- This topic has 8 replies, 6 voices, and was last updated 14 years, 11 months ago by
.
- Posts
Not sure if this subject has been discussed before but has anyone ever investigated into "the bounce" and decided whether such a condition actually exists?
After Walkon bounced at Newbury on Friday I would be interested to think would any trainer use this term as a cop-out for not having their horse fit enough to run.
Hi Phil, I did do some research on this a couple of years ago using Adrian Masseys website, I haven’t still got my findings on my pc but from memory it certainly was a factor for jumps racing but not so much on the flat
- Total Posts 17716
I can tell you about football bounces.
The biggest in history is Germany v.Hungary
3-2 in 1954 (world cup final).
Equally ranking in my books was the Germany v.France
1-0 in the 1986 semi in Mexico.
We see this happening again and again but every time the punters are left gasping.
On both those occasions the exertion was too much and the teams bounced. The culprits who in my examples were Germany on both occasions had some very good players but not quite the same class as their opponents of 1954 and 1986.In horse racing the same thing is bound to happen:
Two-three consecutive good performences followed by an abysmal failure.
The question is how to predict it’s going to happen.
One obvious clue is to look at the last performencesand
the calendar days separating them and take defensive measures.
Don’t be a daydreamer and always pay attention.The condition most certainly does exist.
The trainers sometimes fall victim just like we do, but on other occasions they just want to fill the card and produce tired horses, known to them but unknown to us.Funnily enough this was just being discussed in another place. As I originally understood it , when it first came to my attention in the mid nineties ,the bounce factor was an American term. It refers to 5f and 6f sprinters ,due to their running style . Who after posting a lifetime best performance when coming back from a long lay off , almost always under performed next time out ( bounced ). The thinking at the time was that it didn’t affect middle or long distance runners who were mostly settled off the pace. There was a theory I think put forward by Nick Morden – possibly in Odds On magazine ? That due to the exertions involved and frantic pace , the same theory could be applied to 2m chasers . Nowadays the term seems to have lost all sense of it’s original meaning and simply covers a multitude of sins.
- Total Posts 17716
I can have a look at my racing archives.
I dispute the second after layoff theory – in my opinion on such occasions an improvement is on the cards, if anything.
Sprinters more susceptible to it, maybe.
How do you suggest I go about it ?
After all a series of wins is always going to come to an end even without rise in class.
What is the method ?
Ultimately horses performences are predicted from their speed histograms. If a bounce is considered very likely then you should add a handicap value.
Speed histograms corrected for bounce effect will be good to have.I remember reading extensively about the "bounce" many years ago, can’t remember the sources or the authors, but were nearly all USA based, and, as alluded to earlier, referred mostly (but not exclusively) to sprinters. A statistical analysis of US sprinters affected by the bounce would be much more meaningful than one produced over here. For a start, many of the horses would reappear soon after a good performance and on the same track (and thus going/surface)as many US meetings raced on a large number of consecutive days. Not the case here, so that if a sprinter had a hard race and ran again a few days later and disappointed, it would usually be on a different track; there would be more "x" factors than just "bouncing". (Going, unsuited by next track, travelling, maybe different jockey etc.)
Whilst there’s certainly a potential for some horses under some circumstances to "bounce", I’m sure we can all recall horses posting succesive good performances in a short time span (ie not bouncing).
Oddly enough, soccer is a medium that the "bounce" does become a gambling factor worth consideration, but not perhaps in the manner one might expect.
In the days when "fixed-odds" betters had to include a minimum of 5 selections if a LEAGUE home win was selected, many punters used to stuff /pad-out their selections with the long-odds on shots each week (Celtic / Rangers / Top English teams). This was tantamount to chucking money away. During the 70s, 80s and ealy 90s I used to make a regular and fair profit (you’ll just have to take my word for it) by eschewing the long-odds shots. Chalk-eating (as the yanks call backing odds-on shots) was the road to ruin EXCEPT (and this is the football "bounce"), in cup replays when a top team had been held to a draw in the first match (or leg). In the replay (or second leg), even though the top team would still be odds-on, they would have a much higher success rate. This could be put down to 2 things. First, the underdog team had "bounced" and were incapable of repeating their first match performance and secondly, the top team would not be guilty of complacency after the first game. In those days, replays were nearly always played relatively soon after the initial game. Only under such circumstamnces was it worth backing a clear odd-on fav. Of course, as they were cup games, one could place a single on them, but quite often you could pad-out midweek league matches with such games.
Not so sure the situation is the same nowadays as I seldom bet.- Total Posts 17716
Football is a different sphere of things. I made some remarks for generality’s sake.
In horses suppose we have two speed values t1 = last race and t2 = last but one.
If t1, and t2 are aver different distances eg. 5f and 7f and the next race is over 6f, there are conversion formulas.
What is the best statistical predictor for the next race ?
It’s not (t1+t2)/2.
It’s something like t = (w1*t1+w2*t2)/(w1+w2) where w1, w2 are weight factors and w1 (the most recent) is > w2.
Could be w1=1, w2=0.6.
There are also the so called excusable preformences like when the horse came back with a nose bleed.
Most people say t = best of two, which is wrong.
Or they say t = average of the two but if last is best the t = last. Again wrong.
If you compute the weights w in such a way as to minimize the variance then you reach something and the bounce effect is at least partially taken into account.
But again if you do it using linear regression it’s another mistake. You end up with all the horses on the level and you are underpredicting the winners. You need stochastic regression -a more advanced statistical method.
- You must be logged in to reply to this topic.