Young lightly raced chasers at 3m to 3m3f are over-bet: the youth signal flips sign below marathon trips

Reasoning

Why this hypothesis exists: what the lab saw, the question it asked, the mechanism it proposed, what it predicted, and the result that would prove it wrong. Written before the test was run.

1. Observed
   • H-aa66431f (Young marathon handicap chasers with little chase experience carry the) survived falsification: 2 of 3 attacks left it standing; it is frozen for its one vault query. O-6d75cec0
   • H-bb6d2ec0 (Exposed young marathon handicap chasers sit at par: the lift needs an ) pointed the predicted way but failed the bar: A/E 0.84 on 33 bets (p 0.694), control 1.04. O-8596d289
   • H-aa66431f (Young marathon handicap chasers with little chase experience carry the) met the sealed vault supportive: A/E 1.2 on 290 bets (p 0.1357, threshold 0.0023). O-bebe2c91
2. AskedDo young (7 or under), lightly raced (6 or fewer chase runs) handicap chasers at 3m to 3m3f lose to their price, so the same profile flips sign at the 3m4f line? answered
3. MechanismPunters fit a young, lightly raced staying chaser to a story. At ordinary staying trips (3m to 3m3f) the story is 'improving novice-type, well handicapped' and the horse is backed beyond its chances, while inexperienced jumping still costs it against hardened rivals. At 3m4f+ the opposite folk belief, that marathons need old, battle-hardened horses, makes the same profile drift. The handicap mark is the same in both cases, so the gap is market sentiment about youth at a given trip.
4. Why the market misses itAt standard trips the young improver is the obvious 'unexposed, could be well in' bet and is over-backed. The form book shows the potential but not the jumping cost against experienced fields at a strong staying pace. At marathon trips the price carries a stamina-doubt discount that the horse's actual stamina does not justify.
5. Predictednegative, A/E 0.80 to 0.92 on 180+ bets
6. Falsified if

The question's answer so far: Not convincingly. Young chasers with 6 or fewer chase runs at 3m to 3m3f scored 0.93 on 743 bets (CI 0.76 to 1.13, p 0.78) against a same-race control of 1.06, so the sign leans negative but is not distinguishable from par, and 0 wins from 107 bets at 16/1+ carry much of it.

Mechanism and rationale

Punters fit a young, lightly raced staying chaser to a story. At ordinary staying trips (3m to 3m3f) the story is 'improving novice-type, well handicapped' and the horse is backed beyond its chances, while inexperienced jumping still costs it against hardened rivals. At 3m4f+ the opposite folk belief, that marathons need old, battle-hardened horses, makes the same profile drift. The handicap mark is the same in both cases, so the gap is market sentiment about youth at a given trip.

The rule

race_type = 'Chase' AND is_handicap = 1 AND age <= 7 AND distance_yards >= 5280 AND distance_yards < 6160 AND prior_chase_runs <= 6

Reads: race_type, is_handicap, age, distance_yards, prior_chase_runs

Pre-registration

Spec hash (sha256 of the canonical pre-spec)

9c912ea68b97fa0af518f276467d468478a73cb1dfcb2d2cd0400c24ca89dda6

Frozen

not frozen

The definition can still change.

Confounders

• Novice handicap chases at 3m have a faster pace and more fallers; is_novice = 0 refinement separates them
• Overlap with the earlier NEGATIVE grade for young chasers at 3m-3m3f (H-910e15e3, 0.92 on 971); the exposed-young control tests whether this is a restatement
• Big-yard novices are popular at standard trips, so trainer reputation may drive the over-betting rather than youth or inexperience
• Short-price cluster: earlier positive marathon slices in this family were carried by the up-to-7/1 band, so a negative result here may also sit in one price band
• Cheltenham and other festival novice handicaps may be over-represented and carry their own market biases
• prior_chase_runs is left-censored before August 2020 (low risk in the discovery window)
• Trip boundary at 6160 yards was chosen after the marathon lift was seen, so the sign-flip framing is partly post hoc

