reading some of the methods posts, it seems that some people aren't too keen on methods that are based on last start winners. So to stir the pot I thought I would start a discussion about them.

Here is a method I have been using and I have tested over the past five years. My results are only in Sydney as that is the only venue I bet, others might like to try the system over a larger data sample.

It is quite a simple system but it is very selective, which I don't mind as I like to run several systems at once.

Here goes

SYSTEM SELECTOR RULES SET FOR: G LSW
Printed by Price Predictor at 11:55:05 on 07/02/2004

Race Distance: 1100 - 2000
Horse Win %: 20 - 100
Horse Place %: 80 - 95
Career Starts: 5 - 999
Days to Last Start: 7 - 21
Days to 2nd Last Start: 1 - 42
Last Start Finish Posn: 1 - 1
2nd Last Start Finish Posn: 2 - 24

results for sydney course from 6/2/99 to 20/12/2003

Meetings considered : 259
Win Strike Rate/Seln. : 29.7%
Plc Strike Rate/Seln. : 53.8%
Average Win Dividend : $3.87

WIN PLACE
Races Bet : 89 89
Races Won : 27 49
S.R./Race : 30.3% 55.1%
Outlay ($): 91.00 91.00
Return : 104.50 87.90
$ Profit : 13.50 -3.10
% P.O.T. : 14.8% -3.4%

cheers

dink

__________________
Before you criticize someone, you should walk a mile in their shoes. That way, when you criticize them, you're a mile away and you have their shoes.

sorry full results

WIN PLACE QUINELLA EXACTA TRIFECTA FIRST FOUR
Races Bet : 89 89 89 89 89 65
Races Won : 27 49 32 21 19 11
S.R./Race : 30.3% 55.1% 36.0% 23.6% 21.3% 16.9%
Outlay ($): 91.00 91.00 461.00 463.00 1902.00 4284.00
Return : 104.50 87.90 546.20 530.90 2850.80 10646.50
$ Profit : 13.50 -3.10 85.20 67.90 948.80 6362.50
% P.O.T. : 14.8% -3.4% 18.5% 14.7% 49.9% 148.5%

__________________
Before you criticize someone, you should walk a mile in their shoes. That way, when you criticize them, you're a mile away and you have their shoes.

Due to my small database of 8 mths I can't provide much except encouragement. But for SMBA, days 1-7 inclusive, with the additional rules of 1)56kg or less
2) 2-6 sts since spell
I get 22 bets,7 wins for POT 193%. This will probably get ripped to pieces by those with the large databases. With my additional rules there is one runner tomorrow 14/2 Flem 7, Miabondialee

No cigar on my figures, dink, these were my returns Saturday only, metropolitan everywhere back to '99.
Although I replicated your Sydney profit, that did rely on a couple of 10/1 winners which spread over 5 years is a touch thin.


Races Bet : 262
Races Won : 64
S.R./Race : 24.4%
Outlay ($): 268.00
RETURN : 245.90
$ Profit : -22.10
% P.O.T. : -8.2%

Your filters helped a bit, Sobar, but not enough.

I am a bit fascinated with last start winners but have never been able to make up anything mechanical which is very successful.
There is one thing I run which throws up a lot of bets and is marginally in profit, I will dig up the rules and post it later, perhaps it can be curve-fitted a bit more!

However, things like eyeballing the manner in which they hit the line provide indispensable but unquantifiable data.
Maybe sectional time analysis can substitute, to a degree, this subjective information but is still unhelpful for ppp purposes.

Hi All,

I've run no less than 3 different "systems" based upon last start winners, and all 3 of them have failed miserably... One in particular went great guns for almost 6 months, then fell in the biggest hole you'll ever see, completely breaking a 50 unit bank....

Needless to say, I don't like them anymore :sad:

Cheers,
Chris.

Well here's my last chance at glory. My latest amendments are
1)SMBA
2)Days 1-7
3)Race Dist 1100-1799
4)Race Class 0-57
5)Win% 21-57
6)Sts since spell 2-6
7)weight 48-56kg
last start SP 0-11
9)weight change -3.5-+2.5

Rule 8 is cool man!!!

Rule 8 may well be cool but I want to see if the whole thing is hot. Can somebody find it in the heart of their bottom to run my rules in conjunction with Dink's rules above. Some I've modified and some I've added. I'm not trying to steal Dink's thunder so much as trying to find if I have anything worthwhile to contribute with my small database.
Probably the results were so embarrasing that everybody is just being polite.

Sobar,

These are my results for your criteria run over all my metropolitan data:

Meetings considered : 2079
Win Strike Rate/Seln. : 30.6%
Plc Strike Rate/Seln. : 51.8%
Average Win Dividend : $5.21


Races Bet : 85
Races Won : 26
S.R./Race : 30.6
Outlay ($): 85.00
RETURN : 135.40
$ Profit : 50.40
% P.O.T. : 59.3%



Here are the same rules except substituting win % 20 -60 which is a little less contrived and potentially more robust:

Meetings considered : 2079
Win Strike Rate/Seln. : 25.6%
Plc Strike Rate/Seln. : 51.9%
Average Win Dividend : $5.48

Races Bet : 128
Races Won : 33
S.R./Race : 25.8%
Outlay ($): 129.00
RETURN : 181.00
$ Profit : 52.00
% P.O.T. : 40.3%

Both runs' profit depended on 2 or 3 longshot winners and I don't think I would be holding my breath for repeatability.


Keep em coming.

Tony

Ok thanks Tony. There are just a couple of things, and I bring them up for debate more than anything. I know where you are coming from with the 60% figure because I often feel the same way looking at other systems with unusual cutoff figures. But on the other hand why should a nice round figure command any more respect. I mean the great pyramid of Giza had to slope at 51 deg 52', not 50 deg. As for the longshot winners. Well if they came before my test period of June 2003, I'm prepared to claim them, but yeah, talk can be very cheap in this regard. We'll see what happens.