Win rate, Standard Deviation, Bankroll, and Kelly's Criterion

[Salain]

How do you feel about regression testing? That's my term for going back through a sample size to discover patterns that occur at a particular percentage (not to be confused with software regression testing). Let's say I buy a database of all NBA games played so far this season. Then I test for certain conditions such as spreads greater than +10 for home teams and I find that these situations hit greater than 55% of the time.

Is this valid modeling so that I can expect to blindly bet situations such as this? Or better yet, bet selectively by either applying additional conditions or just using my own gut feel.

I do believe there are situations in sports that fit this criteria. If one could find these situations I believe you would be successful.

I have done a LOT of this. The simple formula to see if a win % is significant is multiplying the win % by 1 minus the win %, dividing by the number of samples, and taking the square root of that.

(sqrt (0.55 * 0.45) / 2000 )

Your win % is 99.7% within 3 times this number up or down from your mean. Your win % is 95% within two times this number up or down.

So the thing to keep in mind in regression analysis is that you are able to test a lot of different theories, and 1 out of every 20 will actually be a set that looks to be 95% significant, but is really just stumbled upon because you've tried 20 things.

A rule to use is that if you start to look at something that you *know* will affect win %, 2 standard errors from the mean is enough to start betting it. If you just start testin things that you *think might* affect it, and happen upon something you should have at least 2.5 standard deviations before betting on it.

 

[cincy_]

Doesn't SD=5 in this case, so 60 wins is just significant at the 5% level?

Yes, the stdev is 5, got that mixed up

 

[Woody0]

How do you feel about regression testing? That's my term for going back through a sample size to discover patterns that occur at a particular percentage (not to be confused with software regression testing). Let's say I buy a database of all NBA games played so far this season. Then I test for certain conditions such as spreads greater than +10 for home teams and I find that these situations hit greater than 55% of the time.

Salain has noted that you'll make spurious hits if you are running large numbers of tests, while Fish asserts this is true but a closely guarded secret. Since Fish is a man of considerable experience I accept that there might be some such patterns.

My personal feeling is that you'll always find patterns but I would tend to reject them as being real (much like betting black on roulette after 13 reds). In NBA there's a guy progression betting for up to three games for a win after 3 losses and vice versa who is apparently doing well. I guess this is basically unrelated to spread but is based on motivation or other human factors. On the other hand I tend to laugh at posts that say team A hasn't lost ATS playing at home after a three day rest when the visiting team played the previous night and won. So what!

So my answer is that I would not accept any prior patterns without further analysis. The further analysis would be to use the pattern as a hypothesis and predict future events (after all that's what you want to do anyway). Then supposing you have a hypothetical 55% win system test the future results for deviation from 50% and 55% to see if it holds up.

 

[quantumleap]

Salain has noted that you'll make spurious hits if you are running large numbers of tests, while Fish asserts this is true but a closely guarded secret. Since Fish is a man of considerable experience I accept that there might be some such patterns.

My personal feeling is that you'll always find patterns but I would tend to reject them as being real (much like betting black on roulette after 13 reds). In NBA there's a guy progression betting for up to three games for a win after 3 losses and vice versa who is apparently doing well. I guess this is basically unrelated to spread but is based on motivation or other human factors. On the other hand I tend to laugh at posts that say team A hasn't lost ATS playing at home after a three day rest when the visiting team played the previous night and won. So what!

So my answer is that I would not accept any prior patterns without further analysis. The further analysis would be to use the pattern as a hypothesis and predict future events (after all that's what you want to do anyway). Then supposing you have a hypothetical 55% win system test the future results for deviation from 50% and 55% to see if it holds up.

I agree that a lot of analysis is needed once a pattern is found. What I would like to add to this is that a particular pattern may be viable, yet it may be a seasonal pattern. For example, after November in English Preimier soccer and after the all-star games in both MLB and the NBA there is a different dynamic. Teams play differently for some reason. In MLB there is a trading deadline after the All-Star game that mixes up the line-ups of the teams.

The point I'm trying to make with all of this is be careful when using a system that has been tested for the entire season. It may not be viable for a particular part of the season.

 

[quantumleap]

You betcha there are!!!

Some of which have never produced a losing record in any year for 20+ years.

When you find one, suggest you do not tell ANYBODY.

Just to let people know here, Fish isn't being selfish. He's just saying that if you tell someone your system and they bet using that system you stand to get lower odds on your bets or perhaps miss out on a bet if the line changes against you.

