Interference Study Reveals Much
Written By Bruce Teague Monday 16th April 2012
Interference decides the outcome of nearly half of all races, according to a study of Australian greyhound meetings between September 2011 and March 2012. The study is ongoing, mainly to allow samples of lightly used trips to build up to a more reliable statistical level.
The report is in four parts: this summary of results, a brief note on most tracks, a list of parameters used to query race files, and the data summary itself. The latter three parts are available for inspection or download â€“ see bottom of this article. Preliminary comments can be seen in the article “Survey Pinpoints Problem Tracksâ€ť published on 12 April 2012.
Readers are invited to make comments to the editor at email@example.com.
The aim was to identify whether, where and how often interference plays a part in race outcomes and, where possible, to suggest specific areas which are contributing to the level of interference.
The points covered in this study are not exhaustive, far from it. Indeed, it will have served a purpose if it leads to state and national authorities commissioning an independent review of the science of track design, to be conducted by a small panel of experts using modern technological aids. The outcome of that review would be unique as, to our knowledge, no such effort has been mounted anywhere in the world, let alone in Australia. Bits and pieces, perhaps, including work on turn injuries, but not a thorough study of the subject.
The achievement of excellence in this field would be of immense value to the sport, challenged as it is today by many competing gambling and recreational options and hampered by ultra-traditional organisational structures.
Here, we have taken a simple approach. We counted up the number of times a fall occurred and the number of times dogs finished more than 20 lengths behind the winner. The percentage of races affected is then shown. Either measure is primarily a function of involuntary interference; not the only cause but by far the main one. In both cases, those proportions are claimed to be a proxy for the integrity of the track â€“ does it maximise safety and does it allow runners to show their best?
Our major support for such an approach is that some tracks â€“ sadly not too many â€“ generate a low level of interference, while others generate lots. We need to know why this is so.
Our analysis has picked out a number of trips where falls occur in 10% or more of races while often also recording high figures for 20 length-plus finishers at the same time. They are Bathurst (all distances), Casino (484m), Gosford (400m and 515m), Grafton (407m), Nowra (520m) and Warrnambool (390m).
A much larger group recorded fall rates of between 5% and 9%, including major tracks at Albion Park,Wentworth Park Sandown Park, The Meadows and Launceston, which also had poor 20 length ratings. So did a large number of trips at provincial tracks.
In contrast, very low fall rates occurred at Devonport, Hobart, Mandurah and Northam. However, before you offer congratulations to our two smallest greyhound states, note that Launceston was poor while Cannington was only ordinary; both have dubious first turns and a strong inside bias.
A question that might be asked in respect to Tasmaniaâ€™s two one-turn tracks is why dogs generally take a straight line after jumping there when comparable tracks elsewhere see dogs crashing left and right. Obviously, Tasmaniaâ€™s success is due to some combination of features or equipment, but what are they? No-one seems to know.
Indeed, the fact that a state may have a mix of good and bad tracks suggests that pot luck methods have been used to build and maintain them. Further, since NSW (with the longest and largest history of greyhound racing has five of the worst six tracks, two of which have been re-built recently, it leads to suspicions that the authority or the clubs lack the means to assess what makes a good track. And the above list does not count Dapto, Richmond and Wentworth Park which all have whopping figures in the 20 length-plus category. In each of these cases the faults are plain to the eye yet, again, all three have had work done on them â€“ supposedly to improve outcomes. Those efforts have not only failed, but the failures have never been acknowledged.
That last comment could also be applied to Ipswich, which has a diabolical bend start for 431m races and a flat first turn for 520m runners (as does Richmond). Race after race sees dogs scattered across the track, desperately trying to find their way back to the bunny. Ipswich scored falls in 7.7% of 520m races and 20 length-plus margins in 40% of them. A huge 66% of 431m races are won by the early leader (easily Australiaâ€™s highest ratio), indicating that the rest of the field has little chance of making up ground.
The evidence at Ipswich, The Gardens and Richmond suggests that designers may have first put the boxes in position and then created a circuit around them. A better way of ensuring the necessary gradients would be to do the opposite â€“ first build the circuit and only then place the boxes where needed. In other words, first create a velodrome effect, and then modify that as required.
