How the TrackLab score works
A plain-English guide to what the number means and where it comes from.
Want more detail? The ⚙️ box at the bottom opens and closes. Open it if you're curious. Skip it if you're not.
📋 What the score is
Every horse in every race gets a score from 0% to 100%. It is our model's estimate of the horse's chance of winning. It is based on data, not gut feeling.
In one line: Our estimated chance that this horse wins this race. All runners in a race add up to 100%.
A score of 15% means we estimate roughly a 15% chance of winning. A score of 40% means about a 40% chance.
📝 Example
2
River King
Kempton Park · 2:30 PM · a jumps race where horses carry different weights
Our model gives this horse about a 28% chance of winning. That is a strong chance and the highest in this race. It can still lose.
40%+ Very strong chance
20–39% Strong chance
12–19% Solid chance
6–11% Could surprise
Under 6% Long shot
How scores workModel estimates, not guaranteed. Past performance is no guide.
💡 The key point
The score is a chance, not a rank. 28% means a strong chance, but the horse still loses more often than it wins. 8% means probably not, but upsets happen. Both are useful.
💎 Underrated horse
An underrated horse is one where we think its chance is better than its odds suggest. We compare our score with the bookmaker's odds. If our chance is higher than the odds imply, the horse gets a label.
At 5/1, £1 on brings back £6 (your £1 plus £5) if it wins.
There are two labels:
Very underrated horse
The odds look too long
Our chance is well above what the odds suggest.
Underrated horse
The odds look a little too long
Our chance is a bit above what the odds suggest.
No label means our chance roughly matches the odds, or is lower. A label is not a promise. An underrated horse can still lose.
🧪 What it looks at
📅 Form
How the horse has run lately, and whether it is improving or fading.
🏇 Jockey
How well the jockey has been doing lately, and how they get on with this horse.
📋 Trainer
The trainer's recent results and record with runners like this one.
🏟 Course
How the horse has done at this course and over this distance.
🌧 Ground
How soft or firm the ground is, and how the horse has run on it before.
🏷 Odds
What the bookmakers think. This is where we start.
🔧 How it's made
The score starts from bookmaker odds, which reflect what many punters think. Then it looks for spots where those odds may be wrong. So we agree with the bookmakers most of the time. We only mark a horse as underrated when the data points to a clear gap.
⚙️ Under the hood
⚙️ A little more detail
- It learned from years of past UK race results.
- It starts from the odds, then adjusts them.
- It is adjusted so that scores behave like real chances. A 20% score wins about 1 in 5 times.
- It holds back scores when key data is missing, rather than guess.