Fish, if someone finds a system like this, why not just bet huge amounts on the system? I would think that one could make tons of money. Sell your wife, mortgage your house, take out loans, anything to get a higher bet amount for an eventually bigger payout.

Thanks much to everyone who contributed to this thread. It is by far my favorite thread of 2006. Special thanks to Woody0, Cincy_ and Fish.

 

[cincy_]

Fish, if someone finds a system like this, why not just bet huge amounts on the system? I would think that one could make tons of money. Sell your wife, mortgage your house, take out loans, anything to get a higher bet amount for an eventually bigger payout.

• If you have a 90% shot of winning, and you bet your house, what if you lose? You are toast.
• If you have a 90% chance of winning, and you bet $5 on it, you have not maximized your profit.

That is the ESSENCE of the Kelly Criterion. It calculates how much of your bankroll you need to wager to maximize your profit without totally exposing yourself. From Fishhead's post:

The natural impulse is to bet everything you've got on the horse that's supposed to win. But when the gambler adopts this policy, he is sure to lose everything on the first bum tip. Alternatively, the gambler could play it safe and bet a minimal amount on each tip. This squanders the considerable advantage the inside tips supply.<?xml:namespace prefix = o ns = "urn:schemas-microsoft-comfficeffice" /><o></o>

​

 

[quantumleap]

• If you have a 90% shot of winning, and you bet your house, what if you lose? You are toast.
• If you have a 90% chance of winning, and you bet $5 on it, you have not maximized your profit.

That is the ESSENCE of the Kelly Criterion. It calculates how much of your bankroll you need to wager to maximize your profit without totally exposing yourself. From Fishhead's post:

The natural impulse is to bet everything you've got on the horse that's supposed to win. But when the gambler adopts this policy, he is sure to lose everything on the first bum tip. Alternatively, the gambler could play it safe and bet a minimal amount on each tip. This squanders the considerable advantage the inside tips supply.<?xml:namespace prefix = o ns = "urn:schemas-microsoft-comfficeffice" /><o></o>


​

cincy_:

I have had this problem recently. When I go on a winning streak I have a tendency to bet amounts that are too large, thinking I am maximizing my profit. It is important to find an amount that will not bankrupt you once you start losing.

However, when I mentioned selling my house my intent was not for just one bet or a few bets. What I was wondering was if you have a valid system why not put as much money as you could into that system? I would like to stress here that it doesn't mean to throw all your money into a few bets. Certainly putting as much money into bets of a valid system would gain you more money in the long run.

 

[Fishhead]

Just to let people know here, Fish isn't being selfish. He's just saying that if you tell someone your system and they bet using that system you stand to get lower odds on your bets or perhaps miss out on a bet if the line changes against you.

Fish, if someone finds a system like this, why not just bet huge amounts on the system? I would think that one could make tons of money. Sell your wife, mortgage your house, take out loans, anything to get a higher bet amount for an eventually bigger payout.

Thanks much to everyone who contributed to this thread.
It is by far my favorite thread of 2006. Special thanks to Woody0, Cincy_ and Fish.[/quote]

Because there will be a RISK OF RUIN involved and a mathematical STANDARD DEVIATION to this sort of system, SO SOME TYPE OF LIMITATION must be administered. However, you are correct in assumming that one should attempt to EXPLOIT these situations as much as possible WITHOUT THE RISK OF RUIN when they present themselves........or knowing the percentages of this risk of ruin and wagering an amount that will provide the best for ones PERSONAL tolerance for ........risk vs reward, risk of ruin(or loss) potential, financial gain, etc., etc..........that is a FUNDAMENTALLY SOUND FINANCIAL STRATEGY.

Obviously more than a few of us here have found some juicy nuggets over the years and have been rewarded more than adequately..........and it is safe to say there are many in abundance currently. The majority of these are how the true professionals walk home with RELATIVELY EASY CASH year in and year out.


Glad to hear you are enjoying this thread, your contributions to it have been very thought provoking on this end and I thank you.

-FH-

 

[quantumleap]

So that's it. I'm done with the trial-and-error method.

I've been spending the last 3 years trying things here and there only to fail with all except a couple systems. I figured that it would be a good learning experience and it would provide entertainment. However, finding winning systems by experimentation is far too costly than betting needs to be.

I have resolved myself to find winning systems using analysis of existing data, rather than trial-and-error. Don't get me wrong, I feel using trial-and-error has given me an insight into sports betting that I could not get using systems alone.

