Systems question

[Punter76]

Question for people with a successful long term system, say over 250+ bets.
What kind of place ratio.. ie 1-2-3-dnp are you achieving?

Because I believe I am backfitting to where the profit came from in my systems, and as a result am left with a very high ratio of winners.

So I'd say I've been backfitting to achieve the profit historically and as a result have kept the selections that have had a good winning run, and that winning run will not continue, and it never does.

Which is why it is so important to give any system at least 500 bets before handing over your hard earned, and having a high ratio of place getters too to comfirm the system is on the right track.

So what kind of places %'s on win only systems, ie 2nds and 3rds to winners are you achieving?

Any thoughts,

[Grub]

From my (limited) experience it seems to be about double P76. And I have heard many say 500+ selections is there benchmark as far as being comfortable it may continue to perform. Think youre on the money there but in saying that I wouldn't drop something promising purely because the number of bets isn't there.

To me there is 2 types of mechanical systems - Longshot (Lower SR but potentailly higher POT) and High Strike Rate (high POT unlikely). You can't have both over a decent period despite what you might get sent in the mail....

The better longshot systems I've come accross have around 20-25% SR for the win and 50% SR for the place. So your backing something that is usually at a decent price (each way) and half the time your getting a return. Most I've spoken to use these types of systems as bit of fun, small bets, as a means to have enough bets for interest and to stop from chasing more bets .... having bigger bank/bets on the higher SR systems.

Of those systems the better ones I've come across tend to be mid to high 30's% SR and around 65 - 70% place SR. So again about double. All sounds nice but there is still decent runs of outs that test the metal. If you have a system of this ilk you can try running place doubles or trebles but it takes more time and effort on a Saturday afternoon and although the outs may be less so might the return.

You need to find something that suits you personally. As I've stated in other threads I'm not super keen on purely mechanical systems at present, I think they have value but IMHO you need to incorporate speed ratings or time ratings or pricing brackets or some type of form study/handicapping however minor.

Backfitting is a trap but there is no other way to 'improve' a system but it can be done logically or foolishly. I think you need to ask yourself questions when you trim a criteria or bring in a new rule not just say it must be good cause the figures improve.

Could dribble on for hours so better pull up but happy to give my opinion and account of some of my experiences anytime.

Grub

[Punter76]

Excellent stuff, agree with alot you said, and I went over my figures and your comments backed up my figures.

ie. backfitting a high strike rate system
Place to winners ratio - 0.625.
I smell a rat here, maybe this is why I have had many 2nds and 3rds this month, I probably over-filtered to achieve the historical profit.
It is all levelling out, profit turnover still good though, but going down slightly over time.
Needs many more bets in system yet.

ie. backfitting a longshot system
Place to winners ratio - 1.219
This looks on the money, plenty of 2nds, but the profit is still over 50% after 170 bets on the win only. Which indicates to me it has probably been through a slighty bad run, and looking at Feb results. 6 bets, 3 wins and a 3rd so the place ratio is moving slighty back towards 1:1 and the profit is climbing up over 55% now. Will still need another 100-200 bets at least for me though.

Exactly the comments I was looking for, thanks, and I'll now be careful backfitting a system if the new results don't show a place to winners ratio of at least 0.80:1, ideally 1.20:1 with a system showing strong win profits is a great sign that you are on the right path.

Any system (even backfitted, even if it sucks) after roughly 500 or so bets should hopefully have this kind of ratio anyway. It was something I began to think about yesterday when I couldn't work out why every system I backfit begins to fail, this may be the answer. I over filtered to achieve the historical profit and I just kept the winners having a good run mostly, now they are beginning to come 2nd and 3rd.
Murphy's Law hey?

Cheers,

[KennyVictor]

Punter76,

I don't know how big a database you have but if I'm testing or creating a system I don't use the whole database to test results on. Because I have a big one (not many places you can say that and not get censored :-) I tend to test say up to the end of 2002 data to create and then test my results on the last 3 years to see if it holds up on fresh data. Alternatively I create using say NSW races and test state by state to see if there is uniformity across the country (not quite so good theoretically as there may be differences in the way races operate, track condition, etc. state to state). If the data you created the rules with give much better results than the data you do the final test on it looks like you backfitted.

KV