Would you bet this system?

[jackad]

I have been testing a system for the past 30 days. It is a very simple system, taking only seconds per race.
Here are the results: 263 bets, 18.6% win frequency, ROI +13.4%

On the basis of these data, would you bet this system with the expectation of forward success? If not, because the sample size is insufficient, at what point would you begin to rely on this system?

How does the win frequency and ROI compare with other systems that you may be using or are familiar with? Or in other words, on a scale of 1 to 5 with 5 the best, how would you rate this system?

Thanks for your response.

Jack

[JimG]

Jack:

The win frequency, ROI, and sample size are all to small for my liking to expect a forward positive expectation.

Since I am not a "system" or database player I cannot offer an opinion with respect to what would be sufficient to go forward. In all honesty, just because a system has done well over a certain number of races, I do not necessarily believe it will throw out a positive roi in the future.

Good luck in your quest. The betting public is pretty sharp on average and tend to pick up trends at a race track very quickly.


Jim

[Tom]

Throw out your two biggest payoffs and recalculate.
Is it still profitable?
How many place horses did the system pick?
Should be about twice as many place horses as winners.
If both of these ideas still give you a winning system,
play it until it stops working.
Tom

[Bob Harris]

Originally posted by Tom
Throw out your two biggest payoffs and recalculate.
Is it still profitable?
How many place horses did the system pick?
Should be about twice as many place horses as winners.
If both of these ideas still give you a winning system,
play it until it stops working.
Tom

I agree with Tom...once you weed out the bombs (I usually calculate the average mutual and then give any horse over 50.00 credit for only the average mutual. This still gives the system credit for the wins...it did pick the winner after all...but tempers the payoff to a level which is less likely to show aberrant results) take out 200 bucks from your bankroll and make 1% flat bets while you're testing it...it'll certainly make the test more fun!

I once saw a sprint system from Sport Stat in Las Vegas which showed a nice profit after 1000 races but had turned negative after 2000...where it would have been after 3000 races is anyone's guess. Endless testing could take months and still be inconclusive...if the system is based on a logical handicapping principle, I say go have fun with it!

[Lefty]

Jackad, take a $40 BR make $2 bets, go up a $1 every $20 profit.ee where it goes. Have fun, Most you can lose is $40 but no upward limits; ecept those of commom sense. Triple the BR put some in pocket start over with $60 etc.

[Tom]

Originally posted by Lefty
Jackad, take a $40 BR make $2 bets, go up a $1 every $20 profit.ee where it goes. Have fun, Most you can lose is $40 but no upward limits; ecept those of commom sense. Triple the BR put some in pocket start over with $60 etc.

Good idea-if it keeps working, you get the money and if it stops,you don't get hurt. Mayber tracking it for a while will shed light on some refinements to make it better.

Tom

[Rick Ransom]

Lefty and Tom,

Excellent advice! I always say that you should play with "other people's money" if you can. Risk very little and let the system prove itself out. If it's really good it'll keep on winning and increasing your bankroll. If not, you won't have lost much. Money won is easier to lose than money earned. The odds are against the majority of systems so you should take this pessimistic approach but give yourself a chance to benefit in case it's for real. One thing in your favor is that most of the systems that are profitable in the long run (according to my experience) fall in this general range (15-20% wins and 10-20% profit).

[GR1@HTR]

Jackad,


Congradulations on finding a positve play. Few thoughts good and bad.


The Bad: The ROI is poor for a just a one month sample. If it was based on a one year sample then yes it is very good. Many times in the past I have found plays that show 1.60 or so ROI on 3 months of data that still don't hold up.

The good:
You have found something that gives you lots of plays. That is very good!

Suggestions for further backfitting to get the ROI up:
1) Filter further by MLO ranges.
2) Filter by class. ie Claiming vs Alw vs Stakes vs Maiden
3) Filter by surface. ie Dirt vs Wet vs Turf.
4) Filter by finish position in last race
5) Filter by days since last race.


Best of luck with your new spot play.