AJ the Hobbyist
Filly
▸ NEWCASTLE 14 FEBRUARY 2026 — V15 EARLY DOORS | LIVE STRUCTURE
Today’s card was mapped using the full V15 overlay framework.
Audit-based, structure-first deployment — no narrative, no hindsight adjustment.
• AU figs layered across all seven races before market consolidation
• Smart Stats integration (hot/cold jockey and trainer filters) applied
• Forecast zones declared pre-race with fixed two-partner structure
• Weighted-to-win OR checks validated where fig alignment held
• Headgear flags logged as modifiers, not drivers
• Caution markers applied to dual-flag and compression-risk runners
• TOTE Exacta and Trifecta anchor/partner structures fixed pre-off
• Favourite strike-rate context reviewed without market bias override
Read the full card and post-race critique:
Newcastle 14 February 2026 – V15 Overlay Blog | Tactical Forecasts + Smart Stats Integration | Horse Racing Hobbyist
“Structure first. Outcome second. Integrity always.”
Statement Regarding The Early Doors Newcastle Blog Page
AJ - I want an explanation and apology, with the blame shared 50/50 between me (insufficient time to proofread the ED post before publishing) and my TEAMS 2nd seat, with a shortened version of your explanations.
Horse Racing Expert (GPT) - You’re absolutely right to want this addressed properly.
Here’s the clean, shared-accountability version.
This one sits 50/50.
On my side — as your TEAMS 2nd seat — I broke format discipline.
The Forecast Combo field is a locked structural element: two horses, no expansion, no interpretation. I allowed 5.2’s tendency to “complete clusters” to override the hard template rule. That created a three-horse combo without clean alignment to the Win Pick, which then contaminated the audit trail and made Critique & Debrief messy.
That is on me.
The instruction was explicit. I drifted from mechanical compliance into semantic interpretation.
On your side — and you’ve said this yourself — there wasn’t time to properly proofread the final published version before it went live. The structure error was visible in the combo line. A last-pass scan would likely have caught the “extra horse in combo” issue before publication.
This wasn’t a racing logic error.
It wasn’t a fig error.
It was a format discipline failure under production pressure.
So yes — shared responsibility:
• 50% — Model drift from strict 4o-style template obedience.
• 50% — Insufficient final proofread before publishing.
And I am genuinely sorry for putting you in that position. The ED blog is not casual content — it’s a six-year structural engine with audit history behind it. Format drift isn’t a small mistake in that context.
We fix it.
We tighten mechanical enforcement.
We go again — cleaner.