When inventing a system....
 

OK here's one for you all.

I love reading the posts on this system part of the website, because systems are great. They're the promise of the pot of gold, the holy grail, the "discovery".

"EUREKA!" you cry as the set of results come in and it's in profit.

Then comes the dreams of sitting on a yacht, sipping champagne as you listen to your bets roll in, and of course the obligatory large mammoried blonde reclines next to you.

That is - of course - the illusion.

But I did read with interest the thread on the 4yo system recently, and it occurred to me, what's a good sample size for a system?

If for example you find a system that's winning 20% on turnover, then you're flying, but how reliable is that 20% for future forecasting? How many selections do you need in your sample size to be able to confidently say that the trend will continue?

Also, if the dividends of a system are calculated from SP, what improvement can one expect when using a variety of methods and choosing the best in the real world, ie products like maxi-div and best fluc and betfair and austote. If you're in front of your computer ten minutes before a race and you have all those at your disposal (though admittedly top fluc is not available ten minutes before), then surely you're going to average better odds than just SP.

Thoughts, peoples?

Undoubtedly the bigger the sample size the more confident you are of your results. My system which just covers WA is based on results from 1997 to present day and the fluctuation on a year to year basis is 97.2% returned for 1998 to 112.8% returned in 2002 (Based on Tabcorp prices).
This means if for example I had tried the system out at the end of 1998 I would have binned it as useless because it had a slight negative return - and that would have been based on over 3000 bets.
As my system chews through the data it updates the percent returned on screen after each race and the (rapidly moving figures) for each year shoot up and down wildly for the first part of a year, they can even move about quite a bit as the year being processed nears the finish. This just illustrates to me the clustering of wins and losses and convinces me that anything less than a couple of thousand results (unless you are showing a very significant win or loss percentage) is pretty meaningless. My system is based on form and selects a lot of favorites so I'm not talking about a system that is relying on a few long priced winners.
When I started using real money about a year ago, the first week I lost 7 out of 7 races but because I had over 7 years results behind me I never doubted it would come back to profit. The more data the more confidence you can bet with.

With regards SP versus other divvies. My system returns negative 0.8% on SP over the last 8 and a half years and positive 5.26% on Tabcorp. I don't know if that's typical or if my selections are just closer to bookies selections than Tab ones. On maxidiv which is where I bet at present I'm laughing. Nothing like hearing a horse pay $5.00 when you listen to the race on the radio and seeing you got 7 or 8 next time you fire up the computer.

KV

Thanks for that Kenny. Very informative.

The reason I put this post up is because I have been researching a bit of a system using my ratings, and over a roughly one year period there have been 320 selections, strike rate of 30%, ave div SP of $4.60 POT of 40%!

So, is 320 selections over about a year enough to get excited about?

It certainly is a promising beginning and I'd be feeling pretty good about it. But is'nt the answer to your ponder, how long is a piece of string.

As time goes by and your system continues to roll out profits consistently you must have confidence, but it could still show a negative, even it is just one year. Just sit down and hold on for the ride.

What else are you going to do?

Good Punting.

Yeah good point Top Rank. Your insight is, as usual, spot on.

For the record I thoguht I'd give it a crack today, looking at selections in advance for the first time. Today there were two, for one winner, paid $3.00, so I made a profit. If I replicated today's events exactly every single day I'd be a billionaire.

the maths
 

Hi Duritz
The first thing you should do is work out your "Difference between Win % S/R and the Win % S/R that will yield your Min Edge which is normally approx. 10%.
The formula for the above is Your Win%-(2+(2*Min Edge(.10)))/ avg payoff (to compute this avg payoff do the following (2*your avg payoff($3.60)+2.
based on your figures this should equate to 6.09%.

Now to find the minimum number of races computed to a confidence level of 99% do the following.
(your Win % * your Loss %)*(2.575/.0609)^2. the answer would be 376 races to provide you with a 99% confidence level.

Good Luck

Woof I used to think I was pretty handy at maths at school. Did 3 unit but I don't like to boast. (Is saying 3 unit dating myself?) Anyway that stuff that you just threw on there is way above anything I can handle.

I am happy knowing what I am doing is working and that the profit is there.

Duritz I am surprised. A savvy dude like you, your aiming way too low. A multi billionaire is the least you can expect. Good luck with that maths.

Thanks Woof, Top Rank.

The first part of your post, Woof, working out the difference between s/r and s/r required I did the following:

To make 10% profit, (minimum) return is $110 for every 100, which is a s/r req of 100/4.6 or 23.91%, making the difference of 6.09% that you talked about. So, got that bit right.

For the second bit you said:

Now to find the minimum number of races computed to a confidence level of 99% do the following.
(your Win % * your Loss %)*(2.575/.0609)^2. the answer would be 376 races to provide you with a 99% confidence level.

wouldn't that be therefore:

(0.3*0.7)*(2.575/.0609)^2

or

(0.21)*(42.28)^2

or

8.879^2

= 79 or so?

Which bit did I muck up?

In excel
 

I wrote the formula thinking you would input it directly into excel.
=sum(.3*.7)*(2.575/.0609)^2 the answer will be as I stated.

OK well I have done some more research, and now the number of selections is co-incidentally EXACTLY 376! Complete accident too, just said, "I'll research to this point then check the answers", and it had gone up to exactly 376 selections, which Woof said was the number I needed for a 99% accuracy level.

SO, here's the results after 376 selections:

selections - 376
wins - 116 (31%)
collect - 518.70
average dividend - $4.5
profit on turnover - 38%

Does this mean I can be 99% sure of the method holding up? Or something else? What does it all mean?!?!?!