EXTERNAL LINK
EXTERNAL LINK
EXTERNAL LINK

A track can boast top money for A1 winners if they rarely stage them or offset the cost by reducing the Prize money for the also rans.You need to look at the overall Rating of the Grade, the percentage of races in each Grade and the overall prize money for each Grade to assess the generosity of Prize money at individual tracks. If you look at the Graded data on GStatsUK you will see how the Grading structure, and therefore the related Prize money, has changed at some tracks since racing recommenced in June. Not a simple task and don't the tracks know it!!

They must have the strongest pool of dogs in the country!

Sheffield ratings were absurd too, theyve been toned down a tad but are still too high

Ditto Central Park

By far the best way to establish time relationships between one track/distance and another is to start with an individual dog, then another and another, etc.

Identify those which have raced at tracks A and B within, say, a 45 day period and then check the time difference. Do this 500 to 1,000 times and average those differences. And so on for C, D, E

This presumes that you are working with times that have already been corrected for slow/fast tracks. We do that after each meeting by comparing times run by each dog finishing within 3 lengths of the winner and relating that to its career performances.

I see Irish corrections of 20 or 30 for the going but they appear to be just someones guess (?). These are extremely high figures and would rarely be justified. Outside of a muddy bog I would be sceptical, especially of the fast figure. Dogs are usually more flexible than that.

Of course, this is just a portion of the puzzle.

But you need a good database and a query facility to do the job. Mine takes a couple of seconds to process the query. For example, there is a 0.05 difference between Sandown 515m and Meadows 525m based on some 2,000 samples. Users might well make their own corrections as a function of each dogs personal habits but generally it works well.

I have no idea how G-D does theirs but I do know most of their Oz form records are incomplete.

Ideally, the national authority would have a form database which anyone can access and then download race results into their own program. Oz does not have that due to various state by state jealousies. (The horses do, though).

And, yes, Ireland has just won this years Melbourne Cup (again). The greyhound version will be on shortly.

That's interesting, I had a difference of between 8 and 9.5 lengths between the two tracks when the lure was on the rail over a number of yrs. One of us was wrong.

Also noticed lately, in the the section between the post the first time around to the back straight at the Meadows in 500m races, top dogs were breaking 12.40 when 4yrs ago they were struggling to break 12.50 - that's too fast, so something has changed there.

Paul - I well know the problems with GData Ratings as they once used my Ratings but couldnt adapt their systems to cope with changes and for some time now generate their own Ratings. As you suggest many of the Standard Times from which GData generate their Ratings are wrong but the real problem is that since losing their feed from the GBGB the British results have had to be entered manually with, understandably, many consequent errors.

Tracks like Nottingham and Sheffield are a particular problem for GData as large going allowances are commonplace and easily missed on manual entry a problem not helped by the GBGB sometimes posting results first and adding Going Allowances later. This is a fact rather than a criticism as I fully sympathise with the GData volunteers who do such an incredible job in manually entering such a large volume of data.

I agree that Central Park 480m Ratings are far too high but the Crayford 380m and Hove 500m Ratings much less so and whilst the Standard Times on which the Ratings are based are easily corrected there is little point doing so with a flawed database which it would be a monumental task to correct.

So Paul I think you may be reacting to missing going allowances such as I referred to earlier in this Thread where at a Hove meeting all of the 500m races bar one showed the Going as Normal when it was officially 30 Slow. Pennys Bobby's fastest of the year 23.06 at Crayford shows a Rating of 127 which is way too high partly because the 20 Fast going called is missing. Errors such as these are commonplace for British tracks on GData and for this reason alone GData Ratings should in my view be treated with extreme caution.

Bruce - I have not posted your full post as I just wanted to explain that the Going Allowances you refer to in Ireland are not 'going allowances' as you may understand them but rather an adjustment applied in an attempt to make times comprarable between tracks for the purposes of the IGB's Grading system. Sometimes the Grading Allowances are adjusted to reflect the going on the night but there is no consistency in doing so. Irish Ratings have to be viewed with this inconsistency in mind.

At British tracks the going on the night is reflected in 'going adjusted' times which analysis shows are generally reliable. Of course, such allowances cannot be precise because the going may vary for Railers, Middle runners and Wides.

There are a reliable databases for British racing which I use to develop Ratings largely based on a statistcal analysis of 'going adjusted' times recorded by mature greyhounds winning races with a clear run at different tracks. As I explain in a previous post the problem GData has is that they lost the results feed they once had from another site and have to enter race data manually which as you can imagine is a formidable task prone to error.

Thanks for that - although I won't pretend to completely understand it.

In any event, the underlying objective is the same ... to relate performances from one track/distance to another and thereby make more sense of a dog's overall form.

If there is one thing that emerges from all the to and fro it is that any country must have a primary source of form data - complete and professionally done. Anybody can then access them with some confidence.

We have two but they are done differently and with different inclusions and omissions. There are also major differences between publishers of formguides (several) which then get picked up for use at various clubs and pubs to put on the wall or on various touchscreen devices. We also have a digital stud book but it does not talk to the form data people.

None of this helps progress the industry.

In our case, we plucked bits from everywhere to insert into our own database (even including "encrypted" files). No doubt many others do the same. That gets us to near-99% accuracy but it is laborious and time consuming, especially when one or more sources change what they show and we have to go back to scratch again. Physical changes to tracks are also a burden.

The outcome is if you make it hard to do you lose customers or turn off those you do have (like me). It is also why tri-code racing's share of the gambling dollar has gone from 50% a quarter of a century ago to some 10% now.

Bruce - I sympathise with you as the Irish Grading system isn't easily explained in a few words but the objective is as you suggest and allows greyhounds to contest Graded races in their designated Grade at any Irish track.

As you will know the standard trip at British tracks, unlike Irish tracks, is not standardised and each track has its own Grading system and the relatively few greyhounds switching tracks have to have qualifying trials before being allocated a Grade.

As I said in my previous post Going Allowances have long been a reliable feature of British racing and there are a number of databases which are freely available for individual access and also reasonable access to race videos which support a strong but declining Betting market.

The main GBGB database is encrypted and they control feeds and when GData lost their direct feed rather than the solution you opted for have relied on manual updates which given the amount of British racing is a formidable task which understandably falls well short of the 99% accuracy you have achieved and so GData's British Ratings are unreliable hence the origin of this Thread.