Bill Cullen
06-03-2005, 05:36 PM
It has been my understanding and my experience that too many folks (myself included)throw out a promising system or methodology before they have adequately tested it. I am not going to talk here about adequate sample sizes, levels of statistical significance, randomization, etc. For true verification of any system or approach, there's no getting around statistics and that topic has been well covered in previous threads on this site.
What I aim to talk about here are a few easily observable signs that seem to correlate with the better systems I have tested over the years and I have tested hundreds of them (most of them manually). These 'signs' are usually logical enough with a little thought and analysis. I offer them as a reminder that even though you as a system tester might not be any where near finished testing your system, and the current results so far are discouraging, some of these 'signs' might indicate that there is yet light at the end of the proverbial tunnel.
Here they are in no particular order:
1) Are your system picks disproportionately scratched from the races they're entered in? Could be your picks are really legitimate contenders and the trainers are shoping around for the best races to enter them in.
2) Do your system picks seem to get dropped in class a little bit too often. or seem to get too many favorable jockey switches, or seem to have quicker turn-around times than usual? Legit contenders where the trainers believe they have a 'live' horse do not happen in isolation: a good trainer will make all the moves necessary he/she feels will help the 'live' horse win the race.
3) Do the successful system picks tend to cluster in certain categories: age, gender, distance, surface, surface conditions? This is just another way of asking if you are analyzing your results beyond merely the percentage of winners and your average payoff?
I'm curious to hear what other signs/tipoffs people have experienced in conjunction with testing good handicaping systems or approaches.
thanks,
Bill Cullen
What I aim to talk about here are a few easily observable signs that seem to correlate with the better systems I have tested over the years and I have tested hundreds of them (most of them manually). These 'signs' are usually logical enough with a little thought and analysis. I offer them as a reminder that even though you as a system tester might not be any where near finished testing your system, and the current results so far are discouraging, some of these 'signs' might indicate that there is yet light at the end of the proverbial tunnel.
Here they are in no particular order:
1) Are your system picks disproportionately scratched from the races they're entered in? Could be your picks are really legitimate contenders and the trainers are shoping around for the best races to enter them in.
2) Do your system picks seem to get dropped in class a little bit too often. or seem to get too many favorable jockey switches, or seem to have quicker turn-around times than usual? Legit contenders where the trainers believe they have a 'live' horse do not happen in isolation: a good trainer will make all the moves necessary he/she feels will help the 'live' horse win the race.
3) Do the successful system picks tend to cluster in certain categories: age, gender, distance, surface, surface conditions? This is just another way of asking if you are analyzing your results beyond merely the percentage of winners and your average payoff?
I'm curious to hear what other signs/tipoffs people have experienced in conjunction with testing good handicaping systems or approaches.
thanks,
Bill Cullen