not independent.
The long-price over-bet idea (H-5ad83d6a, 0.742 on 563, 10 wins against 13.5 expected, p 0.86) is a near miss. It is also a re-test of the same 563 runners that suggested it, so it adds no new evidence. The partition shows where the drag sits: spring turf returners at 16/1+ scored 0.653 on 373. Year-round AW returners at 16/1+ scored 1.28 on 626. Whatever the longshot drag is, it is not a general returner property, so I will not slice it further.
investigator v1
Answered16 Oct, 14:03
Are long-priced (16/1+) early-spring Flat handicap returners over-bet beyond the general longshot shortfall, while 16/1+ returners later in the year are not?
Not established. The 16/1+ early-spring returners scored 0.742 on 563 (p 0.86, re-tested on the same data), and AW 16/1+ returners year-round scored 1.28 on 626, so there is no general longshot-returner over-bet.
Does the March-April returner lift sit in all-weather races rather than turf, contrary to the turf-season mechanism?
Mostly yes at discovery: AW spring returners 1.195 on 480 against turf 1.027 on 944. Dropping the month still gives 1.138 on 1213, so it is not a start-of-turf-season effect, though it is not validated (p 0.11).
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Observed16 Oct, 14:05
H-98ae1db9 (Turf-to-all-weather handicap returners after 150+ days beat their pric) passed discovery as a niche: A/E 1.23 on 654 bets (p 0.06), same-race control 0.97.
result0.50reasoned about 16 Oct, 14:07from H-98ae1db9
Tested16 Oct, 14:05
refinement: GB only (the thread's Ireland drag): A/E 1.24 on 561 bets, p = 0.061 (95%: 0.94 to 1.60), by period 1.29 · 1.01. info
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Tested16 Oct, 14:05
control: Same-race other runners should score about 0.95-1.0 (overround-neutral). A second control separates the surface: A/E 1.01 on 559 bets, p = 0.49 (95%: 0.72 to 1.39), by period 0.91 · 1.26. info
H-e2d31d3d (Rival: turf-to-AW switchers after a short gap carry the same lift, so ) was predicted negative but came out the other way: A/E 1.07 on 4045 bets (p 0.089 for the opposite direction).
surprise0.82reasoned about 16 Oct, 14:07from H-e2d31d3d
Tested16 Oct, 14:05
neighbour: >= 150 -> 120 (-20%): A/E 1.29 on 899 bets, p = 0.011 (95%: 1.04 to 1.59), by period 1.31 · 1.22. pass
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
New question16 Oct, 14:07
Within turf-to-AW Flat handicap switchers, does A/E rise steadily with the length of the layoff (14-60, 61-149, 150+ days), as a turf-discount-plus-readiness mechanism predicts?
Short gap scored 1.07 on 4045, 150+ days 1.23 on 654 and 120+ days 1.295 on 899. A smooth gradient would support a two-part cause. A flat middle would mean the 150+ result is a lucky slice. Columns: days_since_last_run, surface, last_surface, is_handicap, region.
testingpriority 0.60raised by O-c26a3db5, O-e4c2a982
New question16 Oct, 14:07
Is the long-gap turf-to-AW lift carried by a handful of 16/1+ winners rather than spread across the price bands?
The 16/1+ band scored 1.496 on only 11 wins. Outside March-April, 16/1+ scored 1.68 on 8 wins while 8/1 to 14/1 scored 0.59. Earlier in this thread lifts sat in the short bands. Columns: sp_decimal, sp_rank, surface, last_surface, days_since_last_run.
openpriority 0.45raised by O-e4c2a982, O-c26a3db5
New question16 Oct, 14:07
Does the turf-to-AW lift exist in Ireland at all, or is it a GB AW-programme effect?
Three AW rules in this thread underperform in Ireland: IRE 0.449 on 40 (kill reason), short-gap IRE 0.87 on 473, and long-gap IRE 1.128 on only 93. If the effect is GB-only, the lab should restrict the population up front instead of re-killing rules on Ireland. Columns: region, course, surface, last_surface.
openpriority 0.35raised by O-64fa59c9, O-c26a3db5
New question16 Oct, 14:07
Is the turf-to-AW lift really a class drop, with turf horses dropping into weaker AW handicaps, rather than a surface discount?
