TrapMetrics
Try the demoAccess →View access options
HomeLearnA trap-bias percentage needs a second question

A trap-bias percentage needs a second question

··2 min read

What sits under a trap win percentage: the sample, the window, the fair share and whether the pattern survives a period that was not used to find it.

trap biastrack geometryevidence

Ask what is underneath the percentage

A headline trap win percentage can look precise while telling you surprisingly little. Before interpreting it, look for the number of runners and races, the dates included, and the distance being measured.

A result drawn from a small sample can move sharply after a few more races. A large sample can still answer the wrong question if it combines conditions that are not relevant to yours, such as a sprint trip and a middle-distance trip at the same track.

Measure against the fair share, not one in six

The benchmark matters as much as the number. Not every race fills all six traps, so a trap that is often empty has fewer runners and a smaller fair share of the wins. The honest comparison is each trap's share of wins against its share of runs.

At Romford over the last two years, trap 1 won 19.0% of races against a fair share of 17.3%: +1.8 points on 7,048 runs, which is a clear bias. Trap 6 sat 1.3 points under its fair share, which is inside noise. Both numbers look similar in size. Only one of them is a finding.

A draw statistic is not a causal explanation

A high winning percentage from a trap does not, by itself, prove that the trap caused the wins. The runners drawn there may have been stronger, or the period may contain other differences, such as a change to the running rail.

Treat the percentage as a reason to investigate, not as a selection rule.

Make the comparison relevant, then check it again

Check whether the track, distance, period and race population match the question you want to ask. Write the filter down before you look at the result. The more variations you inspect, the easier it becomes to find a pattern that is only chance.

Then ask whether the pattern persists in a period that was not used to find it. If it disappears, that is informative too.

Bring it back to the race

Track information is one part of the picture. Read it with the runner's form and the rest of the field. Even a stable historical pattern does not decide the next race. Every track page on this site carries its race count, its window and a clear-or-noise read for each trap, so you can see how much a number is worth before you use it.

Questions this piece answers

Why is one in six the wrong benchmark for a trap?

Because not every race fills all six traps. A trap that is often vacant has fewer runners, so its fair share of wins is its share of the runs, not a sixth. Comparing it with a sixth invents a bias that is not there.

How many races does a trap statistic need?

More than most sites show. A difference of one or two percentage points needs thousands of runs to separate from noise. Every TrapMetrics track page says whether each trap's difference is clear or inside noise.

Does a trap bias mean I should back that trap?

No. A bias tells you where to look, not what to bet. The runners drawn there may be stronger, and the market already prices the obvious patterns.

T
The Edge AnalystGreyhound racing and statistical analysis

The analyst behind TrapMetrics: the person who built the rating model, runs it every morning, and writes up what it finds. About the writer

Also worth reading

  • How to read a model's record without being dazzledWhen the prediction existed, what the denominator is, whether the chances are calibrated, and why winning more often is not the same as making money.
  • How we rate a raceThe unique elements behind a TrapMetrics race page, in plain English: point-in-time features, a calibrated chance, the TMR scale, measured going, track geometry, selection classes and the consensus forecast.
  • Trap bias at Central Park: which traps actually winTrap bias at Central Park measured on 5,627 races over 2 years: share of wins by trap against fair share, by distance and by grade, with the noise called noise.
Your next step

Find your place in the research.

Try the tools, read the evidence and choose the access that fits your questions.

View access optionsOpen a real race

Register before accounts open and Research is £19.99 a month, locked in.