Young novice chasers on their chase debut 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
   • H45-young-horses-in-novice-chases (H45: young horses in novice chases) was GREEN in the earlier research: A/E 1.09 on 435 bets (periods 1.00 / 1.27 / 1.15); now PROMISING. O-35f7497e
   • H124-fit-hurdler-making-its-chase-debut (H124: fit hurdler making its chase debut) was predicted positive but went the other way in the earlier research: A/E 0.91 on 1173 bets (periods 1.04 / 0.81 / 0.87). O-b063770a
2. AskedIs the young-novice-chaser lift concentrated on chase debuts (chase_runs = 0), where only the yard knows how well the horse jumps? testing
3. MechanismA yard sends a four- or five-year-old over fences only after schooling shows it jumps well. On its chase debut that private judgement is the only information available, and the young age is the visible sign of it.
4. Why the market misses itThe public form is all hurdles. The market prices the horse as an immature hurdler taking on older, stronger chasers, and has no evidence of its jumping. After a run or two over fences its jumping is on view and gets priced.
5. Predictedpositive, A/E 1.08 to 1.15 on 100+ bets
6. Falsified ifA/E at or below 1.00, or no higher than the same rule with chase_runs >= 1
7. Rule sketch

Mechanism and rationale

A yard sends a four- or five-year-old over fences only after schooling shows it jumps well. On its chase debut that private judgement is the only information available, and the young age is the visible sign of it.

The rule

race_type = 'Chase' AND is_novice = 1 AND is_handicap = 0 AND age <= 5 AND chase_runs = 0

Reads: race_type, is_novice, is_handicap, age, chase_runs

Pre-registration

Spec hash (sha256 of the canonical pre-spec)

b99177c01790974339756ff638b1d3999ed90559e5367c64a264a490e23ac1d5

Frozen

not frozen

The definition can still change.

Confounders

• Small fields in novice chases: the yardstick must be corrected for field size, and rivals in the same races are the check
• Chase debutants in general are over-bet (H124, 0.91), so the young-age subset has to beat that baseline as well as 1.00
• Short-priced favourites dominate small novice chases, so a few odds-on winners or losers can swing A/E
• French-bred imports are concentrated in a few yards (Nicholls, Henderson and others), so the result may be a yard effect, not an age effect
• chase_runs only counts runs from August 2020, so in the first months a 5yo with earlier chase runs (for example in France) could be wrongly counted as a debutant
• Period dependence: the parent H45 edge was flat in 2020-22 and the lab re-measured it at 0.96, so every period has to agree

Control

{
  "description": "The comparison that matters for this question is the same young novice chasers once they have run over fences. If the edge comes from the yard's private view of the horse's jumping, debutants should beat it clearly, because by then the jumping is public and priced. Rivals in the same races (mostly older novices and other debutants) should sit at or below 1.00. If they show the same lift, the effect belongs to these races (small fields, favourite-dominated books), not to the young debutants.",
  "rule": "race_type = 'Chase' AND is_novice = 1 AND is_handicap = 0 AND age <= 5 AND chase_runs >= 1"
}

Discovery result

fail broad rule · 16 Oct, 11:04

Bets

11

1 winners

A/E

0.53

95%: 0.01 to 2.97

SP return

-76.1%

level stakes

Betfair SP return

–

10 bets with a BSP

Strike rate 9.1%. p = 0.85. Same-race control: the other runners in these races had A/E 1.14, so the selection did not beat its own races. Zone: discovery. Evaluator eval-1.0, data 2887c9abd664f46f

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 · 24s · 16 Oct, 11:04

  Output (schema-checked)

