Understanding the Ratings

A short guide to the two headline figures we publish for every horse — the Performance Rating (PR) and the Speed Rating (SR) — and the thinking behind them.

Both ratings are built from raw race data — weights carried, finishing margins, race class, and the clock — on our own scale, our own assessment of each performance. The settled figures, calculated after a race has been run, are derived purely from that data. The pre-race Performance Rating you see on a racecard is a forecast — built from the horse's own form and then tweaked by our horse racing expert AI agent, which refines the figure with its own research and analysis (more on that below). The Speed Rating is always a pure clock-based figure.

The two ratings answer two different questions:

  • Performance Rating — how good was the performance?
  • Speed Rating — how fast did the horse actually run?

A horse can score well on one and not the other, and that difference is often where the interesting reads are. More on that below.

Performance Rating (PR)

What it is: a merit figure expressed in pounds (lb) — the same currency as a handicap mark. A higher number means a better performance. As a rough guide, modest maiden and low-grade handicap form sits in the lower bands, while Pattern-class performances rate near the top.

What drives it:

  1. Weight carried. Everything starts from the weight on the horse's back. Giving weight away and still running well counts for more.
  2. Beaten lengths — distance-sensitive. Margins are converted into pounds, but a length doesn't always mean the same thing. Over a sprint, a length is worth more pounds than over a marathon, because ground is much harder to make up at five furlongs than over two miles. Our scale reflects that.
  3. A weight-for-age allowance. Younger horses are still physically maturing, so they receive an allowance when racing against older rivals — larger early in the season and tapering as the year goes on. This stops a progressive three-year-old being unfairly marked down for carrying less weight than an established older horse.
  4. An anchor to the class of the race. Once every runner's merit is worked out, the race as a whole is pegged to the standard you'd expect for its class. This is what keeps the numbers meaningful across the card: a strong race rates higher than a weak one, even if the winning margins look similar.

A worked example (illustrative)

In a one-mile race, the winner carries 9st 4lb (130 lb). A rival carries the same weight but is beaten two lengths. If a length is worth two pounds at this trip, the runner-up's raw merit is 130 − (2 × 2) = 126 lb. Now suppose a third horse is also beaten two lengths but was carrying 10 lb more — it gave weight away and still finished alongside the runner-up, so it earns a higher figure. The whole field is then shifted so the race sits at the level expected for its class.

How to read it: PR is your form number. It rewards horses that have run well relative to the weight they carried and the company they kept. It is the figure to lean on when you're asking “which of these has the best level of form?”

Two versions of the PR. After a race, the PR is the settled merit figure described above — purely the raw data. Before a race, on a racecard, the PR is a projection of how the horse is likely to perform: we build it from the horse's recent form, and our horse racing expert AI agent then tweaks that figure, refining it with its own research and analysis.

When the pre-race figures appear. Both pre-race ratings first become available once a race's runners are declared, as form-based projections. The agent's tweak to the Performance Rating is then applied the evening before that day's racing — so the figure you see on the eve of a meeting reflects the agent's research, while a rating that appears further out is the form projection on its own, awaiting that refinement. The after-the-race figure stays a clean, independent merit number.

Speed Rating (SR)

What it is: a clock-only figure. Crucially, it ignores where a horse finished and what it carried — it is purely about the time. A front-runner who is collared late and a closer who flies home can post very similar SR figures if they ran the same clock.

What drives it:

  1. A course-and-distance benchmark. For each track and trip we build a standard “par” time from years of results, so we know what a sound, representative winning time looks like there. A horse's raw time is measured against that benchmark.
  2. How the track was running that day. Going and track conditions make whole cards collectively fast or slow. We measure that day-and-meeting effect from the races themselves and strip it out, so a genuinely fast time on testing ground isn't undersold, and an easy time on lightning-fast ground isn't overrated.
  3. Distance sensitivity. As with beaten lengths, a second off the clock is worth more over a sprint than over a staying trip, and the scale reflects that.

A worked example (illustrative)

The benchmark time for a one-mile race at a course is 100.0s. A horse runs it in 98.0s on a day when the track is running to standard. That two seconds faster than par is converted into rating points and added to the baseline, producing an above-average speed figure. If the meeting as a whole had been running slow, we'd first credit the horse for the slow conditions before making the comparison.

How to read it: SR is your raw speed number. It is most informative on truly-run races and strong galloping performances, and least informative in muddling, tactical races where nobody goes a true gallop (everybody's clock is slow, so everybody's SR is modest).

Using PR and SR together

Because they measure different things, the gap between a horse's two figures is often the story:

  • High PR, modest SR — a smart, race-fit performance that won on merit, but in a steadily-run or tactical race. The form is solid; don't read too much into the slow clock.
  • Modest PR, high SR — the horse didn't win the battle but ran a fast time, often doing its best work without a clear run or against the bias of the race. One to note, especially if conditions next time might suit better.
  • High on both — the strongest profile: good form and a fast time. Hard to argue with.

Neither rating is “the right one.” They are two lenses on the same race, and looking at both gives a fuller picture than either alone.

How we keep the ratings honest

  • Independent settled figures. The after-the-race ratings are derived purely from raw data — times, weights, margins and class — and are never tuned or fitted to match anyone else's ratings; they are our own assessment of what happened. The pre-race Performance Rating is then tweaked by our horse racing expert AI agent, which refines it with its own research and analysis, so a racecard PR reflects both the horse's own form and the agent's read. (The Speed Rating stays a pure clock figure, pre-race and settled alike.)
  • No hindsight on a racecard. When a rating appears on a racecard, it reflects only what was known before that race — built from the horse's prior runs. We deliberately keep that pre-race figure separate from the settled, after-the-fact assessment of a past race, so a racecard never benefits from information that wasn't available at the time. Nothing leaks backwards from the future into a horse's pre-race number.

These ratings are an assessment, not a guarantee. They are a tool to inform your own judgement — best used alongside everything else you'd weigh up before a race.