Hey,

A lot of systems use "Days since last run" as a filter to their systems. Personally I haven't ever used it and I have never seen any stats on it.

Does anyone have any stats on the following :

Winners avg days since last run ( and broken into subsections ie 1-7 days, 7-14 days, 14-21 days, 21-28 days, 28+ days )

Average days since last run for all horses.

In particular I am interested to see if there is the same % of winners in each subsection as taken in relation to the % of horses running in that subsection.

An example might help:

1-7 days has 2300 Winners 10000 horses ( 23% Winners )
7-14 days has 460 Winners from 2000 horses ( 23% Winners )
14-21 days has 200 Winners from 1000 horses ( 20 % Winners )
21+ days has 50 winners from 100 horses ( 50% winners )

The details above are just made up to show what I am looking for. In particular my figures (which were made up) would show it is better to have a break of 21+ days.

If anyone has any details it would be greatly appreciated. I want to know whether or not the factor is valid.

Thanks

Hi Wesmip1

This is taken directly from an online database for UK racing. While not strictly applicable to Australian racing, race horses are the same all over the world. Its the training methods that are different.

'Days since' are generally used as an indication of racecourse fitness. Reasearch in a book published in the eighties (The Punters Revenge) indicates that a horse with a bad recent run is more likely to lose, where as a hores with a good recent run is more likely to win, and concluded that 'days since' as a varible on its own was not much use unless linked with the horses last performance.


1997-2006
Days..............Bets....Wins......Strike Rate
1 or 2............5043....623.......12.4%
3,4 or 5.........19877...2383.......12.0%
6,7,8 or 9.......59665...5964.......10.0%
10 to 14.........85599...7813........9.1%
15 to 19.........63625...5960........9.4%
20 to 25.........53628...5095........9.5%
26 to 31.........31109...2763........8.9%
32 to 40.........26661...2262........8.5%
41 to 60.........27170...2284........8.4%
61 to 90.........14897...1111........7.5%
91 or more.......48385...3364........7.0%
Not Run Before...34866...2263........6.5%

Hi Dolus,
The figures you show there are the sort of thing that make statisticians exited. There is a clear trend towards the more recent the horse ran the better its chance of winning.
When trying to form a correlation myself in the past though I have to say that as your quoted research said there wasn't much to be gained from just looking at the days since a race. The only clear conclusion I came to was that the horse was at an advantage if it had raced in the last 6 days. I would suggest this is because it shows the horse to be fit and showing no ill effects from its last race.

KV

Hi,
IMHO I have found it worthwhile to convert each performance for each runner to a Z score rating, then carryout the same analysis re Days from Last Start, further analysis could be by Subsets of Class, Distance and trainers, then average the scores in both the +/- groups and this will provide a performance envelope.

All,

thanks to everyone who responded. Especially to Dolus for providing the figures and to KV and woof43 for some ideas.

Thanks again.

In the past when betting on metro racing on Saturdays, I found that there was a definite improvement when sticking to only horses that had raced at a metro track exactly 7 or 14 days ago. Now that there are strong metro meetings being held on days like Friday and Sunday, I feel this is possibly not so much the case now.

Just following on from that theme, here's something you could think about:

1 - Saturday metro races only (Bris, Syd, Melb, Adel, Perth)
2 - Races of 1200m or more
3 - If there is only 1 horse in the race which last started exactly 7 days ago, it becomes the selection, providing it won or was placed at that last start.
4 - If there are no horses in the race which last started exactly 7 days ago, then look for horses which last started exactly 14 days ago. If there is only 1 in the race, it becomes the selection, providing it won or was placed at that start.

Yesterday's results:
BR7-8 Babinda Half UNP
SR4-1 Fooram UNP
SR5-2 High Priority WON $4.80/$1.80
SR8-4 Kincharm UNP
MR2-10 Ulfah WON $4.80/$1.80
MR9-5 Tesbury Jack WON $9.60/$2.20
PR1-2 Graduate Boy WON $2.00/$1.30
PR6-5 Wave Rock WON $2.40/$1.40

Yesterday's results may be a flash in the pan. Who knows? Anyway, whatever the case, after looking at the results of all metro races from yesterday, I have to say that what I thought many years ago still seems to be pretty close to the mark. That is that if you bet on Saturday metro meetings and concentrate mainly on those which last ran exactly 7, 14 or 10 days ago, you shouldn't go too far wrong.

interesting guys, sportz i suspect the apparent success of following sat metro runners those whose last starts were exactly 7,14 days would be because the nag is racing in similar class to last start, has had no real setbacks injury wise etc & with several starts under the belt has attained the fitness necessary to perform well.

just an afterthought, perhaps the results could be improved by factoring in a number of starts since a spell, maybe 2 or 3?

Yes, I would recommend 2+ starts from a spell.

I also think I would perhaps make another change to those little rules I mentioned previously. Instead of the horse having to run a place last start, make it simply one placing from the last 2 starts. That gives us a chance at getting a few roughies home with horses who disappointed last start, but are backing up quickly. They sometimes get up at nice odds.

1 - Saturday metro races only (Bris, Syd, Melb, Adel, Perth)
2 - Races of 1200m or more
3 - If there is only 1 horse in the race which last started exactly 7 days ago, it becomes the selection, providing it has had at least 2 starts since a spell and has been placed at least once from it's last two starts.
4 - If there are no horses in the race which last started exactly 7 days ago, then look for horses which last started exactly 14 days ago. If there is only 1, it becomes the selection, providing it has had at least 2 starts since a spell and has been placed at least once from it's last two starts.

Over the past couple of weeks, I think those rules would have produced 13 selections for 6 wins and a win return of $37.40.