Is a Positive ROI Attainable and Sustainable- How Would You Know?

[Wizard of Odds]

By (1) fitting prior races, (2) validating with hold-out races and (3) further testing with additional races, you might find some subset of conditions that produce a positive ROI (even before rebates). How would you be reliably assured that you could reproduce a similar positive ROI on future races? After testing thousands of models and combinations of variables, how do you know that you merely found that positive ROI subset by randomness?

Even if your conditions reliably produced a positive ROI, how are you assured that the game won't change such that your edge disappears? That your own bets reduce or eliminate your edge?

Are your expected dollars earned for the year greater than you would definitely earn by investing in risk-free U.S. Treasuries?

...Just thinking

[acorn54]

Originally Posted by Wizard of Odds

By (1) fitting prior races, (2) validating with hold-out races and (3) further testing with additional races, you might find some subset of conditions that produce a positive ROI (even before rebates). How would you be reliably assured that you could reproduce a similar positive ROI on future races? After testing thousands of models and combinations of variables, how do you know that you merely found that positive ROI subset by randomness?

Even if your conditions reliably produced a positive ROI, how are you assured that the game won't change such that your edge disappears? That your own bets reduce or eliminate your edge?

Are your expected dollars earned for the year greater than you would definitely earn by investing in risk-free U.S. Treasuries?

...Just thinking

for what it is worth cx wong in his book "precision" and who was , (is?) the modeler for the benter team in hong kong, measured r squared of the parimutuel market efficiency of what their team was betting into. the conclusion was that the market is evolving towards being more and more efficient as time passes. so yes profits are attainable, but less and less so as time and technology of statistical analysis progresses.

[Buckeye]

We all only have only so much time . . .

Anything is possible but I've been down that road of trying to fit future outcomes to the past and it simply does not work.

[Wizard of Odds]

Originally Posted by Buckeye

We all only have only so much time . . .

Anything is possible but I've been down that road of trying to fit future outcomes to the past and it simply does not work.

Then, are we all lunkheads trying to beat the fair roulette wheel?

[jay68802]

Originally Posted by Buckeye

We all only have only so much time . . .

Anything is possible but I've been down that road of trying to fit future outcomes to the past and it simply does not work.

Find a new game.

[Tom]

Keep records of evet bet.
If you don't, you don't know where you are.
You might be killing them in win bets and giving it all back in exactas.

[taxicab]

Originally Posted by Tom

Keep records of evet bet.
If you don't, you don't know where you are.
You might be killing them in win bets and giving it all back in exactas.

This.
Every bet.
Every dollar wagered.
If you lose 20-24% of your annual amount wagered,you're decent for sure(takeout don't ya know).

[castaway01]

Statistically, I don't know how you would "prove" it's sustainable. I've seen the arguments with the stat guys (who never bet) here and you'll never get a large enough sample to prove a method is sustainable. So I think you just have to look at your ROI and your bankroll. If you've got more money at the end of the year than you did at the start, you're doing well. Sure it's attainable. Sustainable? Only for the rare few.

[Dave Schwartz]

Originally Posted by castaway01

Statistically, I don't know how you would "prove" it's sustainable. I've seen the arguments with the stat guys (who never bet) here and you'll never get a large enough sample to prove a method is sustainable. So I think you just have to look at your ROI and your bankroll. If you've got more money at the end of the year than you did at the start, you're doing well. Sure it's attainable. Sustainable? Only for the rare few.

What a mouthful! And spot on.



My New Years resolution was to stop reminding people to just look at their withdrawals vs deposits. It pisses people off.

[headhawg]

Originally Posted by castaway01

Statistically, I don't know how you would "prove" it's sustainable. I've seen the arguments with the stat guys (who never bet) here and you'll never get a large enough sample to prove a method is sustainable. So I think you just have to look at your ROI and your bankroll. If you've got more money at the end of the year than you did at the start, you're doing well. Sure it's attainable. Sustainable? Only for the rare few.

