Rival: the debutant lift is handicap marks after heavy hurdle beatings, in both countries

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-448cb451 (Rival: chase debutants after any heavy hurdle beating beat their price) passed discovery but was falsified: IRE A/E 0.906. O-2930585a
   • H-551aa0a1 (Test-mechanism: chase debutants beaten heavily last time over hurdles ): the country 'GB' carries the effect (A/E 1.86) while the rest score IRE 0.73. O-4137a04c
   • H-2e457254 (Rival: the debutant lift is the yard's chase-debut skill, not misread ): the country 'GB' carries the effect (A/E 1.27) while the rest score IRE 0.79. O-5c620d47
   • H-551aa0a1 (Test-mechanism: chase debutants beaten heavily last time over hurdles ) passed discovery but was falsified: IRE A/E 0.735; fewer than two periods with enough bets to compare (thin: the forward record will judge). O-6a4e2da9
2. AskedIn all three chase-debutant cuts, GB scored 1.27 to 1.86 and IRE 0.73 to 0.91. Is the split explained by handicap status or novice status, rather than by the country? answered
3. MechanismA horse beaten 20 lengths or more over hurdles has its rating dropped. When it first runs over fences in a handicap, it carries a mark set on form in a discipline it may not have suited. If the horse jumps fences better than hurdles, that mark is lenient. GB runs many novice handicap chases, so GB debutants would land in this setting more often than Irish ones.
4. Why the market misses itThe form figures show heavy recent defeats, and punters read them as a lack of ability. The mark looks fair on hurdle evidence, and nothing in the form shows how the horse will jump fences.
5. Predictedpositive, A/E 1.15 to 1.30 on 200+ bets, with Ireland's handicap subset at or above 1.0

The question's answer so far: Handicap status does not explain the split. Within handicaps GB scored 1.38 on 182 and IRE 0.37 on 43, and non-handicap debutants (mostly IRE) scored 1.22 on 110, level with the handicap group. Country and handicap status are nearly collinear in this population. Novice status is still untested (see Q-33ece0c3).

Mechanism and rationale

A horse beaten 20 lengths or more over hurdles has its rating dropped. When it first runs over fences in a handicap, it carries a mark set on form in a discipline it may not have suited. If the horse jumps fences better than hurdles, that mark is lenient. GB runs many novice handicap chases, so GB debutants would land in this setting more often than Irish ones.

The rule

race_type = 'Chase' AND COALESCE(chase_runs, 0) = 0 AND last_race_type = 'Hurdle' AND last_beaten_lengths IS NOT NULL AND last_beaten_lengths >= 20 AND is_handicap = 1

Reads: race_type, chase_runs, last_race_type, last_beaten_lengths, is_handicap

Pre-registration

Spec hash (sha256 of the canonical pre-spec)

6e0bc9e50e888b27f282eebc2cece0274fffb59594442daa6bed8e3d854ca86a

Frozen

not frozen

The definition can still change.

Confounders

• Region mix: GB stages far more handicap chases, so a pooled lift could still be the GB effect already seen. The IRE refinement settles this.
• Novice handicap status: the lift may come from novice handicap chase conditions (small fields, weak opposition) rather than the mark itself.
• Long-shot concentration: earlier cuts drew their winners from 16/1 or bigger. A few long-priced winners can inflate A/E, so the sub-17 refinement is checked.
• Period and season: late-period summer jumping has a different race mix and weaker fields.
• Trainer: a few yards that place chase debutants well off lowered marks could drive the result, though an earlier yard-skill rival did not explain the debut lift.
• Size of the rating drop: the rule does not read the official rating change, so the 'lenient mark' is inferred from the beating, not measured.

Control

{
  "description": "Same-race control: the other runners in these handicap chases should score about 1.0 (0.95 to 1.02). If they also score high, the lift belongs to the races (for example novice handicap chases or a summer race mix), not to the debutants. Mechanism control: the same debutant profile in non-handicap chases (novice, beginners' and graded races), where no handicap mark applies. The lenient-mark mechanism predicts these score clearly below the handicap group, near 1.0. If the non-handicap group matches or beats the handicap group, handicap status does not carry the lift, and the rival fails even if the main rule passes.",
  "rule": "race_type = 'Chase' AND COALESCE(chase_runs, 0) = 0 AND last_race_type = 'Hurdle' AND last_beaten_lengths IS NOT NULL AND last_beaten_lengths >= 20 AND is_handicap = 0"
}

Discovery result

pass broad rule (niche) · 17 Oct, 03:02

Bets

225

26 winners

A/E

1.25

95%: 0.81 to 1.83

SP return

+3.5%

level stakes

Betfair SP return

+25.5%

222 bets with a BSP

Strike rate 11.6%. p = 0.15. Same-race control: the other runners in these races had A/E 0.98, so the selection beat its own races. Zone: discovery. Evaluator eval-1.0, data 6a4e98be18e83933

Falsification

Each attack is a deliberate attempt to kill the idea. 4 of 5 survived.