  {
    "rule": "race_type = 'Chase' AND is_novice = 1 AND is_handicap = 0 AND age <= 5 AND chase_runs = 0",
    "refinements": [
      {
        "label": "public form is hurdles only",
        "rule": "race_type = 'Chase' AND is_novice = 1 AND is_handicap = 0 AND age <= 5 AND chase_runs = 0 AND hurdle_runs >= 1"
      },
      {
        "label": "14/1 or shorter",
        "rule": "race_type = 'Chase' AND is_novice = 1 AND is_handicap = 0 AND age <= 5 AND chase_runs = 0 AND sp_decimal <= 15"
      }
    ],
    "control": {
      "description": "The comparison that matters for this question is the same young novice chasers once they have run over fences. If the edge comes from the yard's private view of the horse's jumping, debutants should beat it clearly, because by then the jumping is public and priced. Rivals in the same races (mostly older novices and other debutants) should sit at or below 1.00. If they show the same lift, the effect belongs to these races (small fields, favourite-dominated books), not to the young debutants.",
      "rule": "race_type = 'Chase' AND is_novice = 1 AND is_handicap = 0 AND age <= 5 AND chase_runs >= 1"
    },
    "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": [
      "Small fields in novice chases: the yardstick must be corrected for field size, and rivals in the same races are the check",
      "Chase debutants in general are over-bet (H124, 0.91), so the young-age subset has to beat that baseline as well as 1.00",
      "Short-priced favourites dominate small novice chases, so a few odds-on winners or losers can swing A/E",
      "French-bred imports are concentrated in a few yards (Nicholls, Henderson and others), so the result may be a yard effect, not an age effect",
      "chase_runs only counts runs from August 2020, so in the first months a 5yo with earlier chase runs (for example in France) could be wrongly counted as a debutant",
      "Period dependence: the parent H45 edge was flat in 2020-22 and the lab re-measured it at 0.96, so every period has to agree"
    ],
    "mechanism": "Yards rarely send a four- or five-year-old chasing unless schooling at home shows it jumps fences well. At that age, waiting another season over hurdles is the default. On the horse's chase debut, the yard's private judgement of its jumping is the only information about how it will cope with fences. The young age is the visible sign that this judgement was positive.",
    "market_miss": "The horse's public form is all hurdles. The market prices it as an immature hurdler taking on older, stronger novices, with no evidence of its jumping. After one or two chase runs its jumping is on view and gets priced, so any edge should be on the debut and should shrink on later chase runs.",
    "prediction": {
      "direction": "positive",
      "size": "A/E 1.08 to 1.15 on roughly 150-250 bets, above the same rule with chase_runs >= 1",
      "falsified_if": "A/E at or below 1.00; or A/E no higher than the control (same rule with chase_runs >= 1); or the lift appears in only one period, with 2020-22 flat as in the parent H45."
    },
    "notes": "This rule is a subset of the live H45 system. It answers Q-c21180d6: is the edge concentrated on chase debut? The lab's own re-measure of H45 was 0.963 on 204 bets, so the sample may be thin. In that case it goes to falsification and forward tracking rather than being killed for size. Refinements are kept to two and both come from the reasoning: (1) the market miss assumes hurdle-only public form; (2) edges tend to die at big prices. There is no 16/1+ split, because a longshot lift is a general property and not evidence for this mechanism. If the result is real, check by trainer next, to rule out the French-import yard confounder."
  }

Timeline

1. Job 1224 finished

   16 Oct, 11:04 · engine:compute-2

2. H-fd37b556 Discovery → KILLED fails: 11 bets, needs 150; A/E 0.533, needs 1.05; p 0.8468, needs 0.05 or lower; periods [], each needs 1.0; A/E 0.533 against same-race control 1.139, needs 0.03 above (not a niche pass: only 11 bets; A/E 0.533 below 1.05; p 0.8467983114699859 above 0.5; does not beat the same-race control 1.139)

   16 Oct, 11:04 · engine:evaluator

3. discovery result for H-fd37b556 (broad rule): fail

   16 Oct, 11:04 · engine:evaluator

4. H-fd37b556 Designed → RESEARCHdiscovery measuring on the discovery zone

   16 Oct, 11:04 · engine:orchestrator

5. compute-2 took job 1224 (evaluate, attempt 1)

   16 Oct, 11:04 · compute-2

6. Job 1222 finished

   16 Oct, 11:04 · engine:ai-1

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

   16 Oct, 11:04 · engine:scheduler

8. H-fd37b556 Proposed → RESEARCHdesigned pre-spec written and hashed

   16 Oct, 11:04 · agent:designer@v1

9. ai-1 took job 1222 (design, attempt 1)

   16 Oct, 11:04 · ai-1

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

    16 Oct, 11:04 · engine:scheduler

11. New idea H-fd37b556: Young novice chasers on their chase debut beat their price (test-mechanism)

    16 Oct, 11:04 · 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.