Weekly Performance Scorecard

NYRA & Regional Tracks — TrackSmart AI Cumulative Analysis

TrackSmart AI | Prepared for Saratogabets | Week Ending September 6, 2026

1. Overview & Methodology

This report covers the cumulative performance of TrackSmart AI's handicapping model across NYRA and regional tracks for the reporting period of December 26, 2025 through September 6, 2026, spanning 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), and Kentucky Downs (KD).

The model's primary performance benchmark is the Top 4 Strike Rate — the percentage of races in which the eventual winner appeared among the AI's top four selections. This week's update adds 109 races run across Finger Lakes, Saratoga, and Kentucky Downs between August 31 and September 6, bringing the season-to-date dataset to 1,894 total races.

2. This Week at a Glance

Volume Expansion

The tracking dataset grew to 1,894 total races this week, with 109 new races added across Finger Lakes, Saratoga, and Kentucky Downs between August 31 and September 6.

Strike Rate Resilience

The cumulative Top 4 Strike Rate held steady at 77.1% (1,461 for 1,894), essentially flat against last week's 77.2%. The model turned in flawless 100% cards on both August 31 and September 1 at Finger Lakes, then closed the week with an explosive 91.7% strike rate (11 for 12) at the notoriously chaotic Kentucky Downs meet on September 6.

Value Normalization

The Average Payout for AI hits ticked upward to roughly $8.00, with deep-value identification continuing to drive outsized exotic returns. Notable overlays captured this week include Diablo Shiraz ($55.00), Kupuna ($26.22), Perrylu ($19.40), Carmen Amalia ($19.32), Jean Valjean ($18.08), Rum Twist ($17.64), and Bless the Broken ($19.26).

System Integrity & Scratch Protocol

The automated scratch adjustment protocol was the standout performer of the week. Post-scratch step-ups repeatedly moved winners into the Top 4 actionable slots, salvaging what would otherwise have been Fringe-tier misses — most notably Diablo Shiraz ($55.00), Perrylu ($19.40), Glycogen ($10.32), and Nyikos ($7.42).

3. Daily Performance Breakdown

The table below details the week's card-by-card results, with all prior reporting condensed into the Prev. Totals row and the bottom row reflecting the new cumulative totals through September 6.

Note: Prior totals reflect cumulative data through August 30. Daily breakdowns from previous reports have been condensed into the Prev. Totals row. Late scratches and altered race flows dynamically shifted runners up in rankings across all cards this week, contributing to the strike rate's exceptional stability.

4. Winner Distribution by Rank (Cumulative)

Top Pick and 2nd Selection combined account for 63.7% of all Top 4 hits, while the 3rd and 4th slots combine for the remaining 36.3% — a reminder of how much of the model's back-half performance still depends on the scratch-adjustment protocol and deep-dive value overrides.

5. Cumulative Statistics Summary

6. This Week's Structural Findings

The Next-Man-Up Doctrine

Adhering to the rule of dynamically moving alternates into the Top 4 upon scratches proved to be the bedrock of this week's profitability. Mechanically sliding backups into vacated spots salvaged highly-priced tickets on nearly every card, capturing direct winners like Nyikos ($7.42) and Glycogen ($10.32), and the colossal $55.00 bomb Diablo Shiraz at Kentucky Downs.

The First-Time-Starter Blind Spot

A recurring pattern continued in 2-year-old maiden events, particularly on the turf. Leaning too heavily on published morning workouts, or penalizing outside post draws, repeatedly blinded the model to raw on-track ability and elite sire pedigree — missing Slickback at $58.18 and Pharocity at $10.66. Wide variance bands and pedigree overrides remain necessary for turf first-time starters.

Cracking the Kentucky Downs Code

The model was initially challenged by the quirky Kentucky Downs layout but adapted quickly as the week progressed. By correctly identifying that tactical inside trackers holding a garden-spot trip carried a massive advantage on the temporary 25-foot rail, the AI fully dissected the September 6 card for a 91.7% hit rate (11 for 12). Elite turf pedigrees, such as Twirling Candy's, also consistently outperformed raw incoming speed figures.

Reading the Pace Script

The model thrived whenever it could isolate the race's pace flow. It seamlessly adapted to sloppy and sealed tracks at Finger Lakes, producing back-to-back 100% days, and correctly diagnosed pace meltdowns when claiming-level horses engaged in early speed duels. The AI reliably separated true "Alpha Controls" speed-holding scripts from "Turf/LP Priority" closing setups.

The September 4 Data Gap

A data mapping and parsing flaw on September 4 (Race 10) substituted a horse from another race into the card, completely masking a $65.06 longshot winner. Stricter data validation and gate-checking protocols are needed to guarantee correct horse-to-race alignment before the finalized sheets are run.

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

1. Keep the scratch-adjustment protocol running exactly as configured — it was the single biggest driver of this week's profitability and should not be tightened or second-guessed.
2. Build a dedicated turf-FTS override that widens variance bands and weights pedigree class more heavily for first-time turf starters, rather than defaulting to workout times and post position.
3. Carry forward the Kentucky Downs inside-trip logic (garden-spot trackers on the 25-foot rail) as a standing adjustment for future visits to quirky, temporary-rail courses.
4. Continue trusting the pace-flow separation between "Alpha Controls" and "Turf/LP Priority" scripts — it is producing the model's most consistent top-pick hits.
5. Add a mandatory horse-to-race alignment check before finalizing sheets, closing the gap that let the September 4 Race 10 error slip through.

8. Data Notes & Caveats

• Reporting period reflects cumulative data from December 26, 2025 through September 6, 2026; the Prev. Totals row condenses all prior weekly reports into a single cumulative line.
• This week's additions cover 109 races run at Finger Lakes, Saratoga, and Kentucky Downs between August 31 and September 6, 2026.
• “% of Hits” in the Winner Distribution table is calculated against total Top 4 hits (1,461), not total races run (1,894).
• The September 4 (Race 10) data mapping error affected only that race's record; it does not change the cumulative Top 4 Strike Rate reported above.