Open any race card and you can list a dozen things that matter: speed, pace, class, form, trainer, jockey, going, days off, draw, weight. Now ask which of them tells you something in this race, on this ground, in this field. The answer usually turns into a shrug.
That is the practical problem. There are too many factors, and the ones that help in a Class 4 chase at Sedgefield are not the ones that help in a Listed sprint at Newmarket. The skill is less about knowing every factor and more about knowing which ones earn their place.
Which horse racing factors actually predict winners? Plenty of them pick winners. The harder test is whether a factor picks winners at a price that pays. Across 13 years of GB and Irish racing, the Betfair favourite, the top of our VDW ratings and the top of our Race Predictor all landed within a few percent of a fair price. A factor that finds winners but is already priced in has not found you anything the market missed. So choose a few factors, check each one against the odds in races like the one in front of you, and expect most of them to fail that check.
Why most factors are really categories
When you say you look at speed, you have chosen a category, not a factor. Inside it sit several different questions: recent speed, average speed over the trip, best speed on similar going, peak speed, whether the horse is trending up or down. Each has its own answer.
The same goes for class, pace, form and connections. Trainer form is a category. The number of winners that trainer has sent out in the last fourteen days, in this type of race, is a factor. Treat the shorthand as a starting point, not the analysis.
That is also why looking at everything tends not to work. You are not trying to weigh hundreds of factors. You are trying to find a few that suit the conditions and make each one justify itself.
The ranking systems worth comparing first
Before hunting for bespoke factors, it helps to see how the big ranking systems behave on the average race. These produce a top-rated horse in almost every race:
- Betfair SP favourite, the market’s own verdict at the Betfair starting price.
- Race Predictor top horse, the number one in our model, which turns several factors into a win chance for each runner.
- VDW Rank 1, the top of VDW, one of our own in-house rating systems.
- Official Rating, the handicapper’s number, which is mainly useful inside handicaps because weights are set against it.
Here is how the first three have performed across GB and Irish racing from 1 July 2013 to 14 October 2026, at Betfair SP, level stakes, after 2% commission, with fallers and pulled-up horses counted as losers. They describe that stretch of racing and will not repeat exactly.
| Signal | Races | Win rate | A/E | Return on stakes |
|---|---|---|---|---|
| Betfair SP favourite | 167,257 | 34.4% | 1.00 | -1.4% |
| VDW Rank 1 | 153,920 | 21.6% | 0.99 | -3.8% |
| Race Predictor top horse | 167,314 | 25.5% | 1.01 | 0.0% |
A/E means actual wins divided by the wins the odds imply, so a figure near 1.00 says the price was about right. Our A/E guide explains it in plain English.
The win rates are very different, from 21.6% to 34.4%, yet the A/E barely moves. In this sample the price had already taken the rating advantage into account. Our PR Rank 1, the top of our own PR rating, is also the Betfair favourite in 81% of races, so it tells the same story as the favourite and is not listed separately.
Agreement is a confidence tool, not a value tool
It is tempting to trust a horse more when two ratings agree. Agreement does raise the win rate, because you are narrowing to horses that many measures like. But those are also the horses the market likes, so the price shortens in step. Rating agreement tells you which horses are likely to run well. It does not tell you whether the price is wrong.
Use agreement as a way to build a shortlist. Do not treat it as proof of value. Our guide to combining ratings goes deeper on this.
How to choose factors for the race in front of you
The overall figures tell you which factors are worth shortlisting. The race in front of you tells you which to use. This is a way of working, not a tested system, so judge it by your own records.
- Pin down the conditions. Race type, course, distance, class, going and field size. These define what similar past races means.
- Pull a sample of matching races. Aim for at least a few hundred runners. If the filter is too tight, widen distance or class a notch. If it is too loose, add field size or course. As a working preference, treat samples under about 1,000 runners as suggestive only, because small samples swing a lot. That is a rule of thumb, not a hard line.
- Look at contenders, not just winners. A horse beaten a short head after being short of room ran a contender’s race. A horse that won because the favourite was hampered did not. Contenders can give a fairer view of whether a factor was doing its job, but deciding who was a contender is a judgement made after the result, and hindsight bends those calls. Use it as a supplementary check, not a replacement for the price-based record, and write down your reason for each call before you look at the outcome.
- Check each factor against the price. Win rate is the easy view. A/E (actual wins against the wins the odds imply) at Betfair SP is the more honest one. If a factor’s top horse sits near 1.00 in your slice, the price has already caught up with it there.
- Keep the working set small. A short list you understand is easier to check than a long one. A sensible target is around ten factors, though that number is a rule of thumb and not a result from the data.
Treat anything you find as a hypothesis. A factor that looks good in one slice of past races may be there by chance, particularly if you tried many slices before finding it. The cure is to record what it does on races you have not yet seen.
The honest read
Choosing factors is not a hunt for the one true rating. It is a habit. You start with the ranking systems that are easy to compare, check how they behave in the kind of race you care about, and keep only the ones that hold up against the price.
On the figures above, the market is hard to beat with any single obvious factor, so expect most of your ideas to come back at about a fair price. The ones that do not are the ones worth a longer look, and a longer test.
If you want somewhere to run the comparison, open today’s race cards and put the favourite, the Race Predictor top horse and the VDW ratings side by side for a race or two. See where they agree, where they split, and which runners are left for your own form work.
