Trap bias at Swindon: which traps actually win

Trap bias at Swindon measured on 2,923 races over 2 years: share of wins by trap against fair share, by distance and by grade, with the noise called noise.

Over the last 2 years at Swindon, 2,923 races, trap 2 has been a clear plus at +2.7 points against its fair share and trap 6 a clear minus at -2.3; the other traps sit inside noise. The track is 422 metres round, measured from our own split times.

All distances

Trap 1 sits inside noise at -0.2 points. Trap 2 is a clear plus at Swindon: it wins 20.5% of races against a fair share of 17.8%, +2.7 points on 2,892 runs. Trap 3 sits inside noise at -0.2 points. Trap 4 sits inside noise at -0.5 points. Trap 5 sits inside noise at +0.6 points. Trap 6 is a clear minus: 10.2% of wins against a fair share of 12.5%, -2.3 points on 2,031 runs.

By distance

Only distances with at least 1,500 runs get their own chart; a bias that shows at one trip and not another is usually the run to the first bend at work.

476 metres, 2,427 races

Trap 1 sits inside noise at +0.4 points. Trap 2 sits inside noise at +1.9 points. Trap 3 sits inside noise at +0.1 points. Trap 4 sits inside noise at -0.4 points. Trap 5 sits inside noise at +0.1 points. Trap 6 is a clear minus: 9.9% of wins against a fair share of 12.0%, -2.1 points on 1,611 runs.

262 metres, 465 races

Trap 1 sits inside noise at -3.1 points. Trap 2 is a clear plus at Swindon: it wins 24.7% of races against a fair share of 17.8%, +7.0 points on 461 runs. Trap 3 sits inside noise at -1.3 points. Trap 4 sits inside noise at -1.8 points. Trap 5 sits inside noise at +2.8 points. Trap 6 sits inside noise at -3.5 points.

By grade

Grade letters grouped: A for graded races, D for sprints and dashes where the track uses them, S for stayers, and so on.

Bias in points against fair share; * marks a clear bias.

The market and the fields

The Betfair favourite won 36.0% of 2,917 priced races here over the window. 65% of races went off with six runners; the rest had a vacant trap, which is why fair share is measured on runs and not assumed to be one sixth.

How to read this

Fair share is each trap's share of the runs, so a vacant trap 6 does not hand trap 5 a bias. A bias is called clear only when it is at least 1.5 points and at least 2.5 standard errors from fair on the runs behind it; anything smaller is a difference inside noise, and the page says so. The window is the last 2 years, refreshed whenever this page is rebuilt. The live version, with the going and our own accuracy at this track, is on the track page.