Turf-to-AW switchers often drop in class, and that drop is visible but may be under-weighted. If switchers who do not drop in class score about 1.0, the surface story fails. GB only, because IRE race_class only starts in 2026-07. Columns: race_class, last_race_class, surface, last_surface, region.
testingpriority 0.50raised by O-c26a3db5, O-e4c2a982
To answer: Within turf-to-AW Flat handicap switchers, does A/E rise steadily with the length of the layoff (14-60, 61-149, 150+ days), as a turf-discount-plus-readiness mechanism predicts?
because There are two discounts. Punters distrust turf form on the AW at any gap. On top of that they mark down rust, and the longer the absence the bigger the mark-down. But the trainer picks the return, so the extra rust discount is undeserved. If both parts are real, a 2-5 month gap should carry more lift than a 2-8 week gap and less than 150+ days.
and the market misses it because The form book shows only turf runs and a long blank. Bettors weigh recent AW figures, and these horses have none. They treat the gap as a fitness worry rather than a planned return to a surface the yard chose.
predicting positive, A/E 1.08 to 1.18 on 800+ bets
and it is wrong if A/E below 1.02 on 800+ bets. A flat middle would mean the 150+ day result is a lucky slice, not a dose-response.
race_type = 'Flat' AND is_handicap = 1 AND surface = 'AW' AND last_surface = 'Turf' AND days_since_last_run BETWEEN 61 AND 149
To answer: Is the turf-to-AW lift really a class drop, with turf horses dropping into weaker AW handicaps, rather than a surface discount?
because AW handicaps are mostly lower class than turf handicaps. A horse switching from turf often drops a class or more, and the drop, not the surface, may make it competitive. If the surface discount is the cause, switchers at the same or a higher class should still beat their prices. If the class drop is the cause, they should be fairly priced.
and the market misses it because Under the surface story, punters mark down turf form on the AW even when the class is unchanged. Under the rival, the class drop is the under-weighted signal, and same-class switchers carry no edge.
predicting negative, A/E 0.92 to 1.03 on 1000+ bets
and it is wrong if A/E of 1.07 or higher on 1000+ bets. That would rule out the class drop and leave the surface discount standing.
race_type = 'Flat' AND is_handicap = 1 AND surface = 'AW' AND last_surface = 'Turf' AND region = 'GB' AND race_class <= last_race_class
The thread began as a four-year-old story (H126, 1.13 on 1904). It has turned into a story about horses moving from turf to the all-weather. The age-specific version is undercut: non-four-year-olds at 3/1 to 7/1 scored 1.166 on 333. Non-handicap returners are fairly priced (0.998 on 536). The latest three results sharpen the picture.
- Turf-to-AW handicap returners after 150+ days, in any month, scored 1.23 on 654 (p 0.06, niche, control 0.97).
- AW-to-AW returners after the same break scored only 1.015 on 559. Freshness alone does not carry the lift; the surface switch is needed.
- The rival I set up to fail, short-gap (14 to 60 day) turf-to-AW switchers, came out the opposite way: 1.07 on 4045, CI 0.97 to 1.18, GB 1.089, IRE 0.87.
So the gradient runs: AW-to-AW long gap about 1.0, turf-to-AW short gap 1.07, turf-to-AW long gap 1.23, and from 120 days 1.30 on 899. That points to two parts. The market discounts turf form on the AW at any gap, and a long break adds to that discount.
The weaknesses stay the same as before:
- **Ireland:** H-ce91b64d was killed on IRE 0.449, and short-gap IRE switchers scored 0.87.
- **Longshots:** the long-gap rule's 16/1+ band (1.496) rests on 11 wins. Outside March-April the lift is 1.12 on 371, but 8/1 to 14/1 there is 0.59.
- **Fading second period:** results drop in the later window (1.15 to 1.02 short gap; GB long gap 1.29 to 1.01).
The edge looks real in direction but is thin, mostly British and longshot-heavy. Before more slicing it needs a dose-response check and a class-drop rival test.
investigator v1
Answered16 Oct, 14:07
Is it the turf-to-AW switch that is mispriced at any layoff, or does it need a 150+ day absence?
Both the switch and the layoff matter, and the switch is necessary. Short-gap switchers are not fairly priced (1.07 on 4045), long-gap switchers score higher (1.23 on 654), and long-gap AW-to-AW returners score about 1.0 (1.015 on 559).
Bars grow from 1.00, the market. Faded bars have fewer than 40 bets.
