KennyVictor
8th March 2006, 11:15 AM
No, don't get your hopes up I haven't found a way to che&t the bookies. If you get through this though there may be something worthy of your interest. I've capitalised the juicy bits for those with a short attention span.:)
I've noticed as part of some systems on here that people recommend things like "Bet on the favorite but only after a favorite has won at that meeting". Other's have said "Bet in such a fasion but stop after the first win". These two almost opposite approaches, both possibly valid in their own context, interested me but weren't of any immediate use so I stored them away at the back of the mind until later.
Anyway, I've been betting quinellas lately and noticed that wins seemed to be clustering within meetings. If you had five wins across 7 meetings they wouldn't be evenly spread there would often be a couple of meetings with 2 wins or maybe one meeting with three. So I decided to investigate. I ran my system as it would have worked over races since 1996 to find the returns I would have got and then ran it again only betting when a quinella had turned up at the meeting.
NSW Races showed an increase in ROI from 117% to 123% (100% means getting your money back, 200% means doubling your money).
WA Races went from 119% to 121%, QLD from 116% to 126% and Vic races from 114 to 119%.
So, a definite improvement on four independant sets of results. Interestingly the strike rate in three of the four actually went down a little so the dividends were improving at the expense of strike rate. Also the returns on races since January 2006 were much better in 3 of the four sets of figures which is probably why I noticed the clustering.
Encouraged by these results I decided to see if it worked on win bets too. These figures by the way are based on my ratings, I'll get to some figures on favorites later.
On win bets across the four states I didn't seem to improve much. WA went up a couple of percent, NSW 1% and Victoria down 2%. Again though with all results the strike rate went down. 1% less winners in all states except WA where it was about 0.1%. Divies were compensatingly higher after the first winner so I was making about the same ROI but just betting less.
I was intrigued by this point as to why the results were as they were. Why did the strike rate always seem to go down and why weren't my winners improving if the quinellas did. So I decided to analyse the same thing betting on favorites.
The program I have which analyses bets on favorites isn't as versatile as my other program so these results are only done on NSW races but there are over 70,000 of them so it's still a pretty valid sample.
If you were to bet on the SP favorite at all NSW races you would score 34.35% winners, 33.32% if you only bet on races where the NSW tab recognises the meeting. If you wait until the first favorite has won at the meeting you would only score 33.14% and 32.03% winners respectively. Again the strike rate goes down by 1%. The ROI from these bets drops from 89.47 to 88.58 for all races and 89.84 down to 89.55 for Tab Races. So, YOU DON'T IMPROVE YOUR RETURN ON FAVORITES BY WAITING FOR THE FIRST FAVORITE TO WIN.
Next I tried the other end of the scale. What happens if you bet on the longest priced (SP) horse on all races and after the longest price horse has won.
Well, at all meetings your Return from backing the longy is about 36% of your cash and at Tab meetings about 44% of your cash. If you wait until the longshot has won a race and then bet those that run after it rises to 43% and 56% respectively. Even more significantly the strike rate IMPROVES after the first longshot has won, the only time that happens in all my sets of figures. The number of bets is drastically reduced, it's a long wait for the longest priced horse to win at a meeting before you make a bet but: IF YOU'RE BETTING LONGSHOTS YOU HAVE A BETTER CHANCE OF WINNING AFTER A LONGSHOT HAS ALREADY WON AT A MEETING.
Finally, I'm a programmer and I know we put bugs in programs just so we can make money from doing software support so I ran another test to see if the program seemed to be working. Besides I was still intrigued about the strike rates falling consistently after a winner.
I did the test (on NSW races) using the earliest and latest horse in alphabetical order as a sort of random horse thingy. On both start of the alphabet horses and end of the alphabet horses the strike rate dropped from about 12 to about 11.5 percent winners if you waited until one of their fellows had won a race before having a bet. The ROI didn't really change much (up a little on one, down a little on the other).
So what did I learn from this? I suddenly thought that the strike rate must be dropping because there are more horses in races later in the program although the ROI on a random horse stays about the same.
OK, Draw what conclusions you like from these results (if any). I found them interesting and I've concluded that:
1. Favorites win at the same rate independant of whether a favorite has already won at a meeting.
2. If a longshot has got up there is now a better chance of a longshot getting up from now on.
How do I explain the quinella improvement but little win tote improvement which started this whole train of thought? Well if lesser fancied horses are winning in clusters that is more likely to affect exotics where you are relying on the performance of multiple shorter priced horses (i.e. there is more potential for a longshot to find its way into the first two or three).