• perturbation: Perturb beating threshold down to 16 lengthspass

  273 bets · A/E 1.11 (0.75 to 1.60) · control 0.99 · 17 Oct, 03:09

  Breakdown

  betsA/E0.71.01.3

  2025-10-01 to 2026-04-271581.25

  after 2026-04-271150.94

• perturbation: Perturb beating threshold up to 25 lengthspass

  182 bets · A/E 1.29 (0.79 to 2.00) · control 0.97 · 17 Oct, 03:09

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.

Other experiments

• neighbour: >= 20 -> 16 (-20%)pass

  273 bets · A/E 1.11 (0.75 to 1.60) · control 0.99 · 17 Oct, 03:02

  Breakdown

  betsA/E0.71.01.3

  2025-10-01 to 2026-04-271581.25

  after 2026-04-271150.94

• neighbour: >= 20 -> 24 (+20%)pass

  189 bets · A/E 1.21 (0.74 to 1.87) · control 0.98 · 17 Oct, 03:02

  Breakdown

Lineage

Agent runs

• designer v1ok · 32s · 17 Oct, 02:59

  Output (schema-checked)

  {
    "rule": "race_type = 'Chase' AND COALESCE(chase_runs, 0) = 0 AND last_race_type = 'Hurdle' AND last_beaten_lengths IS NOT NULL AND last_beaten_lengths >= 20 AND is_handicap = 1",
    "refinements": [
      {
        "label": "IRE handicap debutants (the decisive test of the rival)",
        "rule": "race_type = 'Chase' AND COALESCE(chase_runs, 0) = 0 AND last_race_type = 'Hurdle' AND last_beaten_lengths IS NOT NULL AND last_beaten_lengths >= 20 AND is_handicap = 1 AND region = 'IRE'"
      },
      {
        "label": "GB handicap debutants (checks that the GB lift survives once handicap status is fixed)",
        "rule": "race_type = 'Chase' AND COALESCE(chase_runs, 0) = 0 AND last_race_type = 'Hurdle' AND last_beaten_lengths IS NOT NULL AND last_beaten_lengths >= 20 AND is_handicap = 1 AND region = 'GB'"
      },
      {
        "label": "Shorter than 16/1 (a lenient mark should show up at backable prices, not only through a few long-shot winners)",
        "rule": "race_type = 'Chase' AND COALESCE(chase_runs, 0) = 0 AND last_race_type = 'Hurdle' AND last_beaten_lengths IS NOT NULL AND last_beaten_lengths >= 20 AND is_handicap = 1 AND sp_decimal < 17"
      }
    ],
    "control": {
      "description": "Same-race control: the other runners in these handicap chases should score about 1.0 (0.95 to 1.02). If they also score high, the lift belongs to the races (for example novice handicap chases or a summer race mix), not to the debutants. Mechanism control: the same debutant profile in non-handicap chases (novice, beginners' and graded races), where no handicap mark applies. The lenient-mark mechanism predicts these score clearly below the handicap group, near 1.0. If the non-handicap group matches or beats the handicap group, handicap status does not carry the lift, and the rival fails even if the main rule passes.",
      "rule": "race_type = 'Chase' AND COALESCE(chase_runs, 0) = 0 AND last_race_type = 'Hurdle' AND last_beaten_lengths IS NOT NULL AND last_beaten_lengths >= 20 AND is_handicap = 0"
    },
    "pass_criteria": {
      "min_bets": 150,
      "min_ae": 1.05,
      "max_p": 0.05
    },
    "forward_bar": {
      "min_bets": 100,
      "min_ae": 1.05,
      "max_p": 0.2
    },
    "confounders": [
      "Region mix: GB stages far more handicap chases, so a pooled lift could still be the GB effect already seen. The IRE refinement settles this.",
      "Novice handicap status: the lift may come from novice handicap chase conditions (small fields, weak opposition) rather than the mark itself.",
      "Long-shot concentration: earlier cuts drew their winners from 16/1 or bigger. A few long-priced winners can inflate A/E, so the sub-17 refinement is checked.",
      "Period and season: late-period summer jumping has a different race mix and weaker fields.",
      "Trainer: a few yards that place chase debutants well off lowered marks could drive the result, though an earlier yard-skill rival did not explain the debut lift.",
      "Size of the rating drop: the rule does not read the official rating change, so the 'lenient mark' is inferred from the beating, not measured."
    ],
    "mechanism": "A horse beaten 20 lengths or more over hurdles has its official rating dropped. On its first run over fences in a handicap, it carries a mark built only on hurdle form, in a discipline it may not have suited. Horses that jump fences better than hurdles then run off a lenient mark. The edge needs a handicap: in non-handicap chases no mark applies, so the effect should be weaker or absent there. Because the effect works through the mark and not the country, it should appear in Irish handicaps as well as GB ones. GB's higher share of novice handicap chases would explain why the earlier pooled debutant lift looked GB-only.",
    "market_miss": "Punters see heavy recent defeats in the form figures and read them as a lack of ability. The mark looks fair on the hurdle evidence. Nothing in the visible form shows how the horse will take to fences, so the market prices the hurdle form and not the chase potential the dropped mark gives away.",
    "prediction": {
      "direction": "positive",
      "size": "A/E 1.15 to 1.30 on 200+ bets; IRE handicap subset at or above 1.0; non-handicap control clearly lower (near 1.0)",
      "falsified_if": "Any of: broad rule A/E below 1.05; IRE handicap debutants below 0.95; non-handicap control at or above the handicap group. Each would mean country, not handicap status, carries the split, or that handicap status is irrelevant."
    },
    "notes": "I kept the rule sketch and added an explicit check that last_beaten_lengths IS NOT NULL. Pull-ups and other non-finishers are left out on purpose, as in the earlier no-pull-up debutant cut (1.24 on 335). The country refinements are not data-mining: the IRE subset is the main falsification test the hypothesis names, and GB is its pair. I raised the forward bar's A/E to 1.05 because an A/E of 1.0 cannot tell a lenient-mark edge from no edge. Ireland may produce only a small sample, so read the IRE refinement mainly by direction (at least 0.95, ideally at least 1.0), not by significance."
  }

