What Can Winning Boxes Tell You?

Many years ago I wrote that “A well laid out track should produce a set of winning box percentages where no one box delivers less than 10% of all winners and where box 6 is the worst of the lot. If that does not apply, look for some peculiarity in track design which might be causing the problem”.

That principle still holds true today. In fact, you could enlarge on it. Three guidelines will quickly identify odd ones out:

1. Box 1 should be at or below 18%.
2. Box 6 (the worst) should not be less than 10%.
3. Box 8 should be around 12%.

All those rules hold good for one country – New Zealand – but not for the USA and not for Australia, where outcomes are erratic.

A survey of the seven main NZ tracks, using long term data, confirms that winners are more evenly spread with perhaps just one exception at Forbury in Otago, where 1 and 8 do better than expected over its 545m distance (which is worth checking).

Considering box 1, six of the nine sample tracks we looked at in Australia produce well over 18% of winners – Albion Park (20.1), Launceston (20.7), Sandown (18.4), Meadows (19.4), Wentworth Park (19.3), and Cannington (18.8). Most of those also had either many more or many fewer winners from box 8 than the expected average. In four of the nine, the worst box was not 6 but either 5 or 7. In total, that’s a disease.

Our underlying principles do not just apply in Australia and New Zealand. 10 year figures for English tracks, where only six runners take part, show the same thing. Here is an 8,000 race sample taken from a U.K. publication (Win at Greyhound Racing, Oldcastle Books, 1977):

Box 1 2 3 4 5 6
Winning % 19 16.6 15.3 14.9 16.7 17.1


Certainly, the English track managers seed dogs according to their railing ability. Even so, the wins are nicely spread, indicating a low interference level, and the saucer-shaped set of figures conforms to the basic principles of good design. Irish data is probably similar but their authority does not publish any figures.

So what’s wrong with Australia?

The first clue is obvious – cutaway first turns (the turn within the turn) at Wentworth Park, Bulli, The Meadows, Launceston and Cannington always do two things; they give the inside dog an additional advantage, one that it does not really need; and some dogs cannot handle the more complex turn, thereby causing interference as they move off.

The next clue is that some turns bring dogs together, rather than keeping them apart. That causes an unholy mess: The Meadows and Albion Park seem to be the worst there. Allied to that is the Sandown experience, which does all sorts of funny things, such as causing dogs to bump on the way to the turn and then prompting some inside dogs to dance to the right when they get there.

Horsham and Angle Park are better than most, although some mixing is present on the turn. Both have relatively low but still dominant box 1 figures.

Flat first turns at Richmond and Ipswich cause no end of trouble while Dapto’s boxes are jammed in against the line of the running rail. All generate interference.

In other cases, notably Bulli, a flat home turn throws dogs off and significantly changes the running order.

Generally, Australian track designs display a suck-it-and-see syndrome. In other words, the designs have no backing in engineering or any other discipline. They have just happened.

The foundation of New Zealand track designs is unknown, although their relative young age suggests that they may have learnt from experience elsewhere. Observations suggest turn banking is kinder than in Australia while the big swinger is the use of the follow-on-lure system which appears to offer dogs a wider and higher view.

Indeed, lengthy FOL trials in Brisbane and Adelaide produced lower levels of failing to chase convictions and lower injury rates as well. That outcome is consistent with obtaining a good sight of the lure and also with keeping dogs further apart. It is more than a shame that such experience has been wasted.

So, are winning boxes the answer to good track designs? No, not by themselves, but they are the prime evidence that tells you to look further. They are the red flags.

The trick is to find a solution that meshes the desires of a random and changing mix of dogs with a man-made collection of steel, loam and fluff. That’s not a simple task. To do it well demands some thousands of hours of study. However, we do have some basic principles to work with.

1. Never put boxes on a bend.
2. Make turns simple, even and well banked.
3. Switch to the follow-on-lure. There is nothing to lose.
4. Don’t guess – get the evidence first.
5. And listen to the dogs – they know best.

The above comments all refer to performances over the track’s main distance, where a good run to the first turn is available. Winning boxes for other distances will be more erratic as they are located on or near a turn.

Good data on winning boxes is not as easy to find as it should be. A major problem is that Victoria and Queensland some time ago shifted from long term data to providing it only for the previous 12 months. That means the majority of trips offer too few samples to justify statistical accuracy. A minimum of 400 is needed, but preferably 1,000 or more. In any case, current data for Queensland disappeared when it switched over to the NSW Ozchase system. The same thing happened when Tasmanian data was inserted.

In terms of racing information generally, the switch to the Ozchase system has put the industry back by 20 years.

The only reliable box information comes from National Tab form although even there the odd trip is doubtful as it does always not re-start data collection when the track undergoes a physical change but continues to use the same distance. Such changes always produce different winning box numbers. One example of that is the Maitland 400m trip.

Past Discussion

  1. Bruce,
    Check Grafton results from the main race distance of 407 mtr. it currently shows from the last 931 races box 1 has the best winning %of 17,2 and the worst is box 3, 10.1 % box 6 11.0%
    This confirms your theory in regard to track and design giving reasonable fairness to all starters.
    Laurie Arnott

  2. Bruce,

    Check Grafton results from the main race distance of 407 mtr. it currently shows from the last 931 races box 1 has the best winning %of 17,2 and the worst is box 3, 10.1 % box 6 11.0%

    This confirms your theory in regard to track and design giving reasonable fairness to all starters.

    Laurie Arnott