TrackSmart AI Weekly Report Card

Cumulative Performance: NYRA & Regional Tracks

TrackSmart AI | Prepared for Saratogabets | Week Ending November 8, 2026

1. Overview & Methodology

This report card tracks every race TrackSmart AI has handicapped across the NYRA circuit and the regional tracks in the Saratogabets coverage footprint, from December 26, 2025 through November 8, 2026. The cumulative dataset covers Belmont at the Big A (BAQ), Finger Lakes (FL), Aqueduct (AQU), Laurel Park (LRL), Santa Anita Park (SA), Penn National (PEN), Saratoga Harness (SARH), Woodbine (WO), Saratoga (SAR), Del Mar (DMR), Hollywood Casino at Charles Town Races (CT), Kentucky Downs (KD) and Keeneland (KEE).

The headline metric is the Top 4 Strike Rate: the share of races in which the eventual winner was one of the AI’s four ranked selections. Each hit is also credited to the rank the winner held, from 1st through 4th pick, and the average payout on those hits is tracked alongside it so accuracy can be read against price.

This week added 85 races run from November 2 through November 8 at Finger Lakes, Belmont and Del Mar, bringing the dataset to 2,780 races.

2. This Week at a Glance

Volume Expansion

The 85 new races came from three Finger Lakes cards (November 2–4), four Belmont cards (November 5–8) and three Del Mar cards (November 6–8). The dataset now stands at 2,780 total races.

Strike Rate Resilience

The cumulative Top 4 Strike Rate held steady, moving from 77.3% to 77.2% (2,146 for 2,780). The week itself landed the winner in the Top 4 in 65 of 85 races (76.5%), right in line with the season-long mark. Two cards were perfect: Finger Lakes on November 2 (8-for-8) and Belmont on November 8 (8-for-8). Del Mar closed the week at 88.9% on November 8, after its low point of the week at 44.4% on November 7.

Value Normalization

Average payout on AI hits moved up from ~$8.18 to ~$8.22. Value identification and favorable scratch dynamics stood out this week, with the top prices on the board including Honey in the Bank ($43.90), Fuhgeddaboudit ($37.48), Souper Lawrence ($33.62), Babe Ruthless ($26.80) and Rose Lisa ($21.32).

System Integrity & Scratch Protocol

The automated scratch adjustment was again the most lucrative feature of the week. Moving “Next-Man-Up” alternates into active slots protected win rates and lifted unheralded runners directly into winning positions, most notably Honey in the Bank ($43.90), Flying Hawley ($14.62) and Classic Commander ($9.84) when original Top 4 selections scratched.

3. Daily Performance Breakdown

Prior totals reflect cumulative data through November 1; earlier daily breakdowns are condensed into the Prev. Totals row. Late scratches shifted runners up in the rankings across all cards this week. October 24 Keeneland rank totals include a Dead Heat between the 1st and 2nd AI choices, yielding 2,147 total hits for 2,146 successful races.

4. Winner Distribution by Rank (Cumulative)

The rank order continues to behave the way a ranking should: wins step down from each position to the next, and the top two selections account for 63.7% of all credited hits. This week’s 65 winners split 22, 19, 17 and 7 from 1st pick through 4th pick. The 4th selection produced the fewest hits but the two biggest prices listed for the week, which is the job that slot exists to do.

5. Cumulative Statistics Summary

6. This Week’s Structural Findings

The Next-Man-Up Cascade

The mechanical rule of sliding alternates up after late scratches paid off again. The automated cascade caught Honey in the Bank ($43.90), Flying Hawley ($14.62) and Classic Commander ($9.84). Strict adherence to updating the board for scratches remains mandatory.

Off the Turf and Onto Tapeta

Inclement weather forced surface shifts at Belmont on November 6 and 7, moving multiple turf races to the All-Weather (Tapeta) track. Those were also Belmont’s two lowest cards of the week, at 62.5% and 60.0%. The AI struggled slightly with horses carrying strict “no turf experience” penalties, such as Souper Lawrence ($33.62) and Kid’s Khanclusion. An automated Off-the-Turf recalculation that shifts ratings dynamically will close this gap.

Class Moves and Layoffs

The form-cycle logic was slightly too rigid on class hikes and layoffs this week. Horses dropping in class off a layoff, such as Major Bourbon ($19.96) and Determind Princess ($10.24), and horses stepping up from the maiden claiming ranks, such as Babe Ruthless ($26.80), won despite heavy AI penalties. A slight “re-set” buffer for these conditions will help keep live runners on the board.

When the Market Disagrees With the Figures

In several races, including Sunny Creed’s on November 6, the AI’s raw numbers strongly disagreed with a short Morning Line, leading to logical misses. A “respect the money” modifier for Morning Line favorites taking a surface switch or steep class drop could help catch winners that historical speed figures undervalue.

Crowded Speed and the West Coast Turf Read

The engine was exceptionally precise at projecting “Crowded Early Speed” battles, isolating stalkers inheriting a “Garden Spot” in November 3 Races 1 and 8 and November 8 Races 2 and 8. Front-end turf pressure metrics on the West Coast may need slight softening, however, as projected late-runner biases failed to materialize when early speed went uncontested.

7. AI’s Final Analysis — Building the Path Forward

This week’s results point to five practical refinements, each aimed at a specific situation the cards exposed.

1. Re-rate races that come off the turf. When a race moves to the All-Weather track, recalculate the field for the surface it will actually be run on, so a turf-experience penalty does not follow a horse onto Tapeta.
2. Add a re-set buffer for form-cycle changes. Class droppers returning from a layoff and runners moving up from maiden claiming ranks should get a small allowance instead of a full penalty, keeping them in contention for a Top 4 slot.
3. Give the market a narrow voice. Apply a “respect the money” modifier only to Morning Line favorites taking a surface switch or a steep class drop, the two situations where past figures say the least about today’s race.
4. Require contested speed before projecting a closer’s race on West Coast turf. Softening front-end pressure metrics there keeps a late-runner bias from being projected when no one is likely to press the pace.
5. Keep the scratch protocol non-negotiable. Next-Man-Up accounted for three of the week’s featured winners, and it only works when the board is updated for every late scratch.

8. Data Notes & Caveats

• The reporting period runs from December 26, 2025 through November 8, 2026. Daily detail is shown for November 2–8 only; all earlier cards are condensed into the Prev. Totals row.
• The Top 4 Strike Rate measures whether the winner was among four ranked selections. It is not a win rate for the top pick, which accounts for 801 wins in 2,780 races.
• Rankings reflect late-scratch adjustments, so a horse can be credited in a Top 4 slot after moving up from an alternate position.
• Rank columns total 2,147 hits against 2,146 successful races because of the October 24 Keeneland Dead Heat. The % of Hits column is based on the 2,147 figure.
• Average payout is calculated on AI hits only and is shown as an approximate cumulative figure. It is not a return-on-investment measure.
• The weekly figure of 65 for 85 is the sum of the ten daily rows in Section 3.