streaks in testing and playing

[JJMartin]

Originally Posted by DeltaLover

The problem is not about the accuracy of the results of the paper betting but the fact that positive return will not be duplicated in the future... Almost all the betting strategies, will eventually converge to a specific negative ROI (usually between 0.89 - 0.95)

"Almost all" - That I agree.

[DeltaLover]

Originally Posted by JJMartin

"Almost all" - That I agree.

Given enough betting time the "Almost all" becomes "ALL" though..

[Capper Al]

Here is my reality, my personal time available to do anything with racing is constrained until I retire. The best I do usually is back test about 30 races. My forward testing is with small wagers. Not enough time to double the forward playing. Once in a while, I actually back test 200 races, but not too many times.

[ultracapper]

Back testing is usually motivated by recognizing a positive pattern, and then investigating that pattern in the past. The initial positive pattern that motivated the investigation should be eliminated from the investigation as it will skew results. It's kind of like a head start. These positive patterns are usually an anomaly, and therefore when practicing the study forward, negative results are arrived at.

[JJMartin]

Originally Posted by ultracapper

Back testing is usually motivated by recognizing a positive pattern, and then investigating that pattern in the past. The initial positive pattern that motivated the investigation should be eliminated from the investigation as it will skew results. It's kind of like a head start. These positive patterns are usually an anomaly, and therefore when practicing the study forward, negative results are arrived at.

I agree.
And also invalidating or being suspicious of samples that go negative when outliers are removed.

[Capper Al]

Getting some good results going forward with my testing. Let's see how long they last.

[green80]

Originally Posted by DeltaLover

The problem is not about the accuracy of the results of the paper betting but the fact that positive return will not be duplicated in the future... Almost all the betting strategies, will eventually converge to a specific negative ROI (usually between 0.89 - 0.95)

Your average horseplayer would be happy with an .89-.95 negative roi. It would be an improvement on what he is doing now.

[DeltaLover]

Originally Posted by green80

Your average horseplayer would be happy with an .89-.95 negative roi. It would be an improvement on what he is doing now.

Although Happiness is something very subjective I still have to doubt about the correctness of your statement...

A 0.89 – 0.95 ROI is more the enough to destroy even the largest bankrolls, in a very quick rate..

If you are going to get broke, it does not really matter how 'negative' your ROI is going to be, whar really counts is the bottom line which in both cases is exactly the same..

[green80]

Originally Posted by DeltaLover

Although Happiness is something very subjective I still have to doubt about the correctness of your statement...

A 0.89 – 0.95 ROI is more the enough to destroy even the largest bankrolls, in a very quick rate..

If you are going to get broke, it does not really matter how 'negative' your ROI is going to be, whar really counts is the bottom line which in both cases is exactly the same..

Most horseplayers are losers long term. And most lose more than 5-11% of of their long term wagering total. That is why I say most players would be money ahead with a .89-.95 ROI.

[thaskalos]

Originally Posted by green80

Most horseplayers are losers long term. And most lose more than 5-11% of of their long term wagering total. That is why I say most players would be money ahead with a .89-.95 ROI.

Long term...I think most horseplayers lose 100% of what they wager.

[Stillriledup]

Originally Posted by DeltaLover

Although Happiness is something very subjective I still have to doubt about the correctness of your statement...

A 0.89 – 0.95 ROI is more the enough to destroy even the largest bankrolls, in a very quick rate..

If you are going to get broke, it does not really matter how 'negative' your ROI is going to be, whar really counts is the bottom line which in both cases is exactly the same..

If Blended takeout is about 20 percent and a player is losing 5 to 11%, he's 9 to 15 percent better than the crowd...which is pretty good. But, i think that a handicapper needs to be a "perfect" bettor/gambler to realize the small "Edge" over the crowd.

Horse players need to be almost 50 percent better HANDICAPPERS than the crowd just to have a shot when you factor in that the takeout is 20 percent and there will be a large portion of the remaining 30 percent that gets chopped up by poor wagering decisions and bad money management.

[Capper Al]

Originally Posted by thaskalos

Long term...I think most horseplayers lose 100% of what they wager.

I agree. The loses over time cause anyone with the smallest negative ROI to tap out.

[green80]

Originally Posted by Stillriledup

If Blended takeout is about 20 percent and a player is losing 5 to 11%, he's 9 to 15 percent better than the crowd...which is pretty good. But, i think that a handicapper needs to be a "perfect" bettor/gambler to realize the small "Edge" over the crowd.

Horse players need to be almost 50 percent better HANDICAPPERS than the crowd just to have a shot when you factor in that the takeout is 20 percent and there will be a large portion of the remaining 30 percent that gets chopped up by poor wagering decisions and bad money management.

Well stated. Which was my point, even though a losing roi, still better than the crowd or average horseplayer.

[raybo]

Originally Posted by Capper Al

Here is my reality, my personal time available to do anything with racing is constrained until I retire. The best I do usually is back test about 30 races. My forward testing is with small wagers. Not enough time to double the forward playing. Once in a while, I actually back test 200 races, but not too many times.

Those 30 races that you are back testing, are they from the same time period from the previous year, or the previous 30 similar race types before you go live? My experience is that small data sets, which 30 races is, must be recent and/or very similar race types from the same time period from a previous year. If you're using generic patterns from different race types, or different times of the year, good luck with your small testing samples going forward.

Don't get me wrong, small testing samples are exactly what I use, but they are from the same time period I am about to play, from the previous year at the same track, or they are similar race types from immediately preceding races prior to my live betting. Time of meet, time of year, and similar race types/field dynamics, IMO, are required if the data sample is small.

[raybo]

Originally Posted by thaskalos

Long term...I think most horseplayers lose 100% of what they wager.

Actually, they probably lose more than 100% of what they wager, they have additional related expenses other than their actual wagering.