Threads · trajectory · H85-fell-when-going-well

H85: fell when going well

Hypotheses: 20 (12 alive)
Depth: 5
Best rung: Live
Open questions: 11
Last activity: 17 Oct, 04:10

Where it stands: H-ec26713e (Test-mechanism: chase unseaters that shortened or held their last SP a) pointed the predicted way but failed the bar: A/E 0.95 on 103 bets (p 0.603), control 1.01.

The tree

PROMISINGA/E 1.30 · 165H85: fell when going wellimportgreen
- LIVESystem: Fell or unseated when going wellimport
- SHADOWA/E 1.08 · 270Rival: chase fallers/unseaters NOT going well show the same liftrival-explanationamber
  - SHADOWA/E 1.26 · 190Test-mechanism: unseated (not fell) chasers not going well carry the general F/U lifttest-mechanismgreen
    - SHADOW

Still open

Is the going-well chase edge a long-price effect? Do going-well fallers at 16/1+ beat their price while non-going-well F/U at 16/1+ do not (0 from 98)? open
Should the live chase system be read GB-only in its forward record, given two independent IRE cuts at or below 0.92? open
Why does the unseat lift hold in IRE (1.37 on 87) when the going-well edge fails there (0.92 on 53)? open
Is the bounce-back after a chase unseat general across both countries and the going-well split? testing
How can UR-based chase rules get a sealed test when the vault predates last_non_finish_code? Should the lab reserve a post-2025 vault slice, or rely only on the forward record? open
Is the 16/1+ hurdle unseater band (A/E 2.38 on 126) a sign that the hidden-run mechanism crosses codes at long prices, or just 4 lucky winners? open
Why is a large drift (at least 1.5x last SP, A/E 1.80 on 76) weaker than any drift (2.10 on 110)? Do big drifts reflect a class rise or a stronger, larger field rather than the U discount? open
Why does the drift split point the other way for fallers: does market support after a fall (shortening, especially to 3/1 or shorter) carry information while a drift does not? open
Will the frozen UR rules (H-f5df804d, H-f8d6b9d7, H-a44d6bdc) beat a forward bar? And what bet count should be set before the drift split counts as replicated rather than a forking-path result? open
How much do the well-fancied, drifting, under-16/1 and jockey cuts of chase unseaters overlap? Should the lab count them as one pool when reading forward results?

The notebook

In order: what the lab saw, what it asked, what it proposed and why, what the tests said, and what an investigator concluded. Numbers are aggregates on the discovery data unless a step says otherwise.

Hypothesised15 Oct, 23:10
H85: fell when going well PROMISING green
because Fell or unseated last time while on the front end or travelling well, before coming under pressure.
```
lt_fell_going_well = 1
```
made by import
Tested15 Oct, 23:10
broad: A/E 1.13 on 405 bets, by period 1.15 · 1.11. info
Betfair SP return +36.5%. Every period above 1.00 (spread 0.04). Zone: all. On

Where it started

Fell or unseated last time while on the front end or travelling well, before coming under pressure.

H85: fell when going well

Hypotheses

20 (12 alive)

Depth

Best rung

Live

Open questions

Last activity

17 Oct, 04:10