Young marathon chasers near the bottom of the weights: youth versus weight position

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-ebbea470 (Rival: light weight, not youth, carries the marathon handicap chase li) was predicted positive but came out the other way: A/E 0.48 on 124 bets (p 0.08 for the opposite direction). O-ccf039fe
2. AskedDo young (7 or under) marathon handicap chasers near the bottom of the weights still beat their price, which would separate youth from weight position? answered
3. MechanismIf the edge comes from youth (horses still improving that the market thinks won't stay), it should hold whatever weight they carry. Young light-weights are often unexposed horses on falling or early marks, not declining ones.
4. Why the market misses itA young horse near the bottom of the weights at 3m4f+ looks both unproven in stamina and short of class on its form, so it gets two marks against it and drifts. Older light-weights look 'well in' and are backed.
5. Predictedpositive, A/E 1.08 to 1.20 on 250+ bets
6. Falsified ifA/E at or below 1.0 on 200+ bets, which would mean the system's lift sits with young horses higher in the weights and supports the class rival
7. Rule sketchrace_type = 'Chase' AND is_handicap = 1 AND age <= 7 AND distance_yards >= 6160 AND weight_lbs <= bottom_weight_lbs + 7

Mechanism and rationale

If the edge comes from youth (horses still improving that the market thinks won't stay), it should hold whatever weight they carry. Young light-weights are often unexposed horses on falling or early marks, not declining ones.

The rule

race_type = 'Chase' AND is_handicap = 1 AND age <= 7 AND distance_yards >= 6160 AND weight_lbs <= bottom_weight_lbs + 7

Reads: race_type, is_handicap, age, distance_yards, weight_lbs, bottom_weight_lbs

Pre-registration

Spec hash (sha256 of the canonical pre-spec)

cd821fbfbfd1e8ea7a279fc122636c57e6bbfb13599c0c21b6a23aa0496983a5

Frozen

not frozen

The definition can still change.

Confounders

• Weight carried (weight_lbs) includes conditional/amateur jockey allowances, so some 'light-weights' are higher-rated horses ridden by claimers. Rider skill, not youth, may then drive the result.
• Horses running from out of the handicap (rated below the bottom-weight mark) are included and are usually outclassed. This would pull A/E down regardless of age.
• Big-field festival marathon handicaps (Grand National type races, Scottish/Welsh/Irish Nationals) probably hold most qualifying runners, so the result may reflect a few race types.
• Small sample at 3m4f+: young light-weights could fall below 150 bets, and a few big-price winners could dominate A/E.
• The parent young-stayer edge is weakening by period, so a positive result may be concentrated in the early years.

Control

{
  "description": "Two comparisons. (1) The lab's standard same-race control: the other runners in these races. These are mostly older and higher-weighted horses, so they should sit near par (about 0.95 to 1.10). Earlier work found heavier older horses in these races at 1.15 on 352, so a control above 1.10 would point to a race-level effect and not a youth effect. (2) The comparison that answers the question: young (7 or under) marathon handicap chasers more than 7 lb above bottom weight. If youth drives the edge, both weight groups should show a similar lift (each above 1.05). If the system's lift sits only with the higher-weighted young horses, the class/weight-position rival is supported and youth is not. A further reference is already measured: older horses within 7 lb of bottom weight in the same races scored 0.48 on 124, so young light-weights clearly above that also shows that age changes how light-weights perform.",
  "rule": "race_type = 'Chase' AND is_handicap = 1 AND age <= 7 AND distance_yards >= 6160 AND weight_lbs > bottom_weight_lbs + 7"
}

Discovery result

fail broad rule · 16 Oct, 14:05

Bets

37

4 winners

A/E

0.80

95%: 0.22 to 2.04

SP return

-1.0%

level stakes

Betfair SP return

+28.1%

37 bets with a BSP

Strike rate 10.8%. p = 0.74. Same-race control: the other runners in these races had A/E 1.07, 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 · 1m 12s · 16 Oct, 14:00

  Output (schema-checked)