I'm going to try to attach a results sheet of the full results but don't know how I'll go. Nope, it's 40K so it's too big - I'll email it if anyone wants it.
KV
I've noticed as part of some systems on here that people recommend things like "Bet on the favorite but only after a favorite has won at that meeting". Other's have said "Bet in such a fasion but stop after the first win". These two almost opposite approaches, both possibly valid in their own context, interested me but weren't of any immediate use so I stored them away at the back of the mind until later.
Anyway, I've been betting quinellas lately and noticed that wins seemed to be clustering within meetings. If you had five wins across 7 meetings they wouldn't be evenly spread there would often be a couple of meetings with 2 wins or maybe one meeting with three. So I decided to investigate. I ran my system as it would have worked over races since 1996 to find the returns I would have got and then ran it again only betting when a quinella had turned up at the meeting.
NSW Races showed an increase in ROI from 117% to 123% (100% means getting your money back, 200% means doubling your money).
WA Races went from 119% to 121%, QLD from 116% to 126% and Vic races from 114 to 119%.
So, a definite improvement on four independant sets of results. Interestingly the strike rate in three of the four actually went down a little so the dividends were improving at the expense of strike rate. Also the returns on races since January 2006 were much better in 3 of the four sets of figures which is probably why I noticed the clustering.
Encouraged by these results I decided to see if it worked on win bets too. These figures by the way are based on my ratings, I'll get to some figures on favorites later.
On win bets across the four states I didn't seem to improve much. WA went up a couple of percent, NSW 1% and Victoria down 2%. Again though with all results the strike rate went down. 1% less winners in all states except WA where it was about 0.1%. Divies were compensatingly higher after the first winner so I was making about the same ROI but just betting less.
I was intrigued by this point as to why the results were as they were. Why did the strike rate always seem to go down and why weren't my winners improving if the quinellas did. So I decided to analyse the same thing betting on favorites.
The program I have which analyses bets on favorites isn't as versatile as my other program so these results are only done on NSW races but there are over 70,000 of them so it's still a pretty valid sample.
If you were to bet on the SP favorite at all NSW races you would score 34.35% winners, 33.32% if you only bet on races where the NSW tab recognises the meeting. If you wait until the first favorite has won at the meeting you would only score 33.14% and 32.03% winners respectively. Again the strike rate goes down by 1%. The ROI from these bets drops from 89.47 to 88.58 for all races and 89.84 down to 89.55 for Tab Races. So, YOU DON'T IMPROVE YOUR RETURN ON FAVORITES BY WAITING FOR THE FIRST FAVORITE TO WIN.
Next I tried the other end of the scale. What happens if you bet on the longest priced (SP) horse on all races and after the longest price horse has won.
Well, at all meetings your Return from backing the longy is about 36% of your cash and at Tab meetings about 44% of your cash. If you wait until the longshot has won a race and then bet those that run after it rises to 43% and 56% respectively. Even more significantly the strike rate IMPROVES after the first longshot has won, the only time that happens in all my sets of figures. The number of bets is drastically reduced, it's a long wait for the longest priced horse to win at a meeting before you make a bet but: IF YOU'RE BETTING LONGSHOTS YOU HAVE A BETTER CHANCE OF WINNING AFTER A LONGSHOT HAS ALREADY WON AT A MEETING.
Finally, I'm a programmer and I know we put bugs in programs just so we can make money from doing software support so I ran another test to see if the program seemed to be working. Besides I was still intrigued about the strike rates falling consistently after a winner.
I did the test (on NSW races) using the earliest and latest horse in alphabetical order as a sort of random horse thingy. On both start of the alphabet horses and end of the alphabet horses the strike rate dropped from about 12 to about 11.5 percent winners if you waited until one of their fellows had won a race before having a bet. The ROI didn't really change much (up a little on one, down a little on the other).
So what did I learn from this? I suddenly thought that the strike rate must be dropping because there are more horses in races later in the program although the ROI on a random horse stays about the same.
OK, Draw what conclusions you like from these results (if any). I found them interesting and I've concluded that:
1. Favorites win at the same rate independant of whether a favorite has already won at a meeting.
2. If a longshot has got up there is now a better chance of a longshot getting up from now on.
How do I explain the quinella improvement but little win tote improvement which started this whole train of thought? Well if lesser fancied horses are winning in clusters that is more likely to affect exotics where you are relying on the performance of multiple shorter priced horses (i.e. there is more potential for a longshot to find its way into the first two or three).
I'm going to try to attach a results sheet of the full results but don't know how I'll go. Nope, it's 40K so it's too big - I'll email it if anyone wants it.
KV