- This topic has 48 replies, 14 voices, and was last updated 18 years, 5 months ago by
.
- Posts
The level stake profit for Topspeed is poor, as it is for RP ratings and Postdata. The long term average for those who tip in every race was about minus 10 per cent on stakes when I last looked, but it has almost certainly deteriorated when you consider the SP changes since November 2006.
From the above, I cannot help but agree with you about the poor tipping record of the Racing Post ratings(form and speed). I think that any analysis of tipsters selecting in just about every race will yield the same dismal results. I’ve discovered through experience that ratings based approaches are only useful if you use the ratings as a starting point and work from there.
The VDW guys seem to think that there is some undiscovered formula that might lead to profitable betting and that if they keep analysing his approach, they might discover it.
I used to think a bit like that….but I was so much older then, I’m younger than that now. (Bob Dylan)
This are the figures for this years flat racing, so far;
RP RATINGS – TOTAL 1199 4762 25 818 580 457 -554.34
SR 25%
ROI: -11%TOPSPEED – TOTAL 569 3352 16 469 401 324 -520.11
SR 16%
ROI: -15.5%It doesn’t look like a big difference when you look through it quickly but I reckon that’s quite a bit. Especially when you consider some of the other stuff on the internet has a positive expectation and a higher strike rate.
I think racing is too unpredictable and luck plays a greater part in things than we would care to think to spend forever contemplating your navel about stuff.
IMO you need a practical approach forfollowing time selections as with any other appraoch, which includes being selective. This includes factors such as non-hcap races up to a mile on the Flat and hurdle races not more than 2m 110f over the jumps. Selection to be top-rated and over the same D & G as top-rated in the previous season. Top Speed has some reasonable examples today:
Mus 1.30 Hamish McGonagall. 3.00 ElCoto
Tow 1.40 Moratorium, 2.40 Butterfly Rose
Hun 1.20 Sir HaydnRacing Post Ratings do seem to have had a worse than normal season on the flat.They do better over the jumps though. Topspeed does perform badly and there must be something wrong with their standard times.Other speed ratings do better like Mark Nelson from Racing and Football Outlook who uses Ken Husseys old system.
My own speed ratings are based upon the ‘conventional’ use of standard times as a benchmark. When starting out, I used only courses for which electronically-recorded race times were available over a 10 year period and for which form ratings for the winning horses were also available. This restricted the number of courses but, I believe, helped produce more reliable results. More recently, electronic timing has become the norm and I’ve started to produce (in some cases tentative) standard times for other courses. In calculating the standard times, actual (past) race times were adjusted for age of and weight carried by the winner and it’s ability (form rating) in order to arrive at the familiar concept of the standard time representing that achievable by a horse rated 100 and carrying 9 stones. Races on extremes of going were discarded and accepted statistical techniques applied to arrive at an estimated standard. For some courses, the calculation of standard times at various distances is relatively straightforward. Others can present problems, according to variations in track confirmation – Goodwood being a prime example. I’ve not looked in any detail at the methodology employed by Topspeed, or the RP standard times.However, as discussed previously, they do seem to throw up some unreliable figures. Concerning the failure to consistently apply differing going allowances for straight & round courses, consider the meeting at Newbury on 21 April this year.
My own & Topspeed ratings for the card were as follows, the first figure being my time rating, the second Topspeed:Diamond Tycoon 107 102
Maraahel 64 31
Pinpoint 89 85
Majestic Roi 89 89
Major Cadeaux 89 89
Junior 80 29
Red Gala 86 38.Based on the race times on the day, there was a clear variation in conditions between the straight and round course. However Topspeed applied the same going allowance to both, resulting in ratings for races over the round course being suppressed – erroneously in my mind.
Whilst it may seem that I’m being hyper-critical of Topspeed’s effort’s, the ratings are at least presented with the intention of providing a set of ratings that can be helpful to the user and I don’t envy being in the position of having to rate every race. The scribings of some other so-called ‘time experts’, though, are complete and utter codswallop and should be consigned to the waste bin. As a new user of this forum, I’m unsure if it’s possible to communicate offline but if anyone is interested in having a look at some of some of my ratings and associated comments from earlier in the year as the basis for discussion, please let me know.Welcome to TRF scallywag76.
If you wish to communicate with people in private then, just use the internal messaging system.
Regards – Matron
Scallywag,
The Topspeed ratings from that Newbury meeting do look a bit dubious, having looked at the card. There was no evidence of a dubious pace in the race won by Red Gala, although it was mentioned that R Hills slowed the pace appreciably on Maraahel in the 2.10. Races over 2 miles are usually steadily run affairs, so I would probably be prepared to accept that.
I think Topspeed got it wrong that day based on the Red Gala result, a fairly competitive handicap for its class of 0-85 which should have produced a better rating than 38 and led him to conclude that the round course was riding slower than the straight course.Most of the discussion on the forum is out in the open, which I think is generally the best way as it promotes more discussion. I also appreciate that no-one wants their research or work plagiarised, so occasionally offline communication is preferred.
Personally, I’m just about finished with flat racing this year – there’s not enough quality racing to keep me absorbed on the AW and the jumps are hugely entertaining but difficult to rate.
If you would care to keep the discussion ongoing on this thread, I would be happy to exchange views about races past and present from a speed angle. This part of the forum is probably the best place for it and others would certainly join in. If you would prefer to go offline with stuff you don’t want to publicise, I’m happy to do that, but I prefer the forum as a medium for debate.
Steady on Artemis, your almost saying that topspeed isn’t as good as it should be ..
Formath, why would anyone bother to base any selective system on a rating with such a huge overall negative expectation? Why not start off with something that breaks even?