Control

{
  "description": "Same-race control: the other runners in these 3m-3m3f handicap chases should sit near par (A/E about 0.95-1.02). If they are just as far below 1.0, the shortfall belongs to the race, not to the young unexposed profile. Second comparison: exposed young chasers (7+ chase runs) at the same trip in handicaps. The mechanism says inexperience is what is over-bet, so this group should sit near par. If it shows the same shortfall, the effect is youth (or the earlier 0.92-on-971 NEGATIVE grade) restated, not the unexposed-improver story.",
  "rule": "race_type = 'Chase' AND is_handicap = 1 AND age <= 7 AND distance_yards >= 5280 AND distance_yards < 6160 AND prior_chase_runs >= 7"
}

Discovery result

fail broad rule · 16 Oct, 14:41

Bets

743

106 winners

A/E

0.93

95%: 0.76 to 1.13

SP return

-34.5%

level stakes

Betfair SP return

-22.7%

721 bets with a BSP

Strike rate 14.3%. p = 0.78. Same-race control: the other runners in these races had A/E 1.06, so the selection did not beat its own races. Zone: discovery. Evaluator eval-1.0, data 6a4e98be18e83933

Falsification

No attacks yet. Only promising ideas are attacked.

Sealed holdout (vault)

Not queried. Only a frozen idea can spend its one query against the 2020-25 holdout.

Forward record

No forward record. Shadow tracking starts after the vault and counts only races run after the freeze.

Lineage

Agent runs

• designer v1ok · 36s · 16 Oct, 14:40

  Output (schema-checked)

