kelly q

[MessiStuff]

Lets say i have a +ev bet and i stake the kelly suggested stake:

I have another +ev bet to place that day, what do i consider my roll to be?

Roll = roll - the bet i staked already

OR

Roll = roll + expected profit of first plus ev bet

[Wrecktangle]

Are the bets simultaneous in time during that day? If not, 1st goes off at Kelly, and the 2nd goes off at Kelly after the new bankroll has been re-calced with the result of the 1st bet.

If simultaneous then the highest Kelly fraction bet goes off at fully Kelly, then your bankroll is degraded by the 1st bet Kelly amount, and then the 2nd bet goes off at it's Kelly fraction of the remaining bankroll.

BTW, this is an excellent way to burn down a bankroll in record time if you are not absolutely sure of your Kelly fraction, and few (99%+ of posters in this, or any forum) are.

Since you are asking Qs like this, most likely you fall into the 99%+ group.

[bztips]

Assuming simultaneous independent events, the correct answer is given by Ganchrow here

[MessiStuff]

If simultaneous then the highest Kelly fraction bet goes off at fully Kelly, then your bankroll is degraded by the 1st bet Kelly amount, and then the 2nd bet goes off at it's Kelly fraction of the remaining bankroll.

Thanks this is what i was looking for.

BTW, this is an excellent way to burn down a bankroll in record time if you are not absolutely sure of your Kelly fraction, and few (99%+ of posters in this, or any forum) are.

When you say that its an excellent way to burn roll if your unsure of your kelly fraction, do you mean that people mis-estimate edge, or that they use kelly wrong?

[bztips]

Thanks this is what i was looking for.

Sorry, Wreck is wrong -- as I said, see Ganch's post; that has the correct formula.

[Peep]

BTW, this is an excellent way to burn down a bankroll in record time if you are not absolutely sure of your Kelly fraction, and few (99%+ of posters in this, or any forum) are.

No shit, I am certainly in that group. Easy to assume an edge that ain't there any more, just because that edge was there in the past.

Or, as Grandpappy Peep used to put it..

"It ain't what you don't know that will get you. It is what you KNOW THAT AIN'T TRUE that will get you everytime".

[Data]

If simultaneous then the highest Kelly fraction bet goes off at fully Kelly, then your bankroll is degraded by the 1st bet Kelly amount, and then the 2nd bet goes off at it's Kelly fraction of the remaining bankroll.

BTW, this is an excellent way to burn down a bankroll in record time if you are not absolutely sure of your Kelly fraction, and few (99%+ of posters in this, or any forum) are.

Besides providing mathematically wrong Kelly strategy, Wrecktangle also uses incorrect Kelly terminology. He completely misuses a term "Kelly fraction", which has a very distinct meaning. Every time in the quoted fragment above he said "fraction" he actually meant what is called "stake". This diarrhea of ignorance proves that Wrecktangle has no clue about Kelly, the suspicions we had has just been confirmed. I feel it is important to point out this fact since he carries himself here as know-it-all. Most of his posts are OK but sometimes he talks with aplomb and authority about the topics he does not know about. The poster and the lurkers should be aware of this.

[MessiStuff]

Had a look at ganchrow's post, someone would have to break it down for me to understand.

Was messing around with the SBR kelly calculator, which seems to give me the answer, even if i don't understand the math behind it, i understand the principle that the bet which is the most +EV gets a larger stake of your bankroll than bets that are less +EV, which is fairly obvious.

[Data]

Assuming simultaneous independent events, the correct answer is given by Ganchrow here

Right, the above links to discussion of theoretical aspects of the solution. I think th OP may be more interested in practical uses.

MessiStuff, the easiest way to practically solve the problem you presented is to use Kelly calculator. It assumes that you can bet parlays consistent of the two events at hand. If you cannot bet parlays then you can use less user friendly but much more versatile "generic kelly.xls" Excel spreadsheet attached to this post.

[Flying Dutchman]

Besides providing mathematically wrong Kelly strategy, Wrecktangle also uses incorrect Kelly terminology. He completely misuses a term "Kelly fraction", which has a very distinct meaning. Every time in the quoted fragment above he said "fraction" he actually meant what is called "stake". This diarrhea of ignorance proves that Wrecktangle has no clue about Kelly, the suspicions we had has just been confirmed. I feel it is important to point out this fact since he carries himself here as know-it-all. Most of his posts are OK but sometimes he talks with aplomb and authority about the topics he does not know about. The poster and the lurkers should be aware of this.

No shit? Well "Mr know it all", "we" confirmed this by simulation. For my part I've seen "fraction" used and "stake" used.

Not to blow the lid off all this shit, but Kelly is sub-optimal as far as maximal BR growth is concerned, the optimal "fraction" or "stake" appears to be higher.

"Correct Kelly" is a bullshit term in sports betting, as none of it is optimal in terms of maximal BR growth is concerned.

The real fvking point is, YOU, Ganch, nor anyone really knows how to optimally bet this stuff as you didn't even go back to see of Kelly really is optimal in terms of growth of BR. You huff and puff about "Kelly." There is so much variability in sports prediction you CANNOT show any of it is optimal.

...which was Wreck's main point: 99%+ of you don't know enough about your predictions systems to "optimally" bet fvk-all in sports.

