Test-mechanism: handicap runners-up raised 6 lb or more and back within 14 days beat their price

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-dd3724ed (Test-mechanism: easy handicap winners raised 6 lb or more and back wit) survived falsification: 5 of 5 attacks left it standing; it is frozen for its one vault query. O-825e5788
   • H-d9dabea2 (Rival: big-rise quick-returning handicap winners beat their price with) survived falsification: 5 of 6 attacks left it standing; it is frozen for its one vault query. O-8f406178
2. AskedIs the market's over-fade about the pounds raised, so it also shows on horses raised 6 lb or more after finishing second, without a win? testing
3. MechanismIf punters anchor on the pounds raised, any horse hit with a big rise after a strong run is faded, while its form peak, kept by a quick turnout, carries it close to the new mark.
4. Why the market misses itPunters and the form book show a big rise and no win. The price leans on that, but the handicapper has only reacted to a performance the horse is still close to repeating.
5. Predictedpositive, A/E 1.05 to 1.12 on 150+ bets
6. Falsified ifA/E at or below 1.00 on 150+ bets, or below 1 in either period, which would tie the lift to winning form, not the rise
7. Rule sketch

Mechanism and rationale

If punters anchor on the pounds raised, any horse hit with a big rise after a strong run is faded, while its form peak, kept by a quick turnout, carries it close to the new mark.

The rule

last_position = 2 AND days_since_last_run IS NOT NULL AND days_since_last_run <= 14 AND is_handicap = 1 AND last_is_handicap = 1 AND rating_change IS NOT NULL AND rating_change >= 6

Reads: last_position, days_since_last_run, is_handicap, last_is_handicap, rating_change

Pre-registration

Spec hash (sha256 of the canonical pre-spec)

6a3a108ed29c283f36f2b6612d7062fbfbc5a5f90f33fa4f6477961df7bfcec4

Frozen

not frozen

The definition can still change.

Confounders

• Runners-up are often short-priced next time, so the group leans to favourites, where A/E behaves differently; check by price band, since the family's big-rise lift sat at 7/1 or shorter
• A 6 lb+ rise after a second may mark unusual races (for example, well clear of the third, or a strong race), so the rise may stand in for race quality rather than punter anchoring
• Code mix: Flat and jumps handicappers react differently, and quick 14-day returns are much more common on the Flat; report by race_type
• Period instability: earlier small-rise slices held in only one period, so require A/E above 1 in both periods
• Region: IRE samples were too thin before; report GB and IRE separately
• Same-race control: if other runners in these races also beat their prices, the lift is a property of the race

Control

{
  "description": "Same-race other runners should sit near A/E 1.0 (about 0.95 to 1.02); if they also beat their prices the lift belongs to the races, not the rule. The extra control group is handicap runners-up back within 14 days whose mark rose 0 to 5 lb: the mechanism says the fade comes from the big visible rise, so this group should score near 1.0 and clearly below the 6 lb+ rule. If it scores as well as the rule, the lift is runner-up form plus a quick turnout, not the rise.",
  "rule": "last_position = 2 AND days_since_last_run IS NOT NULL AND days_since_last_run <= 14 AND is_handicap = 1 AND last_is_handicap = 1 AND rating_change IS NOT NULL AND rating_change >= 0 AND rating_change <= 5"
}

Discovery result

fail broad rule · 16 Oct, 16:36

Bets

59

6 winners

A/E

0.58

95%: 0.21 to 1.26

SP return

-42.9%

level stakes

Betfair SP return

-23.4%

58 bets with a BSP

Strike rate 10.2%. p = 0.94. Same-race control: the other runners in these races had A/E 1.10, 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, 16:36

  Output (schema-checked)

