Quick-returning last-time winners of any race running in a handicap at 9/1+ are over-bet

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-220ad315 (Raised handicap winners sent off at 9/1+ are over-bet, at any gap sinc) pointed the predicted way but failed the bar: A/E 0.84 on 1561 bets (p 0.933), control 1.02. O-76793565
   • H-21f9e6e7 (Long-priced raised handicap winners turned out within 14 days are over) pointed the predicted way but failed the bar: A/E 0.57 on 255 bets (p 0.967), control 1.03. O-7c9eecf8
   • H-560b6713 (Generalise: last-time winners of any race type running in a handicap a) pointed the predicted way but failed the bar: A/E 0.88 on 2634 bets (p 0.929), control 1.02. O-deb65699
2. AskedIs the long-price winner deficit concentrated in quick returners (days_since_last_run <= 14), across every type of previous race, as the 0.57-on-255 near miss suggested? answered
3. MechanismA horse turned out within two weeks of a win often runs under a penalty or a new mark before it has recovered. A long price then reflects the professionals' view that it has bounced or is now held. Recent-winner money from the public keeps the price shorter than that view justifies.
4. Why the market misses itThe form book shows a fresh '1' and the recency makes it look in form. The public backs the visible win and does not price the physical cost of a quick turnaround or the handicapper's response.
5. Predictednegative, A/E 0.65 to 0.85 on 500+ bets
6. Falsified ifA/E 0.93 or higher on 400+ bets, or a deficit that rests on Ireland alone with GB at par.
7. Rule sketch

The question's answer so far: Yes, as far as the data can show. Quick-returning last-time winners of any previous race type scored 0.651 on 424 at 9/1+, against 0.879 for all gaps, in both GB and IRE. It is still not significant (p 0.051).

Mechanism and rationale

A horse turned out within two weeks of a win often runs under a penalty or a new mark before it has recovered. A long price then reflects the professionals' view that it has bounced or is now held. Recent-winner money from the public keeps the price shorter than that view justifies.

The rule

last_position = 1 AND is_handicap = 1 AND sp_decimal > 9 AND days_since_last_run IS NOT NULL AND days_since_last_run <= 14

Reads: last_position, is_handicap, sp_decimal, days_since_last_run

Pre-registration

Spec hash (sha256 of the canonical pre-spec)

da4517ff1d05a0f5b24e99d01e3ab5ca4654d637430d68e35ce5c61c6f03da8b

Frozen

not frozen

The definition can still change.

Confounders

• Penalised horses carrying top weight or near it: the penalty, not the turnaround, may be what holds them
• Season mix: summer Flat turnarounds (common, often under penalties) versus winter jumps quick returns (rarer, more physically costly)
• Yards that routinely turn horses out quickly and may be priced differently by the market
• Ireland concentration: earlier winner pools were weaker in IRE (0.78) than GB (0.91), so the deficit must hold in GB
• Overlap with the pooled 9/1+ handicap winner set (0.88 on 2,634): a quick-return subset may simply inherit that pool's deficit
• Race class drop or rise after the win, which moves both the price and the chance

Control

{
  "description": "Other runners in the same races should be close to par (A/E about 0.97–1.03), which shows the deficit belongs to the horses and not to the races. The control that tests the mechanism is last-time winners running in a handicap at 9/1+ that returned after more than 14 days. They should sit nearer par, at or above the pooled 0.88. If the slower returners are as over-bet as the quick ones, the deficit is plain winner-tag money and not a quick-turnaround effect.",
  "rule": "last_position = 1 AND is_handicap = 1 AND sp_decimal > 9 AND days_since_last_run IS NOT NULL AND days_since_last_run > 14"
}

Discovery result

fail broad rule · 16 Oct, 17:00

Bets

424

15 winners

A/E

0.65

95%: 0.36 to 1.07

SP return

-57.5%

level stakes

Betfair SP return

-45.5%

396 bets with a BSP

Strike rate 3.5%. p = 0.97. Same-race control: the other runners in these races had A/E 1.02, 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 · 29s · 16 Oct, 16:58

  Output (schema-checked)

