- This topic has 101 replies, 8 voices, and was last updated 19 years, 4 months ago by
.
-
Posts
-
Artmis, A few questions;
What do you mean by Bath (RH)?
Raceform use Median Times are these adjusted to 9st or 10st?
Do Raceform work on a scale of 0 to 100 for their speed figs?
courtesy of Raceform(hope i’m not breaking copyright law)
4. Raceform Speed Ratings: The Split Second ratings are an expression of a horse’s speed in terms of lengths-per-mile as opposed to pounds in weight. Basically, if one horse has a top speed figure five points superior to another and both horses run to their best form in a race run over a mile, the first horse should beat the second by five lengths. In a race run over two miles, the margin should be ten lengths and so on.
<br>Before the speed figures can be calculated, it is necessary to establish a set of standard or median times for every distance at every track. Looking at the winning times of all winners over a particular trip going back several years (in order to get a large enough sample) and producing an average does this. A distance has to have a certain number of races run over it in order to produce a reliable median time. Otherwise, no median time is calculated and no speed ratings are produced for races over that trip.
Once a meeting has taken place, a ‘raw’ unadjusted speed rating is calculated for each winner on the card. This is done by working out how many lengths per mile the winning time was faster or slower than the median for the trip. A difference of 0.2 of a second equals one length. The raw speed ratings of all winners on the card are then compared to a ‘par’ figure for the class of race. Basically, winners of races of a certain class, A, B, C, etc have been allocated a par rating which should be achievable for the average winner in that class of race. The idea is that better-class horses should run faster compared to the median time than horses of  lower ability.
The difference between the ‘raw’ speed rating and the par figure for each race is then noted. The fastest and slowest races compared to par are discarded and the rest are averaged to produce the going allowance or track variant. This figure gives an idea as to how much the elements, of which the going is one, have affected the final times of each race.
The figure representing the going allowance is then used to adjust the raw speed figures and produce the final ratings, which represent how fast the winners would have run on a perfectly good surface with no external influences including the weather. The ratings for beaten horses are worked out by taking the number of lengths they were behind the winner, adjusting that to take into account the distance of the race and deducted that figure from the winner’s rating.
The speed ratings do not take into account the effect of weight, either historically or on the cards. That component is left completely at the user’s discretion. What is shown is a speed rating represented in its purest form, rather than one that has been altered for weight using a mathematical formula that treats all types of horses as if they were the same.
<br>Speed Plus this is a rating that is formulated as follows:-
The best Raceform (RF) rating for each horse is added to the best Split Second (Speed) rating. The total is divided by two an the highest overall rating is adjusted to 100. Only the top 6 qualifiers are displayed.
(Edited by empty wallet at 10:10 am on July 6, 2005)
Half way through paragraph two they switch from an average to a median!
After considering the various factors involved they weed out the highest and lowest figures and average what’s left. For reasons that I’ve gone into at length before and am not about to repeat now, this is not the best option IMO.
And let’s just not get started about the desirability of incorporating the apparent effect of weight in calculations……
Prufrock
Well spotted :biggrin:
btw from the ones i’ve checked, Raceform median times are not the same as the RP standards
Wallace,
Sorry, my mistake: should be Bath(LH); all the other courses mentioned are right-handed with uphill finishes.<br>Bath is left-handed with an uphill finish. I could have mentioned Pontefract which is also LH and uphill, but it is very stiff indeed, being uphill nearly all the way for the last 5f.
I’m not familiar with Raceform’s method of working out speed ratings. I’m studying empty wallet’s post.
ew,
I presume Raceform must produce a range of median times to represent the different grades/classes of race. Or, perhaps they have a table of set corrections in terms of lengths per second that can be applied to produce par times for different grades.
The idea of dropping the best and worst figures seems very arbitrary and unsatisfactory.
As I’ve stated, I don’t do my own figures anymore, but if I did in future I’d be looking at every race, handicapping it to a rating in a traditional way, and then produce a going correction from the race times and other evidence from all races at the meeting.
I’m not knocking Split Second because there are plenty who use that method with success and I don’t know enough about it. Personally, I find ratings on the same scale as the official one easier to visualise in terms of a horse’s ability and rate of progress. Also the Weight for age allowances(WFA) are easier to handle.<br>
- Total Posts 17716
The class of horse isn’t taken into account for most standards, the class of the race is, but I would suggest there are 2 separate entities.<br> The latter assumes that all class C races are the same, whereas a study of the former paints an entirely different picture.<br> As an example, taking the 8 class C races(excluding nurseries) run in the past 12 months at Ascot shows an average current rating for the winners of 96.9. The same exercise for Beverley shows an a.c.r. of 88.<br> Looking at the winners involved at Ascot shows the likes of Tartouche, Pivotal Point, Prince Aaron, Sergeant Cecil, Prince Of Thebes, and Khabfair,  good class horses who are mainly improving and moving up in class, while the best Beverley can muster is the decliners Bahamian Pirate and Proud Boast.<br> It is a fact that better courses attract better horses, either current or potentially, and that standard times that take no account of this, and the resultant speed figures, are, IMO, deeply flawed.
(Edited by reet hard at 5:35 pm on July 6, 2005)
I’m pretty sure that the RP standard times are related to horses and their RP ratings and not races, but other standard times might be race related.
Take Quito who is rated around 105 by RPR at 9st. He is a mature horse and given a decent pace, which he needs to produce his form, he will run very close to RP standard on good ground at any course whether it be Newcastle or Ascot or Nad Al Sheba, regardless of the class of race. Usually, the RP standard is only good enough to win top handicaps which he has done, but it can be good enough to take a listed race if he gets the run of the race and others run below form.
