30th December 2005, 02:54 PM
|
|
Member
|
|
Join Date: Jan 1970
Posts: 4,437
|
|
Dale,
I agree that one has to have a cut-off point obviously, but when that cut-off point is designed to eek out a profit by twisting and contorting results so that they grab less than a handful of bolters, something is wrong.
When your rules are logical and you have seen a trend completely seperate to the system you are using, then yes, you are right, you could then apply it as a filter. What I'm getting at is that most of this is backfitted and tailor made to throw up longshots, nothing wrong really, except that the rules applied to short priced horses do not increase the strike rate.
In other words it is chasing prices, not horses.
The fact that all the profit is made up from four horses should ring alarm bells in anyone's language.
Your method is different, you are using a cutoff because of research not because it throws up a 50/1 winner every 5% of bets.
A rule I live and die by:
Divide the profit by the maximum dividend, if the number is less than 20, throw it away!
I've used this rule for over 5 years and every one of my mechanical systems has never failed in real time.
You'll see hundreds of systems fail the test of time which score less than 10.
Some are sold which score less than 5.
The more rules you have, the greater the chance of failure.
|