I have been developing a ratings system on and off over the last 17 years and have finally created an application that is reasonably automated and therefore easy to use and apply.
My main question now is - Over what period of time/number of meetings should I rate races before I get a good indication of the rating method’s success ?
Thus far I have rated 670 races over 120 meetings since May this year. They have been virtually all Saturday meetings in Brisbane, Sydney, Melbourne, Adelaide and Perth. I have applied reasonably strict filters to come up with the following betting strategy.
Win bets – 51 for 22 wins giving a strike rate of 43% and a return of $1.70 for each $1.
Place bets – 51 for 32 wins giving a strike rate of 63% and a return of $1.08 for each $1.
Other bet types (Quinella and Trifecta) show profit, but require more analysis.

My own initial thoughts indicate I would need to rate more races given the low selection rate. Thanks for any comment you may like to provide.


_________________
_Burdekin_

[ This Message was edited by: Burdekin on 2003-12-19 13:28 ]

I'm not sure, but it sounds pretty good so far.

Can I ask, did you have any big priced winners or were they all basically around the 3,4,5 dollar mark???

[ This Message was edited by: sportznut on 2003-12-19 13:47 ]

Initially I saw the 670 and 120 figures, I thought you had a solid base of results, but then the 51 bets rang alarm bells.

It seems like you have a very good method of ratings, but base your stats on the actual bets - 51 is not nearly enough.

I'd be looking at around 500 bets to have any confidence in it.

I've been exactly where you are now and thought I was doing great, but the next 60 bets went downhill fast.

good luck Burdekin.


[ This Message was edited by: Chrome Prince on 2003-12-19 13:54 ]

Hello Burdekin,

in my opinion, you have two different ideas in your post.

1. Your ratings, where you are rating each horse in each race, and
2. A "system", where you put your ratings through a series of filters to produce a bet.

51 bets is nowhere near enough to base a solid judgement on, but it is a start. Monitor it for the next 50-odd bets at least to see how it goes, and then you might be in a position to start making judgements.

I don't like the ratio of 22 wins and 32 places. I would normally expect to see placings to be around twice the wins.

As for the ratings, how do they perform if you were to back every single top rated for the win? 670 races is a reasonable sample here, and I would be seriously monitoring this to see if the basis of any systems you build on them are sounds.

Cheers,
Chris.

__________________
Permanence is an illusion

Thanks for the quick response !

I will answer each of you in turn.

Sportznut,
I did turn up some big priced winners eg. Warrior Trader at Eagle Farm paying $14.40 on May 31 and Poachers Luck, also at Eagle Farm on October 18 paying $20.20. Had 4 others between $3.00 and $6.00. The rest under $3.00. I understand the big prices raise the returns, but they did rate and they did win.

Chrome Prince,
The selection of the top 50 appears to give me the highest returns, but subjects the races to the strictest filter. To get to 500 would take me 10 years, I have patience, but that may test it.
What I have provided below are results based on the top rated bets based on the 670 races I have rated -
Win
Bets/Wins/S.R./Return
498/119/23.90/1.04
405/102/25.19/1.07
307/87/28.34/1.14
206/62/30.10/1.15
103/37/35.92/1.45
51/22/43.14/1.70

Place
498/253/50.80/0.98
405/206/50.86/0.98
307/161/52.44/0.99
206/119/57.77/1.07
103/62/60.19/1.06
51/32/62.75/1.08

As you can see, the returns diminish as I include lesser rated runners.

Stebbo (Chris),
Too true, the ratings has been an ongoing project, being subject to many reviews. It appears to give good results, so I have moved to the next conundrum, how to punt to it.
The win/place ratio is disappointing, but if you look at my reply to Chrome Prince, you will notice it improves as more bets are included.

On the 670 races
Win
Wins : 137 S/R : 20.45 Return : 0.91
Place
Wins : 323 S/R : 48.21 Return : 0.99
As you can see, I need to filter the selection of my rated runners to produce the goods.


Once again, thanks for you comments, I have gathered some good info and will provide an update in a month or so.


_________________
_Burdekin_

[ This Message was edited by: Burdekin on 2003-12-19 18:27 ]

[ This Message was edited by: Burdekin on 2003-12-19 18:30 ]