Dose-response: young handicap chasers at 3m to 3m3f show a smaller lift

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
   • H175-young-horse-in-a-marathon-chase (H175: young horse in a marathon chase) was GREEN in the earlier research: A/E 1.16 on 618 bets (periods 1.17 / 1.21 / 1.05); now PROMISING. O-f49a1279
2. AskedDoes the edge fall away as the trip shortens (a dose-response from 3m to 3m4f+), as a stamina-doubt mechanism predicts? answered
3. MechanismThe market's doubt about a young horse's stamina should grow with distance. At 3m to 3m3f a young horse should still be slightly underrated, but less than at 3m4f+.
4. Why the market misses itThe 'young horses don't stay' heuristic applies more weakly at standard staying trips, so the mispricing should be smaller but still present.
5. Predictedpositive, A/E 1.03 to 1.10 on 800+ bets, below the 1.18 seen at 3m4f+
6. Falsified ifA/E at or above the 3m4f+ figure (no gradient), or at or below 0.97
7. Rule sketchrace_type = 'Chase' AND is_handicap = 1 AND age <= 7 AND distance_yards >= 5280 AND distance_yards < 6160
8. question id

Mechanism and rationale

The market's doubt about a young horse's stamina should grow with distance. At 3m to 3m3f a young horse should still be slightly underrated, but less than at 3m4f+.

The rule

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

Reads: race_type, is_handicap, age, distance_yards

Pre-registration

Spec hash (sha256 of the canonical pre-spec)

65699dc514246b7920994ade66ab69608820fd4ef538465a2d37b4a34b7de340

Frozen

not frozen

The definition can still change.

Confounders

• Novice handicaps: young chasers are often novices with few chase runs and handicap marks that have not been tested. Any lift may come from unexposed marks, not stamina doubt (tested by the is_novice = 0 refinement).
• Race class: young horses at 3m-3m3f may sit unevenly in lower or higher classes, and the A/E of each class differs.
• Field size: big-field staying handicaps create more value at long prices and noisier A/E.
• Period decay: the young-stayer lift at 3m4f+ faded in the latest period (1.17, 1.21, then 1.05), so a lift in this band should be checked by period, not just pooled.
• Seasonal or spring concentration: other threads show lifts tied to one early-spring window, so check whether the excess comes from March-April.
• Price band: the excess may sit at 16/1+ from a handful of wins. Read A/E by price band (the 14/1 refinement).
• Region: an IRE/GB split in opposite directions should not be read as mechanism.
• Trainer concentration: a few yards that place young staying chasers well could produce the effect.

Control

Discovery result

fail broad rule · 16 Oct, 13:46

Bets

971

134 winners

A/E

0.92

95%: 0.77 to 1.09

SP return

-36.1%

level stakes

Betfair SP return

-26.1%

943 bets with a BSP

Strike rate 13.8%. p = 0.84. 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 · 32s · 16 Oct, 13:45

  Output (schema-checked)