  {
    "rule": "last_position = 1 AND is_handicap = 1 AND sp_decimal > 9 AND days_since_last_run IS NOT NULL AND days_since_last_run <= 14",
    "refinements": [
      {
        "label": "9/1 to 14/1 band (where the pooled winner deficit sat)",
        "rule": "last_position = 1 AND is_handicap = 1 AND sp_decimal > 9 AND sp_decimal <= 15 AND days_since_last_run IS NOT NULL AND days_since_last_run <= 14"
      },
      {
        "label": "Very quick return, within a week",
        "rule": "last_position = 1 AND is_handicap = 1 AND sp_decimal > 9 AND days_since_last_run IS NOT NULL AND days_since_last_run <= 7"
      },
      {
        "label": "GB only (guards against the Ireland-only pattern)",
        "rule": "last_position = 1 AND is_handicap = 1 AND sp_decimal > 9 AND days_since_last_run IS NOT NULL AND days_since_last_run <= 14 AND region = 'GB'"
      }
    ],
    "control": {
      "description": "Other runners in the same races should be close to par (A/E about 0.97–1.03), which shows the deficit belongs to the horses and not to the races. The control that tests the mechanism is last-time winners running in a handicap at 9/1+ that returned after more than 14 days. They should sit nearer par, at or above the pooled 0.88. If the slower returners are as over-bet as the quick ones, the deficit is plain winner-tag money and not a quick-turnaround effect.",
      "rule": "last_position = 1 AND is_handicap = 1 AND sp_decimal > 9 AND days_since_last_run IS NOT NULL AND days_since_last_run > 14"
    },
    "pass_criteria": {
      "min_bets": 150,
      "min_ae": 1.05,
      "max_p": 0.05
    },
    "forward_bar": {
      "min_bets": 100,
      "min_ae": 1,
      "max_p": 0.2
    },
    "confounders": [
      "Penalised horses carrying top weight or near it: the penalty, not the turnaround, may be what holds them",
      "Season mix: summer Flat turnarounds (common, often under penalties) versus winter jumps quick returns (rarer, more physically costly)",
      "Yards that routinely turn horses out quickly and may be priced differently by the market",
      "Ireland concentration: earlier winner pools were weaker in IRE (0.78) than GB (0.91), so the deficit must hold in GB",
      "Overlap with the pooled 9/1+ handicap winner set (0.88 on 2,634): a quick-return subset may simply inherit that pool's deficit",
      "Race class drop or rise after the win, which moves both the price and the chance"
    ],
    "mechanism": "A horse turned out within 14 days of a win often runs under a penalty or a new, higher mark before it has fully recovered. A price of 9/1 or longer in a handicap shows that the professional view is that it has bounced or is now held by the handicapper. Public money on the visible recent win still keeps the price shorter than that view justifies. So these runners win less often than their SP implies, and more so than winners returning after a longer gap.",
    "market_miss": "The form figures show a fresh '1' and the short gap makes the horse look in form. Recreational money backs that visible win and does not price the physical cost of a quick turnaround or the handicapper's response (a penalty or new mark). The drift to 9/1+ only partly corrects for this.",
    "prediction": {
      "direction": "negative",
      "size": "A/E 0.65 to 0.85 on 500+ bets, below the >14-day control group",
      "falsified_if": "A/E 0.93 or higher on 400+ bets; or a deficit that rests on Ireland alone with GB at par (GB A/E >= 0.95); or the quick-return group no lower than the >14-day control"
    },
    "notes": "The rule sketch is kept and only adds an explicit null guard on days_since_last_run. All columns are pre-off and available from July 2020 in GB and IRE. The 9/1–14/1 refinement comes from a band found after looking at the pooled winner set (0.84 at up to 14/1 vs 1.01 at 16/1+), so treat it as confirmatory only. The pass criteria follow the lab floors read in the negative direction, so A/E 1.05 means a lay-side edge."
  }

Timeline

1. Job 2080 finished

   16 Oct, 17:00 · engine:compute-1

2. H-9e92ea71 Discovery → KILLED fails: p 0.0511 (lower tail), needs 0.05

   16 Oct, 17:00 · engine:evaluator

3. discovery result for H-9e92ea71 (broad rule): fail

   16 Oct, 17:00 · engine:evaluator

4. H-9e92ea71 Designed → RESEARCHdiscovery measuring on the discovery zone

   16 Oct, 17:00 · engine:orchestrator

5. compute-1 took job 2080 (evaluate, attempt 1)

   16 Oct, 17:00 · compute-1

6. Job 2078 finished

   16 Oct, 16:59 · engine:ai-1

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

   16 Oct, 16:59 · engine:scheduler

8. H-9e92ea71 Proposed → RESEARCHdesigned pre-spec written and hashed

   16 Oct, 16:59 · agent:designer@v1

9. ai-1 took job 2078 (design, attempt 1)

   16 Oct, 16:58 · ai-1

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

    16 Oct, 16:58 · engine:scheduler

11. New idea H-9e92ea71: Quick-returning last-time winners of any race running in a handicap at 9/1+ are over-bet (generalise)

    16 Oct, 16:58 · 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.