Timeline

1. Job 2575 finished

   17 Oct, 03:09 · engine:compute-1

2. H-cc72aea8 Falsification → KILLED falsified: IRE A/E 0.367

   17 Oct, 03:09 · engine:falsification

3. falsification result for H-cc72aea8 (partition: Partition: non-novice handicaps only (strip novice-handicap race conditions)): pass

   17 Oct, 03:09 · engine:evaluator

4. compute-1 took job 2575 (attack, attempt 1)

   17 Oct, 03:09 · compute-1

5. Job 2574 finished

   17 Oct, 03:09 · engine:compute-1

6. falsification result for H-cc72aea8 (control: Ingredient control: handicap chase debutants NOT heavily beaten last time): fail

   17 Oct, 03:09 · engine:evaluator

7. compute-1 took job 2574 (attack, attempt 1)

   17 Oct, 03:09 · compute-1

8. Job 2573 finished

   17 Oct, 03:09 · engine:compute-1

9. falsification result for H-cc72aea8 (control: Mechanism control: same debutant profile in non-handicap chases): pass

   17 Oct, 03:09 · engine:evaluator

10. compute-1 took job 2573 (attack, attempt 1)

    17 Oct, 03:09 · compute-1

11. Job 2572 finished

    17 Oct, 03:09 · engine:compute-1

12. falsification result for H-cc72aea8 (perturbation: Perturb beating threshold up to 25 lengths): pass

    17 Oct, 03:09 · engine:evaluator

13. compute-1 took job 2572 (attack, attempt 1)

    17 Oct, 03:09 · compute-1

14. Job 2571 finished

    17 Oct, 03:09 · engine:compute-1

15. falsification result for H-cc72aea8 (perturbation: Perturb beating threshold down to 16 lengths): pass

    17 Oct, 03:09 · engine:evaluator

16. compute-1 took job 2571 (attack, attempt 1)

    17 Oct, 03:09 · compute-1

17. Job 2556 finished

    17 Oct, 03:08 · engine:ai-1

18. Queued attack on the compute lane: attack: Partition: non-novice handicaps only (strip novice-handicap race conditions)

    17 Oct, 03:08 · engine:scheduler

19. Queued attack on the compute lane: attack: Ingredient control: handicap chase debutants NOT heavily beaten last time

    17 Oct, 03:08 · engine:scheduler