  {
    "rule": "race_type = 'Chase' AND is_handicap = 1 AND age <= 7 AND distance_yards >= 5280 AND distance_yards < 6160",
    "refinements": [
      {
        "label": "non-novice handicaps only",
        "rule": "race_type = 'Chase' AND is_handicap = 1 AND age <= 7 AND distance_yards >= 5280 AND distance_yards < 6160 AND is_novice = 0"
      },
      {
        "label": "14/1 or shorter",
        "rule": "race_type = 'Chase' AND is_handicap = 1 AND age <= 7 AND distance_yards >= 5280 AND distance_yards < 6160 AND sp_decimal <= 15"
      }
    ],
    "control": {
      "description": "Same-race rivals (the lab's default) are mostly older horses in the same 3m-3m3f handicap chases; we expect them to show A/E about 0.97-1.0, slightly below 1 because the young runners take the excess. The key comparison is older horses (age 8+) in the same trip band of handicap chases. If they show the same lift as the young horses, the effect belongs to the race type or trip, not to age, and the stamina-doubt reading fails. The dose-response gradient is read against H175 (young horses at 3m4f+, A/E 1.16 on 618). That rule was not limited to handicaps, so a fair gradient check also needs the handicap-only 3m4f+ figure.",
      "rule": "race_type = 'Chase' AND is_handicap = 1 AND age >= 8 AND distance_yards >= 5280 AND distance_yards < 6160"
    },
    "pass_criteria": {
      "min_bets": 300,
      "min_ae": 1.05,
      "max_p": 0.05
    },
    "forward_bar": {
      "min_bets": 150,
      "min_ae": 1,
      "max_p": 0.2
    },
    "confounders": [
      "Novice handicaps: young chasers are often novices with few chase runs and handicap marks that have not been tested. Any lift may come from unexposed marks, not stamina doubt (tested by the is_novice = 0 refinement).",
      "Race class: young horses at 3m-3m3f may sit unevenly in lower or higher classes, and the A/E of each class differs.",
      "Field size: big-field staying handicaps create more value at long prices and noisier A/E.",
      "Period decay: the young-stayer lift at 3m4f+ faded in the latest period (1.17, 1.21, then 1.05), so a lift in this band should be checked by period, not just pooled.",
      "Seasonal or spring concentration: other threads show lifts tied to one early-spring window, so check whether the excess comes from March-April.",
      "Price band: the excess may sit at 16/1+ from a handful of wins. Read A/E by price band (the 14/1 refinement).",
      "Region: an IRE/GB split in opposite directions should not be read as mechanism.",
      "Trainer concentration: a few yards that place young staying chasers well could produce the effect."
    ],
    "mechanism": "The market discounts a young chaser's (age 7 or under) chance at staying trips because the horse has seldom been tested at the distance. If the young-stayer lift at 3m4f+ (H175, A/E 1.16) comes from this stamina doubt, the doubt and the mispricing it causes should grow with distance. At standard staying trips of 3m to 3m3f in handicap chases, young horses should still be slightly underrated, but less than at 3m4f+, giving a dose-response gradient across trip.",
    "market_miss": "Bettors and layers apply a blunt 'young horses don't stay' heuristic instead of judging each horse's stamina from its pedigree and running style. At 3m-3m3f the heuristic bites more weakly, because many young chasers have already run at around 3m, so the mispricing should be smaller than at 3m4f+ but still present.",
    "prediction": {
      "direction": "positive",
      "size": "A/E 1.03 to 1.10 on 800+ bets: clearly below the 1.16 seen at 3m4f+ but above 1.0",
      "falsified_if": "A/E at or below 0.97 (no residual underrating at standard trips), or A/E at or above 1.16 (no gradient: the lift does not shrink with shorter trip, so stamina doubt is not the driver). Either is also falsified if the age 8+ control in the same band shows a lift as large as the young horses'."
    },
    "notes": "The rule sketch is kept unchanged: 5280 yd = 3m, and 6160 yd = 3m4f, the H175 threshold, so the two bands meet without overlap. All columns are pre-off. The predicted size (1.03-1.10) straddles the discovery floor of A/E 1.05. A correct dose-response result could therefore fail discovery and still support the mechanism; it should be read together with the H175 comparison and go to falsification and forward tracking, not be treated as a plain kill. min_bets is raised to 300 because the predicted effect is small. The refinements come from the reasoning: the novice-handicap confounder, and the lab finding that edges sit in particular price bands."
  }

Timeline

1. Job 1679 finished

   16 Oct, 13:46 · engine:compute-1

2. H-910e15e3 Discovery → KILLED fails: A/E 0.923, needs 1.05; p 0.835, needs 0.05 or lower; periods [0.92, 0.93], each needs 1.0; A/E 0.923 against same-race control 1.099, needs 0.03 above (not a niche pass: A/E 0.923 below 1.05; p 0.835019035378531 above 0.5; does not beat the same-race control 1.099; a period points the other way [0.92, 0.93])

   16 Oct, 13:46 · engine:evaluator

3. discovery result for H-910e15e3 (broad rule): fail

   16 Oct, 13:46 · engine:evaluator

4. H-910e15e3 Designed → RESEARCHdiscovery measuring on the discovery zone

   16 Oct, 13:46 · engine:orchestrator

5. compute-1 took job 1679 (evaluate, attempt 1)

   16 Oct, 13:46 · compute-1

6. Job 1676 finished

   16 Oct, 13:46 · engine:ai-1

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

   16 Oct, 13:46 · engine:scheduler

8. H-910e15e3 Proposed → RESEARCHdesigned pre-spec written and hashed

   16 Oct, 13:46 · agent:designer@v1

9. ai-1 took job 1676 (design, attempt 1)

   16 Oct, 13:45 · ai-1

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

    16 Oct, 13:44 · engine:scheduler

11. New idea H-910e15e3: Dose-response: young handicap chasers at 3m to 3m3f show a smaller lift (generalise)

    16 Oct, 13:44 · 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.