Survey Pinpoints Problem Tracks

A couple of months ago we flagged our intention to analyse race performances all around Australia to try to work out what made a good track and what a bad or disruptive one.

We got some interesting comments from readers but not enough to say they were representative of trainers or punters. However, we now have over six months of data to look at and that is plenty to help analyse the busy trips at least.

On a perfect track we expect eight dogs to go around virtually untouched. This means that, to a large extent, their respective abilities would decide the race. Of course, that rarely happens. Three things will hold them up: any extra distance they have to cover (eg from an unfavourable box), interference, or a lack of field sense.

We can't do much about that last item. Dogs are like people – some dumb, some more shrewd and agile. Having said that, we suspect it is not the most important factor by any means.

Certainly, boxing a runner on the “wrong” side of the track does not help but that influence is spread evenly across all tracks anyway so our results would be comparable with each other. Still, the effect of different boxes is a major factor in race outcomes but, over time, it is a consistent one. It can be easily assessed statistically and you can get a good idea of the size of the penalty it imposes. (Although not so easily in Queensland or Victoria where authorities have decided in their wisdom to publish box data only for the last 12 months. This is grossly misleading and indicates a misunderstanding of statistics. The resulting small sample numbers do no more than lead you up the garden path. Curiously, both states once published long term figures, so why did they change?)

For a given race, the early speed of each runner can be used to work out whether a dog can overcome – or take advantage of – a given box. When Dyna Bert ran second in the Easter Egg despite having box 8, it was able to cross because it ran a smart first section of 5.44. In fact, the three placegetters were also the quickest early, which is not unusual in big race finals, and suggests they missed any significant interference.

Yet another indicator of track performance is the size of dividends. Consistently high average payouts tell us that outcomes are unpredictable. The market is unable to get them in the right order. In particular, large Trifecta or First Four dividends tend to indicate a high level of interference, and to especially affect placings, yet they are still not definitive enough for our purposes.

We are looking elsewhere in this study. Can we get a measure of the amount of actual interference in a race and thereby assess what causes that interference? The answer is – very likely.

Therefore, with a massive amount of help from a colleague skilled in computer analysis, we took race results from every state over the last six months and worked out for each trip (a) the percentage of races involving falls and (b) the percentage of races where dogs finished 20 lengths or more behind the winner. The assumption is that either of those factors were a function of interference. It may well be that a few of each were caused by injuries (which we tried to edit out) but, on the other hand, some margins shorter than 20 lengths would also be interference-related so one may balance out the other.

The big result is that six tracks involving ten different distances figured in fall rates above 10% while a very large number had rates of between 5% and 9%. The worst six were (all distances), Casino (484m), Gosford (400m and 515m), Grafton (407m), Nowra (520m) and (390m). These trips all possess problem first turns, either because of their design or their closeness to the start, or both.

At the same time, the major tracks at , , Sandown Park, and all had fairly high fall rates (5% to 9% of races) and poor 20 length ratings. So did a large number of trips at provincial tracks.

In contrast, very low fall rates were incurred at Devonport, , and . However, before you offer congratulations to our two smallest greyhound states, note that Launceston and Cannington do not rate very well, and both have dubious first turns and a strong inside bias.

Shortly we will be publishing more detailed comments, together with all the data produced by our survey.

Meantime, there are some outstanding conclusions.

First, there is strong evidence that falls are more often due to the layout of the track, usually on the first turn. Visual checks add to that confidence – as, for example, when dogs are unable to maintain an even course around the turn.

Second, the always dicey bend starts are not producing a high fall rate but they do record high figures in “distanced” runners. The theory here would be that in the early part of such races the dogs have not yet had the opportunity to build up to top speed, which otherwise would be a major contributor to falls (just as it is in motor cars). However, the jumbly nature of the early part of these races undoubtedly shoves some runners out the back. Hence they rank highly in the 20 length-plus category.

Third, one-turn tracks tend to be safer than circle tracks, at least where a longer run to the first turn is present.

Fourth, there is a high degree of consistency between the new interference data and other measures, such as dividends and winning boxes. This suggests you can argue about the fine tuning but not the overall outcomes.

Finally, it is inescapable that our tracks can be made better – ie safer, with less interference – if we were to study the track design subject more carefully.

The absence of such information (which is partly why we ran this study) is scandalous given that the means to undertake that study are readily available and that the upside is huge. Three items at the top of that benefit list would be the means to encourage more punters and more total betting, the means to avoid wasting millions of dollars on building bad tracks, and the justification to defend the sport against the sometimes illogical, sometimes fair criticism from anti-racing lobbies.

In short, it aint right but it is fixable.

The study will be continuing in order to develop more reliable data for lightly used trips.

Subscribe
Notify of
guest
0 Comments
Inline Feedbacks
View all comments