A problem with mechanical systems... when to bet, what to eliminate?

I have numerous mechanical systems that are quite good.
Here is an example of one..

BETS RETURN POT% WINNERS S/R APW
126 $192 53% 24 19.05% $8.02

Now normally after 100 bets, 50% POT, I'd lock this system in.
However, I can break it down into 8 sub-systems. They are..

BETS RETURN POT% WINNERS S/R APW
10 $12.30 23% 2 20.00% $6.15
26 $20.00 -23% 3 11.54% $6.67
9 $22.10 146% 2 22.22% $11.05
5 $6.00 20% 1 20.00% $6.00
14 $40.20 187% 5 35.71% $8.04
19 $20.10 6% 3 15.79% $6.70
23 $48.60 111% 5 21.74% $9.72
20 $23.10 16% 3 15.00% $7.70

I wouldn't back a system until at least 100 bets are in it, with at
least 30% POT. Which the system as a whole has achieved.

Would one be happy backing the system as a whole knowing that
you would be confident that at least 4 of the sub-systems will not show
a decent profit in the future based on other data I have? Even though the
whole system is returning 53%, after over 100 bets.

Or would one wait until each sub-system has 100 bets too?
That would take another 15 months at least, I am too impatient for that.



Now,
Another problem with mechanical systems, I have many systems showing a small 1-10% profit after 300 bets.

Example :
BETS RETURN POT% WINNERS S/R APW
260 $289.90 12% 33 12.69% $8.78

I read somewhere that you cannot filter too much of a system because
you would not know then what could be anomolies in the system that you
are keeping. ie, just a good run.

Now if I was to eleminate 150 bets in this system using 3 common-sense filters, those 150 bets returned only $100. Leaving the system with the following :

BETS RETURN POT% WINNERS S/R APW
110 $185.40 69% 22 20.00% $8.43

Now surely after 5 months watching this system (after the 3 filters)
clean up continuously month after month, a mechanical system can work
no matter how heavily filtered it is? Correct?


Last question, mechanical systems. How many bets in the database would
each of you be happy with before you start betting?

It's ****ing me off because I am not confident enough to bet on my selections
religiously yet. All the doomsday articles I read make me unsure if I should
begin yet.

Yeah, that's what I thought too.

Thanks, I'll go with that.

Are you sure?

Well the positive feedback is great. Why not...

But your name is not Sportz, Chrome or xpt,
and you ain't involved in a p1ssing contest..

Silly me, and I thought this was the place and the thread to share thoughts on punting systems.....

The reason I didn't reply to your post, is because you asked the questions and then went on to basically answer them yourself.

Personally, I would not dissect the systems too much, as you risk trimming down the data which will distort the results.

You are right, there are dangers in using too many rules, the best systems are straightforward and have the least number of rules.

Bear this in mind, the more rules you have, the more data you need to try it over, as the probability of backfitting increases with each rule.

__________________
RaceCensus - powerful system testing software.
Now with over 412,000 Metropolitan, Provincial and Country races!
http://www.propun.com.au/horse_raci...ng_systems.html
*RaceCensus now updated to 31/12/2024
Video overview of RaceCensus here:
http://www.youtube.com/watch?v=W821YP_b0Pg

your on the money there sahasastar...you see your post had to much merit and it could not be pulled apart ,,and you made too look like a dik...................cheers.......slowman........ ...............

Morning!! What's going on with all Sahasastar's replies (52-55) to posts which are NOT THERE??? Anyway, when it comes to "systems", I start betting when I genuinely believe I'm onto something (there is no "trial" period of waiting and watching). This means noticing something "odd" and quickly scanning a few records. If this oddity shows up again and I can explain it (to myself) then I BET it continuously, immediately. For example, let's say I notice that a number of Stable A's horses have been winning with Jockey "X" aboard and that they are paying well. I quickly note that this has been the case for at least the past 3 months. I ask myself why this is happening (especially if "X" is not the most accomplished jockey riding Stable A's nags). Is it "random"? I discover that "X" is the stable foreman's nephew (sound like a Melbourne Cup winning hoop now riding in England????). That's it. a system is born. So maybe it doesn't work out but mostly these "hunches" DO, until the rest of the punters wake up to "X". This principle (bet immediately) is the "best waves" surfing theory in action.
When I would go surfing I noticed that snap decisions (if soundly based!!) were best. WAITING for several great waves to go by BEFORE jumping on often meant missing the best rides of the day. You were consequently sitting out the perfect breaks and leaping aboard the "duds", experiencing the troughs, losing confidence, and then seeing ANOTHER "pipeline" go by while dithering. My advice is BE BRAVE and trust yourself to be on the Crests, in the surf and on the punt!!

apologies for not responding earlier, but I was like Sportz..felt like you answered your own questions. In addition to number of bets, I would also feel more comfortable with at least one full racing season, to take into account the different periods of the year, tracks, race quality etc. I personally think the measure of a system (for safety) is not only the number of bets, but also the time period of at least one full season, preferably two seasons, or more...but I understand one's eagerness to get on with it and start throwing real money at it.


It's ****ing me off because I am not confident enough to bet on my selections
religiously yet.


Therein is your answer I think. If you do not feel confident, my suggestion is ,either bet very small to gain confidence, or wait awhile longer until you do feel confident about the data you have built up.

Confidence is a major requirement for success..it is as important as the system itself..one cant really work without the other.

System+confidence+ discipline = success

all the very best to you!