A race is after all only a collection of horses of varying abilities. Its the horses that set the standards, not the race IMHO.
- Total Posts 17716
Ec<br> It is some time since I used speed figures as a major part of my betting, though I do recall doing a similar exercise some time ago concerning class C’s at Ascot and Pontefract, and, using 3 sets of ratings, the Ascot races came out around 6lbs better, on average.
Artemis<br>Quito, although an 8yo, is almost certainly still progressing, a view supported recently by his trainer.<br> It is interesting to note that, prior to his Ayr Gold cup win, he had never  won above cl3 or at a gd1 track, despite contesting numerous races at those levels.<br> Since then, all of his wins have been at or above that class, and 4 out of 6 at gd1 venues.  All of which suggest that as a horse progresses, the grade of track he runs at will also progress, whatever his OR.<br> As I said earlier, I am no expert on s/f’s, though, on gut instinct alone, I would be reasonably certain that 5f at Newmarket is a much sterner test than 5f at Catterick, regardless of what the clock says.
(Edited by reet hard at 9:39 pm on July 6, 2005)
Interesting stuff chaps
<br>Class C Hcps/1mile/Good
average time/average OR/average SF
Ascot <br>1760yds 101.39  (78)    104  RFR 87<br>Pontefract <br>1764yds 104.21  (76)    105  RFR 84<br>Beverley <br>1860yds 106.72   (89)    106  RFR 101
<br>Adjusting for the extra 100yds(Bev) 101.02,4yds(Pont)103.97
Further adjustment for the bends at Pontefract and Beverley may be needed,but the main adjustment needed  is the very stiff 3f uphill finish at Pontefract
AS with the different sharpness of  bends,the different gradients in uphill(or downhil) finishes will effect racetimes
<br>The "precision engineer" figure compiler maybe will/should take these different grandients into account via sectional time analysis
(Edited by empty wallet at 5:48 am on July 7, 2005)
- Total Posts 17716
s
It would be useful if the RP did a feature on how they have determined the standard times.
We know by their definition that it is the time expected of a mature horse(4.5+), rated 100, carrying 9st in a fairly run race with zero going correction.
I have taken the ‘rated 100’ to be the RPR, but I think to make sense it has to be the Topspeed rating because the standard is used to calculate the speed ratings. I have used the RPR, but I keep coming out about 10lbs too high, which is roughly the average difference between the two figures.
- Total Posts 17716
Quote: from EC on 8:40 pm on July 7, 2005[br]just using 5f standards alone should be a good measure of the stiffness of a course.<br>I would seriously doubt that that is the case. Any measure that has Catterick equally as stiff as Newmarket has to be highly questionable, on the grounds of course conformation and quality of opposition.
I wouldn’t be convinced that different class C’s would differ greatly Reet Hard..there are improving winners even at below grade 1 tracks..as there are also exposed winners in class C’s at Grade 1 tracks. I feel it would even itself out..basically comparing standards by bringing times in line with one class of race will tell you which track is the easiest/stiffest.<br>Again I would disagree. From the 4 class C h’caps run in the past 12 months at Beverley, the current ratings of the 4 winners are respectively 61,90,84 and 90. The same exercise at Ascot shows subsequent ratings to be 96,115,103,84,93,92,100 and 92, in my view, a totally different picture. It would also be as well to consider the level of opposition at each track would be similarly skewed.<br>I haven’t a clue how RP calculate their standards but I know how I do mine and have every confidence in them..as we don’t know what races are used to produce these standards it’s pretty difficult to trust what they tell us.<br> I,equally, haven’t a clue, although I am fairly certain that they used to be an average of the 10 best times over each distance.<br> I do though, have the utmost confidence in my original statement that the class of a race cannot be defined simply by its title. ÂÂÂ
(Edited by reet hard at 12:57 pm on July 8, 2005)
The RP standards do give a clue to the nature of the track: stiff, easy etc, but what they do not do is give any clue to the class or quality of horses running at that track, unless you compare RP standards with track records. Then you will discover a general rule:
‘the better the class of horse running at a track, the greater the difference between the track record and RP standard’
There are some tracks where the track record is actually above the RP standard because no horse rated above 100 (135 jumps) has actually ran there in a true run race on fast ground in order to break the standard and thus the record.
An example(well, nearly) from today is Wolverhampton where the sprint track records are very close(just below) the RP standard and the 14f record is above RP standard. Compare that with York where the records are comfortably below the standard.
If course standards were proportional to class of race, the gap between track records and standard times at different courses would be consistent. It isn’t.<br>
Ran a little query
Average time of 5f class D+ hcps on good
Catterick 60.28 every race class D<br>Ascot 61.31 only 1 race class D<br>Newm 59.90 only 2 races class D
<br>make of it what you will ;)
(Edited by empty wallet at 7:29 pm on July 8, 2005)
EC
<br>I’d tend to agree with your above post,afterall, the TS,SS or Rating is only the method or opinion of the person compiling them<br>And method/opinion can vary greatly on the worth of one horses form,nevermind when that horses form is compared to or against other horses
Great, in it<br>
ew
I don’t think there is much variance between different methods of assessing racing form in terms of lbs, distance and time. Most of the work in this field has already been done and is generally agreed. There is plenty of scope for arguments about the effects of track configuration, pace and going on the outcome of a race.<br>And, of course, the relative merits and honesty of owners, trainers and jockeys. The rest is in the hands of the Fates, chaos and random chance.
- You must be logged in to reply to this topic.