Topspeed makes mistakes, like any other compiler of ratings. Scallywag has pointed out an obvious example.
I would say from my own experience of doing my own figures and then comparing them with Topspeed, the RP figures are within a few pounds of my own 95% of the time, probably because I use (almost) the same theoretical model. You could say that 5% are incorrect, which is an awful lot of races, but 95% accuracy is acceptable for something like speed ratings, as it is for much more important things like convicting murderers and some delicate surgical procedures. Mind you, I’m not saying that Topspeed and I are using the best possible method of determining speed ratings. There must be even more sophisticated techniques using sectional times and viewing each race several times to see the path each runner takes/interference etc., but that would be a full time job for several people.
I don’t think it’s mistakes so much, I think its the basic formula that’s poor. It’s hard for me to comment properly because I don’t know what the formula is. Do you have a copy of the formula?
When I did my own figures I was getting a strike rate of around 20% but not really getting enough of an edge to overcome the 5% commission.
It’s not too surprising, perhaps, that the Topspeed speed ratings (SR’s) don’t perform too well on an ‘all selections’ basis. Additional to the problems that I’ve highlighted with Topspeed, I also belive that the use of any SR’s in the same ‘deterministic’ way that form ratings are used is fundementally flawed. Consider, for example, horses that post their best SR, without actually winning the race in question. It seems to me that such horses can often be ‘towed along’ by a strong pace and are incapable, subsequently, of translating their effort into a winning performance. What, then, might be the best way to use SR’s?
In the first instance, the concept of calculating SR’s on a 0 to 140 scale, for use with a comparable set of form ratings allows horses that may be better than their bare form rating to be readily indentfied, e.g., SR 97, form rating 93, suggesting that the horse may be a step ahead of the handicapper and represents a decent bet. In the same way, it also enables the user to quantify the overall pace at which a race has been run, i.e., fast, truly-run, slow, etc., and in this context it pays to ignore the blindingly obvious. For example, this season’s Dewhurst, yet again, threw up the best juvenile SR, the Racing Post Trophy wasn’t far behind and what about that cracking maiden win earlier in the year by Diamond Tycoon? Unfortunately, the world and his wife are well aware of these performances. Delve a little deeper, though, and the effort may unearth some hidden gems. From my own analysis earlier this year:York 16 May 3:50, 12F handicap won by TCHERINA. Comparison 59 / 84. Slow. Pace possibly not suiting lightly-raced and stoutly-bred MISSOULA (8th); worth watching.
The ‘comparison’ figures are my SR and the RP / Raceform form rating for the winner. Missoula, given a test of stamina, certainly was worth watching – 2nd at16/1, wins at 5/1 & 8/1.
Incidentally, Artemis, the 2m event won by Junior at Newbury in April was a modest contest but my SR of 80 for the winner, against the RP form rating (at the time) of 86, suggests it was run at a fair pace. Horses that ran well there won and / or performed creditably on subsequent occasions, suggesting that it perhaps wasn’t the soft race that Topspeed might have led you to believe.
Finally, for what it’s worth, Malt or Mash posted a (possibly conservative) SR of 109 on my figures – RP form rating 105 – for his latest win at Newmarket, suggesting that he has ability bordering on Group class; a 6lb rise in the OR’s looks lenient. Shame, then, about his price in today’s November Handicap, although it’s worth noting that, based on my assessment of ground conditions, the going at it’s fastest has only bordered on good to firm for any of his wins. Perhaps Hannon’s caution about the ground today is justified but I guess we’ll know more at about 3:15 this afternoon!Good call on the November Handicap, although as you say not much of a price.
I’m not so sure about horses getting towed along at a fast pace and then not being able to go on and reproduce that form. A horse that runs a fast time and finishes close up should be able to reproduce the run given the same conditions. Sometimes, this should be good enough to win provided the horse has the ‘winning’ attitude and doesn’t hang in behind other horses.
The point about the relationship between RP rating and SR(Speed Rating) is a vital one. In fact, if the SR is greater than the RP rating in any race, in my book it means the horse has been underrated because the RP rating and SR both measure the same thing, namely the horse’s performance on the day. The RP is bound to be more reliable because it is fairly easily worked out for any race whereas many races are run too slowly to produce meaningful speed ratings.
Thanks – it’s nice when things work out well. Wouldn’t life be dull if that happened all the time?!
Re. the ‘towed-along’ idea, it’s just a personal viewpoint and not based on any hard facts. Even so, there must have been a good many horses that posted their best SR in a Derby or Oaks, for example, and failed to replicate that performance again.
What I was trying to say earlier was that it’s all too easy to become hung up on figures and you often have to look beyond them to derive the benefit of the hard work that goes into calculating the ratings in the first place.
I think that horses running very fast times are sometimes taken beyond their comfort zone to where it hurts, so that they are not so keen to run that fast again. This might explain why some horses record fast times in top class races, but don’t ever repeat that time in lesser events. A horse with a ‘winning’ attitude is arguably a horse that can cross the pain threshold without flinching as opposed to the type that travels well but will not exert itself when pressure is applied.
The analogy with human athletes running at middle distances and beyond is a good one. They reach a point where the body aches, for whatever physiological reasons, and they have to have the mental resources to go beyond this barrier if they wish to compete at the top level.
Temperament or courage is often inbred, but it may also be a matter of instilling confidence……. in humans and horses.
Sanbuch, 2nd at Newmarket to Malt Or Mash off a strong pace, meets the winner again at Doncaster on more favourable terms and yet finishes further behind. An example, perhaps, of a beaten horse being towed along (in the Newmarket race) and being unable to translate that into a winning performance?
- You must be logged in to reply to this topic.