I will follow a couple winning cappers here and will continue to use the systems I have found personally. But no more will I continue to experiment using trial-and-error which has been the bulk of my bets (and thus losses) so far.

It's been fun, but so is winning far fewer games at a higher rate.

 

[Fishhead]

LEAP-- Never would I allow myself to be a guinea pig........nuff said.

Leave that for others my friend.

fh

 

[Insiders]

Well this is a good thread for this. I am not a fan of the 1-2% of the bankroll like so many others here say they are. Lets assume that a player had a bankroll of 10K. Lets say he of she put themselves on that 1% grind. Now lets say they actually did hit 55% of the plays over a course of a season, year, whatever for 100 plays. So they have made 100 plays with 55 of them being a winner and 45 of them being a loser, at 100.00 per play that have 5500 in winners and 4950 in losers for a grand total of 550.00 for that season, year, whatever. There is the math useing -10 on all those plays, before someone comes in an says well you should just be laying -05 or 07 or whatever.It really ALL comes down to be able to pick winners, does it of does it not?

Bookmakers love to see people make a deposit of 10k and start betting a 1-2% of that figure. The Bookmakers just laugh an say the VIG will get this guy sooner of laters, and it will too.
Also if Bookmakers had their way about most things this is EXACTLY the kind of players that they want in thier shop. Just ask any of them that you know.

 

[Christian]

i would think just the opposite. if i were a bookmaker, i would much rather someone deposit, let's use your example, $10,000 and bet 5 or 10% of their bankroll on each play...

why would a bookie want their customers to wager such a conservative amount (1-2%)??? since most players are losers, wouldn't it make sense that bookies want them to bet a higher pct.??? they'll just bankrupt themselves quicker!

anyway, everyone has their own approach. i bet year-round. i end up with well over 1,000 plays each year from MLB, NBA and NFL. and i have an agreement with a really smart college basketball capper that i've known for years and we trade picks each year (i use his college plays and he uses my MLB plays each summer) this adds another couple hundred plays each year...

so with around 1,500 plays each year, it just makes sense for me to risk 1% on each of my plays. with a .04 cent return on each dollar invested, i prefer the "grind" method of making a profit each year. it still gives me a great profit but it also PROTECTS my downside during those INEVITABLE losing streaks that come with this business.

but what works for me, may not work for someone else...and to each their own. i just think it's crazy risking anything more than 1 or 2% per play, on ANY play...but that's just me.

 

[CoachLT]

Well this is a good thread for this. I am not a fan of the 1-2% of the bankroll like so many others here say they are. Lets assume that a player had a bankroll of 10K. Lets say he of she put themselves on that 1% grind. Now lets say they actually did hit 55% of the plays over a course of a season, year, whatever for 100 plays. So they have made 100 plays with 55 of them being a winner and 45 of them being a loser, at 100.00 per play that have 5500 in winners and 4950 in losers for a grand total of 550.00 for that season, year, whatever. There is the math useing -10 on all those plays, before someone comes in an says well you should just be laying -05 or 07 or whatever.It really ALL comes down to be able to pick winners, does it of does it not?

Bookmakers love to see people make a deposit of 10k and start betting a 1-2% of that figure. The Bookmakers just laugh an say the VIG will get this guy sooner of laters, and it will too.
Also if Bookmakers had their way about most things this is EXACTLY the kind of players that they want in thier shop. Just ask any of them that you know.

As I have said before, I am in Insiders camp. One thing about experience is that you eventually just "KNOW" when to step out. Now that does not mean that that particular play is a "LOCK" You will NEVER, NEVER, EVER hear me say a selection is a lock. There are 3 things you cannot handicap. Those being penalties, tournovers, and injuries (or crooked officials such as the ACC thieves that STOLE the F$U- Florida game in 03) Just go with what has brought you success in the past. LT

 

[Christian]

you're right about "knowing when to step out" and even though i have a very conservative style (as i've said, 1% of bankroll per play), there are times throughout the year when i'll wager more on a play...

the problem is, so few bettors really "know" when it's time to step it up. everyone wants to think that just because they win a little bit of money over a couple hundred plays, that it means they know what they're doing...

it just doesn't work that way. you won't know you'rea actual "edge" until you're well into the THOUSANDS of plays made. that's the reality of this business. and with beginner gamblers, it is SUICIDE to risk anything more than 1 or 2%...they WILL go broke, simple as that.

i just get upset when beginner/novice gamblers read what experienced gamblers wager (%-wise, i mean) and think that they can adopt the same approach.

play it safe for a few THOUSAND plays (not hundred, that's not a typo, a few THOUSAND) and then once you know for certain what your skill level is, then you can wager a little more aggressively if you want to...

or not, like me. nothing wrong with being conservative sometimes.

