Going preferences, going descriptions

Bar the Bull said:
I am going to sound like a right pain in the arse, but U-shapes are to be expected, I am pretty sure...due to the differences in the number of runners per race.
Click to expand...

In that case should it not look more like a (sloping) L than a U? The sample doesn't actually include the two extremes of going at all, as it doesn't go any 'faster' than good to firm. I can't see any reason why gd-firm races would attract smaller fields (and the sample size for good to firm seems to actually be broadly consistent with the good ground sample). If the description included 'firm' or 'hard' then I could see why a U would be strongly expected.

A big part of the problem is that the heavy ground sample is so small (incidentally, the mirror image of the same problem exists when looking at NH sires). Green Desert apparently does fairly well on heavy ground from a comparatively small sample (10%). If we leave aside heavy ground, though, the stats show he does consistently better as the ground gets faster (8% on soft and good to soft, 10% on good, 14% on good to firm). The same broad trend applies to Oasis Dream.

In those types of situations I would be more inclined to ignore the heavy ground sample and infer that the sires' progeny prefer better ground than to interpret a 'U'.
 
trackside528 said:
In that case should it not look more like a (sloping) L than a U? The sample doesn't actually include the two extremes of going at all, as it doesn't go any 'faster' than good to firm. I can't see any reason why gd-firm races would attract smaller fields (and the sample size for good to firm seems to actually be broadly consistent with the good ground sample). If the description included 'firm' or 'hard' then I could see why a U would be strongly expected.
Click to expand...

I don't know the answer to that. But given that Grey says that most sires show a U shape, and if this is the case (you can never trust the man), then it should be factored in.

Don't forget how seldom a clerk of the course will post "firm" in the going description.

The long and short of it is that there will be an average shape to the curve. I don't give two shiny shites if it is a U, a V, an L or a swastika. Grey's point about sires liking or disliking extremes of going/tacky ground doesn't make any sense until you find out the average shape, and then work out the deviation from the average shape.

It is automatic that the average shape of all sires in aggregate is an inverse of the number of horses per race. Don't forget that.
 
I am using Racing Post data, Trackside, which only has the five going classifications. Firm and Hard are comparatively rare and are presumably lumped in with Good to Firm.

I agree that field sizes in flat races are unlikely to be any smaller on Good To Firm. After all, it is a BHA instruction to the racecourses that they aim to produce such ground for flat races.

But Bar has a point about the data needing to be scrubbed, so we'll see. I'll get it together and send it to him once I find out he's not going to overcharge me.
 
But given that Grey says that most sires show a U shape
Click to expand...

I'm saying I suspect that those resembling the Us are the largest group and that when you add those tending towards the inverse Us you have a majority
 
Bar the Bull said:
The long and short of it is that there will be an average shape to the curve. I don't give two shiny shites if it is a U, a V, an L or a swastika. Grey's point about sires liking or disliking extremes of going/tacky ground doesn't make any sense until you find out the average shape, and then work out the deviation from the average shape.
Click to expand...

And that's where you come in.

If you send the data on to me, Grey, I'll see what I can do with it as well.

One other issue is Flat and NH sires. They probably both suffer from the same problem (skewed sample sizes) but in opposite directions. Differentiating between the two will probably be needed I would imagine.
 
One other issue is Flat and NH sires. They probably both suffer from the same problem (skewed sample sizes) but in opposite directions. Differentiating between the two will probably be needed I would imagine.
Click to expand...

I was thinking that. Another reason for it is that NH and Flat going decriptions might not alway be the same.

By the way does anyone have a better means of extracting data? The only way for me to do it will be to copy/paste from each sire's progeny stats page.
 
Bar the Bull said:
I am going to sound like a right pain in the arse, but U-shapes are to be expected, I am pretty sure...due to the differences in the number of runners per race.

I think the analysis has the potential to be interested, but the data needs to be scrubbed first.
Click to expand...

Agree that data needs to be scrubbed, but if you suspect that u shape should be expected, then the flatlines are interesting then, as they go against the standard.


Not sure how accurate the data here is, but it goes some way to showing preference and shape of s/r at least and generally ties in with my data:

http://www.drawbias.com/sirestats.html
 
Last edited:
Thanks Lard,

That link could be helpful.

The percentages seem to be a bit different to the RP data, which combine flat and jumps. Do you think the stats on your link might be flat only? Another difference is the extra column for Firm ground.
 
