atr draw check

[gerryk]

sorry lads this may be a stupid question.just wondering if lads use this tool on the site?how accurate is it for form study?my question really has to do with the figures that appear when you hover over the green or red bars on the generated charts.for example for stall 14 it may say 'expected win 1.04, actual wins 2.0=98.78%' when it is a 15 runner race.these figures actually appear for the duke of york stall 14 tomorrow when you click on the draw check and dont change the parametres.can someone explain what these figures mean and how they are generated?obviously its a positive for this stall although a relatively small sample.

[fintron]

Re: atr draw check Hi Gerry, I use this feature of ATR and my understanding of it is that any green bar for a stall is a positive, i.e. that draw has produced more winners than would have been expected if there was no draw bias and all stalls had an equal chance of winning. Anything that is red has produced fewer winners than would have been expected if there was no bias and all stalls had an equal chance of winning. It is sort of like a chi-square test in a way, comparing observed versus expected values. I guess how it works is that if you have a hundred races all containing ten runners then if there is no bias then the expected value for number of winners should be 10 for each stall. However, the number of expected winners per stall changes when you start to get bigger fields since stalls 14 and 15 will be occupied less frequently than stall 1, which will be occupied every race and is thus expected to yield a higher number of winners. The caveat with it is that the sample sizes of data used to make the graphs is often very small like you say, so it if you are looking at the draw bias, purely on heavy for example, you might be looking at a chart generated from only a dozen races (all races are listed at the bottom of the page). The reason stall 15 is included in the graph you mention is because when you selected the parameters at the top of the screen, 13,14 and 15 runners are selected by default. This is to try increase the sample size of the data, however, as you say, no runner goes from stall 15 tomorrow, that is why it is greyed out on the chart that is produced. If you unclick '15 runners) and re run the graph then the greyness disappears, but by default it seems they tend to filter past results with the same number of runners as the race in question, plus or minus one runner, to try give the results a bit more statistical power. Hope that helps.