Spot on!

[Robert Fischer]

Originally Posted by Wizard of Odds

By (1) fitting prior races, (2) validating with hold-out races and (3) further testing with additional races, you might find some subset of conditions that produce a positive ROI (even before rebates). How would you be reliably assured that you could reproduce a similar positive ROI on future races? After testing thousands of models and combinations of variables, how do you know that you merely found that positive ROI subset by randomness?

The problem with the above idea, is that it's possible to do so without an understanding of the major models that actually drive the system.

If your work is great, then you may in essence be breaking up races into categories where certain models are found to be significant. In that way you can mimick the way that the 'thinking' should be done. Then they still have to be applied case by case and signal a play when a significant margin of error is present(meaning the expected odds from the public/market are 'irrational'.

The vast majority of idea above instead looks at stats and trends and tries to find 'success' over a trend of results that would be positive had it been bet... That isn't at all what I am suggesting is necessary.

Once you had that great algorithm, you still need to apply them to current patterns to see where the major models are pointing, and whether you are getting a 'go'(multiple models pointing in the same direction + irrational market price) or a pass (tradeoffs).

Then market price itself has to be analyzed; Is it irrational because you fully expected a 'miscalculation' of the major models?(Go!), or is it irrational for reasons that you don't understand(Pass!).

Then proper allocation; highest and best use in both wager-type(s) and bankroll allocation (%kelly etc...)

Even if your conditions reliably produced a positive ROI, how are you assured that the game won't change such that your edge disappears? That your own bets reduce or eliminate your edge?

You understand the major models that drive the system. And you are proactive. If conditions that drive the major models change, then you anticipate and watch along as those models become more or less significant. Like the first 'clash' of a battle, you can adapt. You also are prepared and expect for clashes and changes.


You can play with statistics and back-fit, and that can be great entertainment in and of itself, and maybe even improve your gambling. Who knows... with top rebates or during a short term, you may even show a nice profit.

A true 'edge' requires competence of certain major models and the ability to see when races are within or without your circle of competence.

[Greyfox]

Originally Posted by Buckeye

Anything is possible but I've been down that road of trying to fit future outcomes to the past and it simply does not work.

Are you saying that past performance is not a predictor of future performance?
If so, I beg to differ.
Past performance may not be a perfect predictor, but in many areas of life, including horse racing, it's the best we have.

[acorn54]

Originally Posted by Greyfox

Are you saying that past performance is not a predictor of future performance?
If so, I beg to differ.
Past performance may not be a perfect predictor, but in many areas of life, including horse racing, it's the best we have.

there are no guarantees in life, as far as me, yes i am a stat guy, only a year or two of making serious profits, in the past 10 years. i am just passing along the conclusions and info other stat guys mention in books or other media, i have never met any in person, that have been verified by the tax collector to have made huge amounts of money. in other words, don't kick me, i am just the messenger.

[Greyfox]

Originally Posted by acorn54

there are no guarantees in life, as far as me, yes i am a stat guy, only a year or two of making serious profits, in the past 10 years. i am just passing along the conclusions and info other stat guys mention in books or other media, i have never met any in person, that have been verified by the tax collector to have made huge amounts of money. in other words, don't kick me, i am just the messenger.

Of course there are no guarantees in life.
I'm not kicking anyone.
I'm just stating that in most areas of life, including horse racing, Past Performance is the Best Predictor of Future Performance.
Obviously other predictors, probabilities, and conditions have to factored in to improve accuracy.

[sjk]

Over the years I have found less opportunity in the low price ranges so my win percentage is lower and the average mutual is higher.

Even when the total results are good, if a few large strikes are responsible for the profit you have to wonder how much luck played a part. Of course there were bad beats too.

I still can't understand how anyone playing at a high volume and with a slim margin can possibly know whether they still have an edge or not when things go south.

Things have changed enough over the years that your good record over the past 20 years may not be enough to guarantee anything.