I don't know, I don't know how up to date it is. I use it for interest but rely on a paid source of sure stats. Just thought it is a fairly nice way to display preference.
Same with the draw - it doesn't say how up to date it is.
 
Just thought it is a fairly nice way to display preference.
Click to expand...

Definitely, a very handy reference tool.

It says somewhere on the page that it was up to date at the start of this year. I think it's flat only, because I only see mainly flat sires on the list.
 
On the basis of the list linked to by Lard, aggregating the average strike % tentatively suggests a very shallow 'U' shape: 11% heavy, 10% soft, 10% good to soft, 10% good, 11% good to firm, 13% firm.

I'm not sure about the data though. Some of it deviates pretty markedly from the RP data, not just numerically but also in what the figures would indicate with regards to going preferences.

It may be explained by splitting flat/NH sires I suppose. That would be nice; we wouldn't want hunter chasers paddling three miles round Downpatrick influencing the sample.
 
Barry, Peter, I've sent some data your way.

I haven't been able to make a complete split, though, between flat and NH. A lot of sires have winning progeny in both codes and the RP data I'm using doesn't provide a breakdown by both going and code of racing. If more than two thirds of progeny wins come in either code I have classified the sire accordingly. Otherwise I have classified them as 'dual' stallions.

Keeping the above reservation in mind, field sizes on the flat appear to be very similar across all types of going (about ten runners per race) except Good to Firm, where the average drops to about nine. The average field size in NH races is also close to ten, ranging from 10.6 on Heavy down to 9.1 on Good to Firm.

My guess is that the U pattern will be found to be more pronounced in NH sire progeny than the flat sires, but I will wait for your findings. Over to you.
 
Last edited:
Take Presenting, a horse whose progeny generally like good decent ground. I was asking a fella one day why so many win on soft ground given the perception that they are good ground horses and he said "they're strong horses who just stay better than anything else". So the physical condition of the horse has something to do with it too. For example, Brave Inca preferred good ground but because he was a strong beast, he handled soft ground better than most and this was a huge advantage to him. How do you figure this into the stats?
 
Presenting is by Mtoto out of a Persian Bold mare. Both sires' progeny won more often on faster ground than slow, but Presenting's win close to the same percentage of races on all types of going.

Sadlers Wells is another sire whose progeny do equally well on all types of going. As it happens, out of the current top 100 sires in the jumps table only Sadlers Wells has had more runs over jumps since 1989 than Presenting (10,650 v 9100 respectively).

Perhaps their progeny have a general ability to cope with any surface, and that is why they are high up the prize money tables and in demand? Or maybe it's because so many mares are sent to them that their progeny are representative of the population as a whole, in its diversity at least if not its quality, and because of this there is a spread among their progeny of mudlarks, fast and dead ground specialists?

Cantoris's anecdotal evidence would point towards the first explanation. If Presenting's progeny are stronger than others, is that not a genetic trait which equips them to do as well on testing ground as they do on faster?

As for Brave Inca, how could you figure him into any stats? Just be grateful he was a freak!
 
Last edited:
Really interesting thread chaps and would be interested to see the results of the analysis.

Is it possible to assess the progeny by sex as well? That would appear the most objective physical measure by which you could cut the data.
 
If I come up with anything useful I'll let you know, Aragorn. Bar has done some nice work on the data and I want to muck around with it a bit more.
 
The going descriptions that are used are quite limiting because for each description there are at least two sub divisions of each slice ...a drying & wettening surface. Drying ground and wettening ground would mean that drying soft is a different test to wettening soft in that horses will go through wettening soft but find drying soft sticky..the same for all descriptions. I can't see that this can be measured now obviously so it has to be ignored.

Its impossible to know what exact type of each ground we are actually looking at in any stats because of that...also..Good to soft could actually be Good or Soft just because the official going is reported incorrectly. Exact isn't possible with anything we look at tbh.

Showing stats with the in between descriptions to me is probably misleading. It would also be the same at the extremes..how many times is hard actually good to firm and heavy actually soft..and visa versa.

Imo i would group the goings into three groups..that way the overlap between each is fudged a bit but possibly more meaningful due to not slicing the going into to many pieces that will in 25% or more cases be incorrectly reported officially anyway.

Fast = Hard+Firm+GtoF
Good = Good+GtoS
Slow = Soft+Heavy

I don't really believe Hard ground exists that much hence slinging it in with fast doesn't lobside them imo..to me Hard and Firm is the same going description.

any suggestions of a better grouping?
 
Last edited:
Back
Top