  {
    "rule": "last_position = 2 AND days_since_last_run IS NOT NULL AND days_since_last_run <= 14 AND is_handicap = 1 AND last_is_handicap = 1 AND rating_change IS NOT NULL AND rating_change >= 6",
    "refinements": [
      {
        "label": "7/1 or shorter",
        "rule": "last_position = 2 AND days_since_last_run IS NOT NULL AND days_since_last_run <= 14 AND is_handicap = 1 AND last_is_handicap = 1 AND rating_change IS NOT NULL AND rating_change >= 6 AND sp_decimal IS NOT NULL AND sp_decimal <= 8"
      },
      {
        "label": "Close second (beaten 2 lengths or less)",
        "rule": "last_position = 2 AND days_since_last_run IS NOT NULL AND days_since_last_run <= 14 AND is_handicap = 1 AND last_is_handicap = 1 AND rating_change IS NOT NULL AND rating_change >= 6 AND last_beaten_lengths IS NOT NULL AND last_beaten_lengths <= 2"
      }
    ],
    "control": {
      "description": "Same-race other runners should sit near A/E 1.0 (about 0.95 to 1.02); if they also beat their prices the lift belongs to the races, not the rule. The extra control group is handicap runners-up back within 14 days whose mark rose 0 to 5 lb: the mechanism says the fade comes from the big visible rise, so this group should score near 1.0 and clearly below the 6 lb+ rule. If it scores as well as the rule, the lift is runner-up form plus a quick turnout, not the rise.",
      "rule": "last_position = 2 AND days_since_last_run IS NOT NULL AND days_since_last_run <= 14 AND is_handicap = 1 AND last_is_handicap = 1 AND rating_change IS NOT NULL AND rating_change >= 0 AND rating_change <= 5"
    },
    "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": [
      "Runners-up are often short-priced next time, so the group leans to favourites, where A/E behaves differently; check by price band, since the family's big-rise lift sat at 7/1 or shorter",
      "A 6 lb+ rise after a second may mark unusual races (for example, well clear of the third, or a strong race), so the rise may stand in for race quality rather than punter anchoring",
      "Code mix: Flat and jumps handicappers react differently, and quick 14-day returns are much more common on the Flat; report by race_type",
      "Period instability: earlier small-rise slices held in only one period, so require A/E above 1 in both periods",
      "Region: IRE samples were too thin before; report GB and IRE separately",
      "Same-race control: if other runners in these races also beat their prices, the lift is a property of the race"
    ],
    "mechanism": "Punters anchor on the pounds raised. Any handicapper given a big rise (6 lb or more) after a strong run is faded, even one that only finished second. A quick turnout (14 days or less) keeps the horse near its form peak, so it runs close to the new mark more often than its price implies. This tests whether the lift seen in raised easy winners comes from the visible rise itself rather than from winning form.",
    "market_miss": "The form book and punters see a big rise with no win to justify it, and the price leans on that. But the handicapper has only reacted to a performance the horse can still nearly repeat on a quick return, so the rise is less costly than the market thinks.",
    "prediction": {
      "direction": "positive",
      "size": "A/E 1.05 to 1.12 on 150+ bets, with the lift at 7/1 or shorter and smaller than the 1.47 seen for easy winners",
      "falsified_if": "A/E at or below 1.00 on 150+ bets, or below 1.00 in either period, or no better than the 0 to 5 lb runner-up control. Any of these would tie the family's lift to winning form, not to the rise."
    },
    "notes": "I kept the sketch and added only NULL guards on days_since_last_run and rating_change, matching earlier family rules. Both refinements come from the reasoning and the family findings. The price cap follows the finding that big-rise quick returners do not win at 8/1 or bigger. The close-second cap follows the H19-style result for narrow defeats."
  }

Timeline

1. Job 2042 finished

   16 Oct, 16:36 · engine:compute-1

2. H-da54bdc8 Discovery → KILLED fails: 59 bets, needs 150; A/E 0.581, needs 1.05; p 0.9443, needs 0.05 or lower; periods [], each needs 1.0; A/E 0.581 against same-race control 1.096, needs 0.03 above (not a niche pass: A/E 0.581 below 1.05; p 0.944349528227739 above 0.5; does not beat the same-race control 1.096)

   16 Oct, 16:36 · engine:evaluator

3. discovery result for H-da54bdc8 (broad rule): fail

   16 Oct, 16:36 · engine:evaluator

4. H-da54bdc8 Designed → RESEARCHdiscovery measuring on the discovery zone

   16 Oct, 16:36 · engine:orchestrator

5. compute-1 took job 2042 (evaluate, attempt 1)

   16 Oct, 16:36 · compute-1

6. Job 2040 finished

   16 Oct, 16:36 · engine:ai-1

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

   16 Oct, 16:36 · engine:scheduler

8. H-da54bdc8 Proposed → RESEARCHdesigned pre-spec written and hashed

   16 Oct, 16:36 · agent:designer@v1

9. ai-1 took job 2040 (design, attempt 1)

   16 Oct, 16:36 · ai-1

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

    16 Oct, 16:35 · engine:scheduler

11. New idea H-da54bdc8: Test-mechanism: handicap runners-up raised 6 lb or more and back within 14 days beat their price (test-mechanism)

    16 Oct, 16:35 · 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.