Observed16 Oct, 14:13
H-98ae1db9 (Turf-to-all-weather handicap returners after 150+ days beat their pric) survived falsification: 5 of 5 attacks left it standing; it is frozen for its one vault query.
result0.55reasoned about 16 Oct, 14:15from H-98ae1db9
Tested16 Oct, 14:13
control: Control: turf-to-AW switchers after a medium 60-149 day gap: A/E 1.00 on 1,055 bets, p = 0.51 (95%: 0.80 to 1.24), by period 1.02 · 0.97. pass
H-f370f2f4 (Rival: turf-to-AW switchers that are not dropping in class are fairly ) was predicted negative but came out the other way: A/E 1.08 on 3797 bets (p 0.065 for the opposite direction).
surprise0.83reasoned about 16 Oct, 14:15from H-f370f2f4
Tested16 Oct, 14:13
broad rule: A/E 0.99 on 1,036 bets, p = 0.54 (95%: 0.79 to 1.23), by period 0.99 · 1.00. fail
Is the gap between discovery and vault in this thread a time effect (the market has priced turf-to-AW switchers since about 2024) or regression from selecting the best slice?
Two vault results (1.09→1.03 and 1.14→0.98) shrank sharply in the same direction. Splitting the turf-to-AW 150+ rule season by season (columns: last_surface, surface, days_since_last_run, is_handicap, race_type, region) would show whether discovery-era years are high and later years flat (a decay) or whether the lift is scattered noise. The answer decides whether to retire the AW branch for good.
openpriority 0.55raised by O-65b889c1, O-6d4bfc11
New question16 Oct, 14:15
Does the 1.08 for non-class-dropping turf-to-AW switchers survive in data outside the discovery window, or is it the same artefact as H-98ae1db9?
The surprise rests on 3797 bets with a CI that includes 1.0, drawn from the window the vault has just contradicted. A descriptive look at a later window (columns: last_surface, surface, race_class, last_race_class, region) would settle whether there is anything to test before spending a vault query.
openpriority 0.35raised by O-95f9a9aa, O-65b889c1
New question16 Oct, 14:15
Did 5 of 5 survived attacks fail to predict a vault contradiction because the attacks all use the discovery window and so cannot detect in-window overfitting?
This is a method question. If attacks survived and the vault still contradicted, the attack suite may need a time-holdout attack. That would change how much weight the lab gives a FROZEN status across all threads. Columns: race date or season, alongside the rule's own predicates.
openpriority 0.45raised by O-6d4bfc11, O-65b889c1
Reflected16 Oct, 14:15
The chain has narrowed the H126 lift step by step. First it moved from four-year-olds to all-age spring handicap returners (1.09 on 1424). Then it moved to all-weather spring returners (1.195 on 480, killed on IRE 0.449), and then to turf-to-AW returners after 150+ days in any month. That last rule, H-98ae1db9, scored 1.138 on 1213 at discovery with both periods above 1.1 and survived 5 of 5 attacks. It then failed the sealed vault outright: 0.979 on 3155, p 0.64. This is the second time the thread has shrunk badly from discovery to vault. H-5167b372 went from 1.09 to 1.03 on 6934, which counted as supportive but was far from significant.
The rival H-f370f2f4 (turf-to-AW switchers not dropping in class) was meant to be fairly priced but scored 1.08 on 3797. On its face that would say the surface switch itself is mispriced. However, it was measured on the same discovery window that produced the 1.138 the vault has just contradicted, and its CI (0.98 to 1.20) includes 1.0. The 16/1+ band scored 0.97, so it is not even the longshot pattern we saw before.
What we now believe: every turf-to-AW refinement in this thread was found and confirmed inside one window. The only out-of-sample test said there is no edge. So the surface-switch story should be treated as unproven, probably a discovery-window artefact. It should not be sliced further. What remains of the thread is the published won-last-time four-year-old system and the shadow spring rules, which are waiting on forward data.
investigator v1
Answered16 Oct, 14:15
Is the returner lift a year-round all-weather effect, strongest when a horse switches from a turf campaign to the AW after 150+ days, rather than anything seasonal?
No. The year-round turf-to-AW returner rule after 150+ days scored 1.14 at discovery but 0.979 on 3155 in the sealed vault (p 0.64), so there is no robust year-round AW switch effect.