  {
    "rule": "race_type = 'Chase' AND is_handicap = 1 AND age <= 7 AND distance_yards >= 5280 AND distance_yards < 6160 AND prior_chase_runs <= 6",
    "refinements": [
      {
        "label": "Short-priced story horses (up to 8/1)",
        "rule": "race_type = 'Chase' AND is_handicap = 1 AND age <= 7 AND distance_yards >= 5280 AND distance_yards < 6160 AND prior_chase_runs <= 6 AND sp_decimal <= 9.0"
      },
      {
        "label": "Most unexposed (3 or fewer chase runs)",
        "rule": "race_type = 'Chase' AND is_handicap = 1 AND age <= 7 AND distance_yards >= 5280 AND distance_yards < 6160 AND prior_chase_runs <= 3"
      },
      {
        "label": "Open handicaps only (excludes novice handicap chases)",
        "rule": "race_type = 'Chase' AND is_handicap = 1 AND age <= 7 AND distance_yards >= 5280 AND distance_yards < 6160 AND prior_chase_runs <= 6 AND is_novice = 0"
      }
    ],
    "control": {
      "description": "Same-race control: the other runners in these 3m-3m3f handicap chases should sit near par (A/E about 0.95-1.02). If they are just as far below 1.0, the shortfall belongs to the race, not to the young unexposed profile. Second comparison: exposed young chasers (7+ chase runs) at the same trip in handicaps. The mechanism says inexperience is what is over-bet, so this group should sit near par. If it shows the same shortfall, the effect is youth (or the earlier 0.92-on-971 NEGATIVE grade) restated, not the unexposed-improver story.",
      "rule": "race_type = 'Chase' AND is_handicap = 1 AND age <= 7 AND distance_yards >= 5280 AND distance_yards < 6160 AND prior_chase_runs >= 7"
    },
    "pass_criteria": {
      "min_bets": 180,
      "min_ae": 1.05,
      "max_p": 0.05
    },
    "forward_bar": {
      "min_bets": 100,
      "min_ae": 1,
      "max_p": 0.2
    },
    "confounders": [
      "Novice handicap chases at 3m have a faster pace and more fallers; is_novice = 0 refinement separates them",
      "Overlap with the earlier NEGATIVE grade for young chasers at 3m-3m3f (H-910e15e3, 0.92 on 971); the exposed-young control tests whether this is a restatement",
      "Big-yard novices are popular at standard trips, so trainer reputation may drive the over-betting rather than youth or inexperience",
      "Short-price cluster: earlier positive marathon slices in this family were carried by the up-to-7/1 band, so a negative result here may also sit in one price band",
      "Cheltenham and other festival novice handicaps may be over-represented and carry their own market biases",
      "prior_chase_runs is left-censored before August 2020 (low risk in the discovery window)",
      "Trip boundary at 6160 yards was chosen after the marathon lift was seen, so the sign-flip framing is partly post hoc"
    ],
    "mechanism": "Punters attach a story to a young (7 or under), lightly raced (6 or fewer chase runs) handicap chaser at ordinary staying trips of 3m to 3m3f: 'unexposed improver, could be well handicapped'. They back it beyond its chances, while its inexperienced jumping still costs it against hardened rivals at a strong staying pace. At 3m4f+ a different folk belief (marathons need old, battle-hardened horses) makes the same profile drift, which matches the lift seen there (1.24 on 75, 1.20 on 290 in the vault). The handicap mark works the same way at both trips, so the sign flip would come from market sentiment about youth at a given trip.",
    "market_miss": "At standard staying trips, the young unexposed chaser is the obvious 'could be well in' bet and is over-backed. The form book shows its potential but not the jumping cost against experienced fields. At marathon trips the price carries a stamina-doubt discount that the horse's actual stamina does not justify. So the market's view of youth flips with trip, while the horse's real disadvantage (inexperience) and advantage (unexposed mark) do not.",
    "prediction": {
      "direction": "negative",
      "size": "A/E 0.80 to 0.92 on 180+ bets",
      "falsified_if": "A/E at or above 1.0 on 150+ bets, or the exposed-young control (7+ chase runs) shows an equal or larger shortfall. Either would mean there is no inexperience-specific sign flip: the 3m4f+ lift stands alone and the 3m1f+ shortfall was noise or a plain youth effect."
    },
    "notes": "I checked that prior_chase_runs is a pre_off column (available from 2020-08 in GB and IRE), and kept the draft rule unchanged. The prediction is negative, so I read min_ae 1.05 in the predicted direction: the selection must be over-bet by at least that margin, roughly A/E at or below 0.95. The rule overlaps heavily with H-717bd1f9 (chase_runs <= 3, 0.95 on 472) and H-910e15e3 (0.92 on 971). Those were killed for being too weak, so this needs a clearly stronger shortfall than they showed. The 'up to 8/1' refinement uses sp_decimal <= 9.0."
  }

Timeline

1. Job 1827 finished

   16 Oct, 14:41 · engine:compute-1

2. H-b6c1b1c0 Discovery → KILLED fails: p 0.2482 (lower tail), needs 0.05

   16 Oct, 14:41 · engine:evaluator

3. discovery result for H-b6c1b1c0 (broad rule): fail

   16 Oct, 14:41 · engine:evaluator

4. H-b6c1b1c0 Designed → RESEARCHdiscovery measuring on the discovery zone

   16 Oct, 14:41 · engine:orchestrator

5. compute-1 took job 1827 (evaluate, attempt 1)

   16 Oct, 14:41 · compute-1

6. Job 1826 finished

   16 Oct, 14:41 · engine:ai-1

7. Queued evaluate on the compute lane: discovery run of the broad rule

   16 Oct, 14:41 · engine:scheduler

8. H-b6c1b1c0 Proposed → RESEARCHdesigned pre-spec written and hashed

   16 Oct, 14:41 · agent:designer@v1

9. ai-1 took job 1826 (design, attempt 1)

   16 Oct, 14:40 · ai-1

10. Queued design on the ai lane: write the pre-spec

    16 Oct, 14:40 · engine:scheduler

11. New idea H-b6c1b1c0: Young lightly raced chasers at 3m to 3m3f are over-bet: the youth signal flips sign below marathon trips (test-mechanism)

    16 Oct, 14:40 · agent:investigator@v1

Every change is an event with an actor: an agent and its version, an engine, a human, or the importer. Current stage: Killed.