  {
    "rule": "race_type = 'Chase' AND is_handicap = 1 AND age <= 7 AND distance_yards >= 6160 AND weight_lbs <= bottom_weight_lbs + 7",
    "refinements": [
      {
        "label": "14/1 or shorter",
        "rule": "race_type = 'Chase' AND is_handicap = 1 AND age <= 7 AND distance_yards >= 6160 AND weight_lbs <= bottom_weight_lbs + 7 AND sp_decimal <= 15"
      },
      {
        "label": "Unexposed over fences (6 or fewer chase runs)",
        "rule": "race_type = 'Chase' AND is_handicap = 1 AND age <= 7 AND distance_yards >= 6160 AND weight_lbs <= bottom_weight_lbs + 7 AND chase_runs <= 6"
      }
    ],
    "control": {
      "description": "Two comparisons. (1) The lab's standard same-race control: the other runners in these races. These are mostly older and higher-weighted horses, so they should sit near par (about 0.95 to 1.10). Earlier work found heavier older horses in these races at 1.15 on 352, so a control above 1.10 would point to a race-level effect and not a youth effect. (2) The comparison that answers the question: young (7 or under) marathon handicap chasers more than 7 lb above bottom weight. If youth drives the edge, both weight groups should show a similar lift (each above 1.05). If the system's lift sits only with the higher-weighted young horses, the class/weight-position rival is supported and youth is not. A further reference is already measured: older horses within 7 lb of bottom weight in the same races scored 0.48 on 124, so young light-weights clearly above that also shows that age changes how light-weights perform.",
      "rule": "race_type = 'Chase' AND is_handicap = 1 AND age <= 7 AND distance_yards >= 6160 AND weight_lbs > bottom_weight_lbs + 7"
    },
    "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": [
      "Weight carried (weight_lbs) includes conditional/amateur jockey allowances, so some 'light-weights' are higher-rated horses ridden by claimers. Rider skill, not youth, may then drive the result.",
      "Horses running from out of the handicap (rated below the bottom-weight mark) are included and are usually outclassed. This would pull A/E down regardless of age.",
      "Big-field festival marathon handicaps (Grand National type races, Scottish/Welsh/Irish Nationals) probably hold most qualifying runners, so the result may reflect a few race types.",
      "Small sample at 3m4f+: young light-weights could fall below 150 bets, and a few big-price winners could dominate A/E.",
      "The parent young-stayer edge is weakening by period, so a positive result may be concentrated in the early years."
    ],
    "mechanism": "Young chasers (7 or under) at 3m4f+ are still improving and their stamina is unproven. If the edge is truly about age (improvement plus untested stamina that the market underrates), it should not depend on handicap weight. Young horses near the bottom of the weights are often unexposed horses on early or falling marks, not declining horses dropping down the handicap, so they should beat their price as much as young horses higher in the weights do.",
    "market_miss": "A young horse near the bottom of the weights in a marathon chase gets two marks against it: stamina unproven at the trip, and apparently short of class on its rating. Punters lump it with the outclassed tail of the handicap and let it drift. Older light-weights are seen as 'well in' and are backed, yet they scored 0.48 on 124. The market gives weight position the same meaning at every age, though it means something different for an improving 6-year-old than for a 10-year-old sliding down the handicap.",
    "prediction": {
      "direction": "positive",
      "size": "A/E 1.08 to 1.20 on roughly 150 to 300 bets, similar to the parent young-stayer rule (1.16)",
      "falsified_if": "A/E at or below 1.0 on 150+ bets, or a clearly lower A/E than young runners more than 7 lb above bottom weight in the same races. Either would mean the young-stayer lift sits with higher-weighted young horses, which supports the class/weight-position rival over youth."
    },
    "notes": "I kept the rule sketch as written; all its columns are pre-off (weight_lbs is weight carried, bottom_weight_lbs is the lightest weight in the race). I set the falsification threshold at 150 bets, not the sketch's 200, because the subset may be thin. A thin result that points the right way goes to forward tracking. Refinements: (1) a 14/1 price cap, because earlier slices of the marathon lift came mainly from shorter prices and light-weights at 8/1+ have been poor. (2) 6 or fewer chase runs, which tests the 'unexposed horse on an early mark' part of the mechanism directly. I did not add a jockey-allowance refinement because no claim-allowance column was confirmed; the allowance problem is listed as a confounder instead."
  }

Timeline

1. Job 1732 finished

   16 Oct, 14:05 · engine:compute-1

2. H-0464cf15 Discovery → KILLED fails: 37 bets, needs 150; A/E 0.796, needs 1.05; p 0.7382, needs 0.05 or lower; periods [], each needs 1.0; A/E 0.796 against same-race control 1.074, needs 0.03 above (not a niche pass: A/E 0.796 below 1.05; p 0.738187844342748 above 0.5; does not beat the same-race control 1.074)

   16 Oct, 14:05 · engine:evaluator

3. discovery result for H-0464cf15 (broad rule): fail

   16 Oct, 14:05 · engine:evaluator

4. H-0464cf15 Designed → RESEARCHdiscovery measuring on the discovery zone

   16 Oct, 14:05 · engine:orchestrator

5. compute-1 took job 1732 (evaluate, attempt 1)

   16 Oct, 14:05 · compute-1

6. Job 1730 finished

   16 Oct, 14:02 · engine:ai-1

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

   16 Oct, 14:02 · engine:scheduler

8. H-0464cf15 Proposed → RESEARCHdesigned pre-spec written and hashed

   16 Oct, 14:02 · agent:designer@v1

9. ai-1 took job 1730 (design, attempt 1)

   16 Oct, 14:00 · ai-1

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

    16 Oct, 14:00 · engine:scheduler

11. New idea H-0464cf15: Young marathon chasers near the bottom of the weights: youth versus weight position (test-mechanism)

    16 Oct, 14:00 · 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.