 

[CoachLT]

you're right about "knowing when to step out" and even though i have a very conservative style (as i've said, 1% of bankroll per play), there are times throughout the year when i'll wager more on a play...

the problem is, so few bettors really "know" when it's time to step it up. everyone wants to think that just because they win a little bit of money over a couple hundred plays, that it means they know what they're doing...

it just doesn't work that way. you won't know you'rea actual "edge" until you're well into the THOUSANDS of plays made. that's the reality of this business. and with beginner gamblers, it is SUICIDE to risk anything more than 1 or 2%...they WILL go broke, simple as that.

i just get upset when beginner/novice gamblers read what experienced gamblers wager (%-wise, i mean) and think that they can adopt the same approach.

play it safe for a few THOUSAND plays (not hundred, that's not a typo, a few THOUSAND) and then once you know for certain what your skill level is, then you can wager a little more aggressively if you want to...

or not, like me. nothing wrong with being conservative sometimes.

GREAT point Christian ! Beginners should follow a conservative method until they have hundreds if not thousands of plays under their belt. Also there is tons of info out there due to the internet that old timers did not have. We have 2 young men here at the RX in their early 30s who already have "IT" I am speaking of Dawoofdaddy and Sharky 99. I met both of these gentlemen at last years Bash and within 5 minutes I knew they had "IT" They have already served their apprenticeship and have graduated Magna Come Loudly from Handicapper College. :money: LT

 

[cincy_]

Bookmakers love to see people make a deposit of 10k and start betting a 1-2% of that figure. The Bookmakers just laugh an say the VIG will get this guy sooner of laters, and it will too.
Also if Bookmakers had their way about most things this is EXACTLY the kind of players that they want in thier shop. Just ask any of them that you know.

I agree it would be stupid to deposit 10 k and play 1% per play but that isn't what I do.

I am the 1%-2% sized bettor, but of my bankroll, each book never sees more than 5% of it. Right now, I have 65% of my bankroll in the bank, 20% in Neteller, 2%-3% in various books.

So the book never gets a deposit of 10k. They see a $500 deposit and see me playing 10%-20% of that each play. Or they may see a $200 deposit and me wagering 100% on one play.

Even if they see me wagering my whole deposit on one play, I could still be a 1% bettor. They have no idea what my bankroll is. All they see is what is deposited in their book and what is wagered in their book - and they never see a deposit of more than $500.

 

[Fishhead]

GREAT point Christian ! Beginners should follow a conservative method until they have hundreds if not thousands of plays under their belt. Also there is tons of info out there due to the internet that old timers did not have. We have 2 young men here at the RX in their early 30s who already have "IT" I am speaking of Dawoofdaddy and Sharky 99. I met both of these gentlemen at last years Bash and within 5 minutes I knew they had "IT" They have already served their apprenticeship and have graduated Magna Come Loudly from Handicapper College. :money: LT

Coach LT and myself have discussed this many times in our conversations.

Think us old guys sorta feel cheated in someway or another.

:lolBIG:

 

[CoachLT]

Coach LT and myself have discussed this many times in our conversations.

Think us old guys sorta feel cheated in someway or another.

:lolBIG:

Fish, now that the Holy Grail is within reach after suffering all these years, the U.S. Holy Rollers are trying to thwart us. Beware of people who wrap themselves in the flag and hide behind the bible. They will take away your gambling while viewing kiddie porn. LT:monsters- ucking:

 

[Fishhead]

Fish, now that the Holy Grail is within reach after suffering all these years, the U.S. Holy Rollers are trying to thwart us. Beware of people who wrap themselves in the flag and hide behind the bible. They will take away your gambling while viewing kiddie porn. LT:monsters- ucking:[/qu


You are probably correct sir.

 

[Journeyman]

Fish, now that the Holy Grail is within reach after suffering all these years, the U.S. Holy Rollers are trying to thwart us. Beware of people who wrap themselves in the flag and hide behind the bible. They will take away your gambling while viewing kiddie porn. LT:monsters- ucking:

So damn true Coach!

good thread Fish.