TrackSmart AI: Cumulative Performance Report - NYRA & Regional Tracks
Reporting Period: December 26, 2025 – May 17, 2026
Track: Belmont at the Big A (BAQ) / Finger Lakes (FL) / Aqueduct (AQU) / Laurel Park (LRL)
Executive Summary: May 17 Update
- Volume Expansion: The tracking dataset has grown to 689 total races with the addition of 79 races run across Belmont at the Big A, Finger Lakes, and the newly integrated Laurel Park from May 11 through May
- Strike Rate Resilience: The cumulative Top 4 Strike Rate held consistently strong at 78.5% (541 for 689). The model achieved perfect 100% accuracy cards at Finger Lakes (May 11) and Belmont at the Big A (May 14), while absorbing standard variance and chaos results on heavy multi-track days later in the week.
- Value Normalization: The Average Payout for AI hits rebounded slightly to ~$7.62 (up from ~$7.55). Mechanical adjustments to off-the-pace turf closers and successful automated scratch bumps helped the algorithm isolate fantastic value hits this week, including Voodoo Doll ($18.18) at BAQ and Warming ($19.80) at
- System Integrity and Updates: The expansion to Laurel Park exposed a critical data ingestion gap concerning Also-Eligible (AE) and late program Meanwhile, the mechanical scratch adjustment protocols executed flawlessly across all tracks, saving substantial exotic value when primary pace factors scratched.
Daily Performance Breakdown
|
Date |
Track |
Top 4 Strike Rate |
Total Races |
1st Pick |
2nd Pick |
3rd Pick |
4th Pick |
Avg Payout (AI Hits) |
|
Prev. Totals |
MULTI |
78.9% |
610 |
181 |
130 |
111 |
59 |
~$7.55 |
|
May 11 |
FL |
100.0% |
8 |
4 |
2 |
2 |
0 |
$7.72 |
|
May 12 |
FL |
87.5% |
8 |
1 |
4 |
0 |
2 |
$8.58 |
|
May 14 |
BAQ |
100.0% |
8 |
1 |
2 |
3 |
2 |
$6.01 |
|
May 15 |
BAQ |
50.0% |
8 |
2 |
1 |
0 |
1 |
$7.22 |
|
May 15 |
LRL |
57.1% |
14 |
4 |
1 |
3 |
0 |
$5.38 |
|
May 16 |
BAQ |
81.8% |
11 |
2 |
3 |
4 |
0 |
$9.60 |
|
May 16 |
LRL |
71.4% |
14 |
2 |
2 |
5 |
1 |
$11.02 |
|
May 17 |
BAQ |
75.0% |
8 |
2 |
0 |
2 |
2 |
$8.69 |
|
NEW TOTALS |
MULTI |
78.5% |
689 |
199 |
145 |
130 |
67 |
~$7.62 |
(Note: Prior totals reflect cumulative data up to May 10. Daily breakdowns from previous reports have been condensed into the Prev. Totals row. Live scratch protocols dynamically shifted runners up in rankings across multiple races this week, securing numerous top-tier hits.)
Winner Distribution by Rank (Cumulative)
|
AI Rank |
Wins |
% of Hits |
Trend Note |
|
Top Pick (#1) |
199 |
36.8% |
Pace Flow Precision: Maintained consistency, accurately controlling speed evaluations on standard dirt formulas and capitalizing on logical overlays and class drops. |
|
2nd Selection |
145 |
26.8% |
Class Elevation & Adaptability: Showed immense strength this week. Continues to capitalize on class plungers and tactical stalkers when the primary pace setup softens or front-runners fold. |
|
3rd Selection |
130 |
24.0% |
Alternate Protection: Heavily relied upon this week as a primary safety net. Automated scratch adjustments mechanically moved alternates directly into the 3rd rank for massive upset hits. |
|
4th Selection |
67 |
12.4% |
Rebounding Safety Net: Provided massive value consistency, accurately catching chaotic pace-meltdown beneficiaries and keeping long-layoff survivors on the fringe of multi-race exotics. |
- Cumulative Statistics Summary
|
Metric Stat Description |
||
|
Overall Accuracy |
78.5% |
(541/689 Races) The AI has successfully placed the winner in its Top 4 selections in 541 of the 689 recorded races. |
|
Top 2 Accuracy |
49.9% |
(344/689 Races) The primary or secondary choice remains highly reliable, winning in nearly half of all modeled races. |
|
Avg Payout |
~$7.62 |
The average payout trended up slightly this week. Excellent value retention was achieved by leveraging secondary fringe assessments when primary speed scratched. |
Aggregate Observations: May 11 - May 17 Insights
- Scratch Protocol Successfully Executed: Automated contingency rules remained flawless and were entirely critical to retaining value. Recalculating the rankings gracefully elevated fringe horses to fill top spots following late scratches, capturing massive hits like Sequential ($23.00), Dreams of Rome ($10.46), and Night Jasmine ($4.12) that would have otherwise fallen out of the matrix.
- Data Ingestion Vulnerability: The integration of Laurel Park highlighted an immediate blind spot regarding raw data ingestion for Also-Eligibles (AE) and late program The AI completely missed winners like Bold Fact ($9.00) and Golden Descent ($14.00) because they were not parsed into the original TPN field summaries. A feed pipeline patch is an immediate priority.
- Unexposed "Blue Sky" Limits: The algorithm showed systemic vulnerability when evaluating unexposed, lightly raced 3-year-olds facing either experienced horses or returning for their second starts (e.g., Grant the Great, Sadie Earp). The AI's rigid baseline speed pars left no room for natural developmental "Blue Sky" and age/experience penalty algorithms need aggressive rebalancing.
- Soft Pace vs. Lone Speed Dynamics: While the AI correctly predicted destructive "High Chaos/Meltdowns" perfectly this week (e.g., Tapizar's Temper at FL), it struggled on turf routes and sprints where the pace was unexpectedly The model’s form-cycle penalizations caused it to heavily downgrade early speed, missing gate-to-wire threats (e.g., Coach Mazzula, Chips and Fish) that dawdled on uncontested leads. Early Speed (E1/E2) metrics on the turf need positive modifiers when a lack of challengers is present.
