How does the Fhor.site Points Engine calculate individual horse selections?

The engine starts each horse at a baseline score of 50 points, adjusting up or down based on Fhorsite master ratings, 14-day trainer metrics, raw speed figures, class shifts, track variants, and draw biases.

What is the difference between an EW Simple pick and an EW Edge pick?

An EW Simple pick selects the highest-scoring runner with forecast odds of 5/1 or bigger. An EW Edge pick uses a strict multi-layered mathematical matrix requiring a top-5 model rank and a market-to-model rank discrepancy of at least 2 full places.

Fhor.site Points Engine: Full Calculation Methodology

We do not guess, and we do not rely on “gut feelings”. Every single automated selection displayed on our site comes from a rigorous data pipeline called our Points Engine. It uses the exact same algorithmic logic we use to backtest years of historical racing data to ensure our strategy actually works.

Here is a simplified look at how our code goes from a raw race card to our final Win, Place, and Each-Way (EW) selections.

Phase 1: Filtering the Field

Before any math happens, the system looks at every horse entered in a race. It instantly cross-references our live databases to flag and remove any non-runners. Only active, confirmed horses move forward to the scoring stage.

Phase 2: The 0–100 Model Score

Every eligible horse starts with a baseline score of 50 points. The code then adjusts this score up or down based on live performance data:

Fhorsite Rating (FSr) & Reliability (FSRR): An FSr over 70 adds points. If our rating reliability score (FSRR) is over 50, it triggers an extra confidence bonus.
Recent Speed: We average out the horse’s last three speed figures (SR_LTO, SR_2, and SR_3) against a standard benchmark baseline of 70 to add or subtract points from their score.
Draw Bias (Flat Only): If a horse is drawn in a starting stall that historically has a winning advantage at that specific track, it gets a helper bonus. This is turned off entirely for National Hunt jump races.
Trainer & Jockey Form: We check recent strike rates. A trainer winning more than 12% over the last 14 days, a trainer with a 10%+ course record, or a strong trainer/jockey combo all add points.
Fine-Tuning Penalties & Bonuses: The engine penalizes horses off the track for over 35 days or those stepping up in class. It awards bonuses for positive Official Rating (OR) shifts, track suitability (Course, Distance, Going wins), beaten favourites, and 2-year-old maturity parameters.

Note: After all weights are applied, the final horse score is strictly capped between 0 and 100.

Phase 3: Finding the Market "Edge"

The real profit in racing comes from finding horses that the bookmakers have underestimated. To calculate our advantage score, the engine performs three rapid steps:

Converts the bookies’ forecast odds into pure market probability.
Compares our internal Model Rank against the bookmakers’ Market Rank (where the favorite is #1).
Calculates the final discrepancy. If our model likes a horse significantly more than the public does, a massive edge value is created.

Phase 4: The Four Final Picks

With all calculations complete, the engine instantly publishes four distinct selection types inside the blue Points Engine Picks box on your race card:

🏆 Win Pick: The single highest-scoring, active horse in the race based strictly on data talent, not public odds.
🥈 Place Picks: The top three highest-scoring horses in the race. (Note: This is strictly our model’s top three talent rating, not a projection of bookmaker each-way place terms).
💰 EW Simple: Selects the highest-rated active horse with forecast odds of 5/1 (6.0 decimal) or bigger.
⚡ EW Edge: Our premium value filter. To trigger an EW Edge selection, a horse must have forecast odds of 5/1+, rank in the top 5 of our internal model scores, and outrank the bookmakers’ placement by at least 2 clear spots with a strong mathematical edge.

Same Engine as Backtests: The exact same core algorithm handles our historical strategy testing and our live cards. The only difference is that live cards pull upcoming data blocks, while backtests access historical tables.
Access Rules: Guest accounts or search crawlers may see restricted access to the main picks box until logged in, though the backend processing logic remains identical.

Summary in One Sentence: Win is our top performance score; Place covers our top three scores; EW Simple finds our highest-rated horse at 5/1+; and EW Edge locks in our best “model vs. market” advantage at 5/1+ using strict ranking filters—all computed after evaluating form, speed, and track data while throwing out non-runners.