RaceWatch Score (RWS) explained: what 0 to 100 means

What does a RaceWatch Score of 80 tell me about a runner?

A score of 80 means the model rates the runner strongly on the data it reads. What it means in practice depends on the field around it, which is why the rank in race matters.

The RaceWatch Score is a machine learning rating on a 0 to 100 scale, given to every runner. The model reads each runner's data and produces one number. High numbers mean the data points strongly towards the runner, low numbers mean it does not.

An 80 in a small field where one horse stands clear is a different thing from an 80 in a big handicap where several runners score in the seventies. That is why the RWS Hub shows the rank in race beside each score, and why the rank is the place to start.

Why does the rank in race matter more than the raw score?

Rank in race orders the runners within their own field, which removes the differences in race quality and field size that make raw scores hard to compare between races.

Think of the score as the distance and the rank as the finishing order. Two races can both have a top ranked runner, but one leads its field clearly while the other is barely ahead of the next horse. Reading the gap between first and second rank tells you how clear the model's view is.

The RWS Hub also shows the key factors behind a runner's score and a confidence gap, so you can see whether a high rank rests on one strong signal or several.

What does the RWS model read?

The model reads each runner's data and turns it into one score, and the RWS Hub lists the key factors behind each score so you can see what drove it.

Think of the key factors as the model showing its working. A careful form student weighs many things about every runner, and the model's job is to weigh its inputs the same way, runner after runner, without the shortcuts and blind spots that creep in when a person reads a hundred races a day.

Like any model built from past data, it is better at spotting runners the data favours than at things the data cannot see, such as a horse's well being on the day, a late change of plan or how the race will actually be run.

When are RaceWatch Scores produced each day?

Scores are produced each morning and then locked, so the racecard and the RWS Hub show the same fixed scores for the rest of the day.

A fixed score has a practical advantage: it gives one reading of each race that does not shift as the day goes on, so you can set it against the prices as they move and see where the two agree or disagree.

Because the scores are fixed in the morning, non runners and going changes that come later are not built into them, so check the racecard for withdrawals and ground changes before leaning on a score. A runner whose main rival has been withdrawn, or whose ground has turned against it since the morning, can look very different by the time its race comes round.

Is the RWS a win probability, a price or a tip?

The RWS is none of those. It ranks runners by the model's reading of their data, and a top ranked runner still loses often, so the score belongs beside form and the market.

A win probability is a chance of winning, and the RWS is a model score rather than a chance. A price reflects a bookmaker's margin and the weight of money. A tip is a recommendation to act, and the RWS is a research aid that tells you where the data points.

The practical meaning is that a high score is a reason to study a runner closely, not a conclusion. Two runners on similar scores can be priced very differently, because the market is weighing things the model does not see, or because one of them is simply a better price than the other.

How should the RWS be used alongside form, draw and the market?

Use the rank to decide which runners deserve a closer look, then check the Form, Draw and Odds tabs to see whether conditions, stall and price support the model's reading.

A sensible routine is to read the top two or three ranked runners first, then look for reasons the model might be wrong: an unsuitable trip, a bad draw, ground the horse has never handled or a long absence. If the reasons stack up against a high score, the score deserves less weight.

Then look at the Odds tab. A top ranked runner at a short price tells you the market agrees; a top ranked runner at a longer price is worth understanding before anything else. Either way, the decision is yours, and the score is one input to it.

• Read the rank before the raw number
• Check the gap to the next ranked runner
• Look for conditions the model cannot see
• Compare with the Odds tab
• Check non runners and going changes after the lock

Keep reading