Skip to content

Reference

Frequently Asked Questions

How Racetrackstar builds a race record, calibrates its probabilities, simulates pace, grades its plays, and supports race day on Pro.

Information, not a guarantee, and not an instruction to wager. If wagering stops being recreational, use the responsible gambling resources.

Research and Calibration

What does the aggregation add to a standard past performance?

A past performance shows the lines. Racetrackstar also reads the chart detail behind them: early position, late gain, trip trouble, and wide trips across recent starts. Race-day changes, workout patterns, weather toward post, and sire aptitude for a surface or distance switch join the same record. A missing input adds nothing. It is not estimated.

What is the research lab?

The research lab is where model changes are tested before they reach a published card. It works from the factor ledger and the play ledger, never from hand-picked races. A change that improves the fit on past races but not on later ones does not ship.

How does a rating become a win probability?

A calibration fitted to recorded finishes converts each rating into a chance, and the field sums to about 100 percent. The goal is that runners published at 20 percent win about one race in five over a large sample. The method and its parameters stay private.

How does a factor gain or lose weight?

Through recorded results, not opinion. Each factor's contribution is measured on later races and adjusted in small steps when the evidence supports it. A factor that looks strong at one meet and weak across the rest is treated as noise until the sample says otherwise.

Why is the model not published?

The factor list and the weights are the product of the research. Publishing them would invite copies and would encourage second-guessing a single race on a single factor. The outputs and every graded result are public, and those are the parts a horseplayer can verify.

Pace Simulation

What is the pace shape?

A call on how the early part of the race should be run, based on how many runners want the lead. Each starter is styled from recent trips as early speed, presser, mid-pack, or closer. The field then reads as Lone Speed, Favors Front, Honest Pace, Duel, Fast Pace, or No Speed.

What does the pace projection show?

Projected running positions at three calls: leaving the gate, around the turn, and the top of the stretch. Each runner starts where its style puts it, then gains or loses ground according to the pace shape, the distance, and its post. A sprint compresses the late moves. A route stretches them.

What does the pace simulation add to a read?

Context the ranking cannot show. Two runners can share a rating while one projects to an uncontested lead and the other to a late run into a slow pace. A Duel or Fast Pace marks a race where the speed may tire. Lone Speed marks one where the leader may never be caught. Each runner also carries a tag showing whether its style suits the shape.

Does the pace simulation change the win probabilities?

No. The simulation sits beside the ranking and does not feed it, so win probabilities and system plays are unchanged. When too few runners have a usable style, no shape is called and no projection is drawn. A horse without enough history is listed as not placed.

Ledgers and Grading

What do the two ledgers record?

The factor ledger records every start with the factor scores held at decision time. The play ledger records every system play with the price pinned when the decision was made. Neither ledger is rebuilt from closing odds, and losing periods are never removed.

Which price is a play graded at?

The price held when the decision was made. The closing price is not available at that moment, so grading against it would flatter the record. The gap between the two, known as closing-line value, is tracked separately as a check on the read.

What is the model measured against?

The public pool. The tote already prices the opinion of every bettor in the race, so beating a random pick proves little. The lab compares model probabilities with market probabilities on the same races, by track and surface, and reports a result only when the race count clears a minimum.

Pro Tools

What does Pro change on race day?

The free morning card is frozen at publication. Pro follows the race after that point: live odds, scratches, and cancellations through post, plus a cross-track queue in post-time order. The Breakdown Lab and the stake tools sit on the same screen.

What does the Breakdown Lab show?

The component scores that combine into each runner's rating, such as speed, form, class, connections, and workouts. The reason a horse ranks where it does is visible at a glance. The weights that combine the components remain private.

What do the Lab sliders change?

They reweight the components for one race and show your order next to the published one. The published card, its win probabilities, the queue, and Stake Assist stay on the system read. The sliders test a view. They do not override the model.

How do Circuit Pad and Stake Assist size a play?

Both convert a bankroll you enter into a stake for each play, scaled to the size of the edge the model sees. Circuit Pad plans a full track card from the queue. Stake Assist plans the race you are reading. The figure is a plan, and copying it never places a wager.