In other cases – Angle Park 515m being a prime example â€“ the figures are not so bad but actual race running is routinely hampered by jumbles at the first turn, where some dogs frequently lose their chance through no fault of their own. Interference is sufficient to change the running order but not to relegate the victims too far back and so affect our statistics. Itâ€™s all a question of degree.
Angle Park and The Gardens (which has very high 20 length figures) share at least one problem in that their home turns are also a little flat, making it harder for a dog to hold the rail or to pass on the outside of the leader. Curiously, the otherwise risky Gosford track has a well contoured home turn. Was this pot luck again?
On a related subject, Bulliâ€™s home turn is by far the worst in the business. Its flatness is so extreme that multiple runners can be seen veering off, frequently affecting other dogs on the way. This disruption does not impact on our figures but does significantly change the finishing order, thereby affecting exotic punters.
The disappointing aspect of such poor home turns is that most could either be readily fixed or could have been avoided in the initial track design process. It is a simple matter of attending to the lateral gradients, and offers evidence of a sloppy attitude to basic design needs.
We could go on, track by track, but the story is much the same anywhere. Over one third of all races involve interference sufficient to relegate some runners so far back as to make it impossible for them even to run a place. Certainly, some may not warrant that anyway. But many dogs are not nippy at the start and rely on strength to make their way through the field. A disrupted field hampers their ability to do that, thereby leaving the race at the mercy of quick starters which have stolen a break. Such events also ruin the competitive element of a race, which is the prime, and most attractive, stimulant for any contest. Itâ€™s what excites both dogs and fans.
Digressing a bit, this bias against the moderate beginner â€“ which itself is harmful to the economics of dog ownership â€“ is further emphasised by the current strong trend towards 300m and 400m races, and away from the more testing 500m-plus trips. Track design can also reduce the success rate of wide running dogs, which make up perhaps one in five of all greyhounds. Angle Park and Albion Park, for example, are unhelpful to wide runners.
Similarly, cutaway turns at tracks such as Wentworth Park, The Meadows, Launceston and Cannington bias the track in favour of railers from inside boxes, even though these have a built-in advantage to start with. Notably at Bulli, where the transition or cutaway turn got its start, dogs routinely find it hard to maintain a even course through the turn, leading to one or more being abruptly shoved out the back. The race becomes less fair as a result, all due to the influence of the track builder.
It is remarkable that those responsible for track building have not only failed to search out the better, low-interference examples available (there are a few), but have copied or repeated examples of disruptive tracks. Dapto, for example, underwent a $0.7 million re-build but came up with a dead copy of what previously existed. Dogs are still splattered all over the first turn. Richmond repeated its flat first turn and 400m bend start after spending nearly as much. The Meadows shifted its 525m boxes twice but still boasts a strongly biased circuit. Sale had a complete re-build but maintained its 520m (was 511m) bend start while the turn into the home straight frequently confuses runners. Above all, Sandown has been worked on several times yet for the past 15 years it has consistently produced the highest average dividends of any major track in the country, and has relatively high fall and 20 length figures as well. Neither dogs nor punters can work it out.
There is a habit in the industry to gauge success on the basis of a good dog jumping clear and leading all the way in smart time. This is nice but it does nothing to illustrate the worth of the track in design terms. Not only do faulty designs put dogs at risk but they deter fans from investing more, whether as owners or punters. It is a difficult enough task to maintain a dogâ€™s fitness for the two years or so of its economic life without adding to that challenge by creating disruptive races.
Nothing will answer all these posers except a major independent investigation into the science of track design. Here we have offered a start.
Finally, an official review would be well advised to study New Zealand tracks at the same time. While we lack data to analyse them properly, race viewing suggests that on average their races are far cleaner than those in Australia.
There are many good things about greyhound racing in Australia, not least being a range of top dogs, skilled trainers and state-of-the-art service providers. However, our tracks fail to do them justice.
Our view is that the interference levels shown in this study are unacceptably high. More to the point, they do not need to be. Remedial work is both possible and desirable.