The Model
Physics, Not Opinion.
Most tools show you results. StrideOdds models the horse — then flags when the market has it wrong.
The Physics-First Algorithm
The Physics-First algorithm calculates true win probability from the physical output of the animal — stride velocity, energy distribution, and stride length regularity — not just raw finishing positions.
This captures form signals traditional speed figure models routinely miss: a horse that ran third while closing faster than the winner, or a front-runner that led on borrowed time at an unsustainable pace fraction. When the internal fair-odds line diverges from the tote board, that gap is a potential overlay.
Three Signals the Market Rarely Prices In
GMI, PSS, and CDD are the edge layer of the model — inputs that public handicappers overlook and that the tote board almost never reflects accurately. Unique combinations of these variables are rarely available or utilized in different balanced combinations to set odds, which is precisely where StrideOdds finds its edge.
Ground Moisture Index (GMI)
Quantifies how wet the ground actually is — not just the official going description — and matches it against each horse's wet-track record. Official going ratings lag reality. GMI measures it directly.
Pace Shape Score (PSS)
Maps the running styles of every horse in the field to predict which runners get squeezed by pace pressure and which ones benefit. When multiple front-runners are drawn together, PSS flags the closers the market undervalues.
Condition Differential Delta (CDD)
Measures how far today's conditions deviate from a horse's historically optimal environment, and prices in the gap. A horse that excels on fast ground drawn into a soft-track race carries a CDD penalty the morning line rarely captures.
Stride Biometrics vs Speed Figures
Speed figures tell you how fast a horse ran. Stride biometrics tell you how the horse ran — and whether that performance is repeatable.
| Signal | Speed Figures | Stride Biometrics |
|---|---|---|
| What is measured | Final time, normalized | Velocity + stride length per furlong |
| Effort vs. output | No distinction | Detects horses running within themselves |
| Surface switch | Unreliable comparison | Biometric signature transfers across surfaces |
| Regression signal | Not captured | Flags max-effort runs likely to decline |
Live Tote Intelligence
Professional syndicates move money late, compressing odds on horses they consider value plays. StrideOdds detects statistically significant late money and uses it as a confirmation signal.
When the model flags a horse as an overlay and professional money enters the pool simultaneously, the Confidence Score for that runner increases significantly — giving you a dual-source edge signal before the gate opens.
Real-Time Environment Modeling
A strong inside speed bias inflates the odds of horses drawn wide — creating overlays for anyone who detects it early. StrideOdds builds a live bias model card-by-card, updating fair-odds lines for every subsequent race as the bias signal strengthens.
Beyond bias, Matrix Modifiers factor in real-time environmental data — temperature, wind velocity, humidity, and precipitation. A track rated "good to soft" at morning inspection can shift meaningfully by post time. The model recalculates automatically.
Sub-150ms Signal Latency
In live betting markets, latency is edge. StrideOdds operates at under 150ms from tote data change to on-screen signal — the full pipeline: data ingestion, fair-odds recalculation, Confidence Score update, and UI flag.
This matters most in the final 90 seconds before the gate opens, when professional money enters the pools, weather conditions finalize, and last-minute scratches cascade through exacta and trifecta tickets.
Frequently Asked Questions
What is the Physics-First algorithm in horse racing?+
The Physics-First algorithm calculates a horse's true win probability starting from the physical output of the animal — stride velocity, stride length regularity, and energy distribution — rather than relying solely on finishing positions. This captures form signals that traditional speed figure models miss, such as a horse that ran third while closing faster than the winner.
What are GMI, PSS, and CDD?+
GMI (Ground Moisture Index), PSS (Pace Shape Score), and CDD (Condition Differential Delta) are StrideOdds' three proprietary signals. They quantify going conditions, pace dynamics, and environmental fit respectively — inputs the public tote board almost never prices in accurately.
How fast does StrideOdds update its odds signals?+
StrideOdds operates at under 150ms from tote data change to on-screen signal. The full modeling pipeline — data ingestion, fair-odds recalculation, Confidence Score update, and UI flag — runs in a single low-latency loop, giving users a meaningful speed advantage over manual odds watchers.
How does StrideOdds adjust for track bias and weather in real time?+
StrideOdds builds a live bias model card-by-card, updating fair-odds lines for every subsequent race as the bias signal strengthens. Matrix Modifiers factor in real-time environmental data — temperature, wind velocity, humidity, and precipitation — and automatically recalculate each horse's probability when conditions change.
Ready to use the model?
See the algorithm surface live mispriced horses — free for 7 days.
Start free trial