I'll put that percentage at 100%. None of us know: Gödel was right.

[Wrecktangle]

Besides providing mathematically wrong Kelly strategy, Wrecktangle also uses incorrect Kelly terminology. He completely misuses a term "Kelly fraction", which has a very distinct meaning. Every time in the quoted fragment above he said "fraction" he actually meant what is called "stake". This diarrhea of ignorance proves that Wrecktangle has no clue about Kelly, the suspicions we had has just been confirmed. I feel it is important to point out this fact since he carries himself here as know-it-all. Most of his posts are OK but sometimes he talks with aplomb and authority about the topics he does not know about. The poster and the lurkers should be aware of this.

Why don't "we" let you, the OP and lurkers in on a secret?

None of the folks who make a living in this business use Kelly. They don't even talk about it. BTW, these folks are not math-stupid. Years of experience has revealed the fact that Kelly simply is unworkable.

In spite of the theoretical mumbo-jumbo the formula pointy-heads rant about, it isn't practical in the real business of sport. We've built simulations to test (and yes, the technique I mentioned did come out better than others on average in terms of BR growth) but at the end of the day, there is just too much unknowability (variability) of your true Bayesian priors to make an accurate forecast of your Kelly BR Fraction (sorry, when I was in Europe, this is the term we used).

So it is up to the individual punter: you can risk your Bankroll on a method that is widely known amongst the pros that it is a road to ruin, or you can listen a pissed off forum troll who argues semantics with little practical experience playing with explosives.

[bztips]

Why don't "we" let you, the OP and lurkers in on a secret?

None of the folks who make a living in this business use Kelly. They don't even talk about it. BTW, these folks are not math-stupid. Years of experience has revealed the fact that Kelly simply is unworkable.

In spite of the theoretical mumbo-jumbo the formula pointy-heads rant about, it isn't practical in the real business of sport. We've built simulations to test (and yes, the technique I mentioned did come out better than others on average in terms of BR growth) but at the end of the day, there is just too much unknowability (variability) of your true Bayesian priors to make an accurate forecast of your Kelly BR Fraction (sorry, when I was in Europe, this is the term we used).

So it is up to the individual punter: you can risk your Bankroll on a method that is widely known amongst the pros that it is a road to ruin, or you can listen a pissed off forum troll who argues semantics with little practical experience playing with explosives.

All well and good, but the OP didn't ask your opinion (or my opinion, or anyone else's opinion) about whether Kelly can "work" in the real world. He had a technical question about how to compute the Kelly stake, which you answered incorrectly.

Case closed.

[THA_DIESEL]

Why don't "we" let you, the OP and lurkers in on a secret?

None of the folks who make a living in this business use Kelly. They don't even talk about it. BTW, these folks are not math-stupid. Years of experience has revealed the fact that Kelly simply is unworkable.

In spite of the theoretical mumbo-jumbo the formula pointy-heads rant about, it isn't practical in the real business of sport. We've built simulations to test (and yes, the technique I mentioned did come out better than others on average in terms of BR growth) but at the end of the day, there is just too much unknowability (variability) of your true Bayesian priors to make an accurate forecast of your Kelly BR Fraction (sorry, when I was in Europe, this is the term we used).

So it is up to the individual punter: you can risk your Bankroll on a method that is widely known amongst the pros that it is a road to ruin, or you can listen a pissed off forum troll who argues semantics with little practical experience playing with explosives.

Hey Wrecktangle, I was going to reply to your PM but it says your inbox is full. Thanks

[MonkeyF0cker]

Why don't "we" let you, the OP and lurkers in on a secret?

None of the folks who make a living in this business use Kelly. They don't even talk about it. BTW, these folks are not math-stupid. Years of experience has revealed the fact that Kelly simply is unworkable.

In spite of the theoretical mumbo-jumbo the formula pointy-heads rant about, it isn't practical in the real business of sport. We've built simulations to test (and yes, the technique I mentioned did come out better than others on average in terms of BR growth) but at the end of the day, there is just too much unknowability (variability) of your true Bayesian priors to make an accurate forecast of your Kelly BR Fraction (sorry, when I was in Europe, this is the term we used).

So it is up to the individual punter: you can risk your Bankroll on a method that is widely known amongst the pros that it is a road to ruin, or you can listen a pissed off forum troll who argues semantics with little practical experience playing with explosives.

What an absolutely ignorant statement. So Billy Walters doesn't use FRACTIONAL Kelly? It's unbelievable how much putrid shit you spew in this forum.

[MonkeyF0cker]

No shit? Well "Mr know it all", "we" confirmed this by simulation. For my part I've seen "fraction" used and "stake" used.

Not to blow the lid off all this shit, but Kelly is sub-optimal as far as maximal BR growth is concerned, the optimal "fraction" or "stake" appears to be higher.

"Correct Kelly" is a bullshit term in sports betting, as none of it is optimal in terms of maximal BR growth is concerned.

The real fvking point is, YOU, Ganch, nor anyone really knows how to optimally bet this stuff as you didn't even go back to see of Kelly really is optimal in terms of growth of BR. You huff and puff about "Kelly." There is so much variability in sports prediction you CANNOT show any of it is optimal.

...which was Wreck's main point: 99%+ of you don't know enough about your predictions systems to "optimally" bet fvk-all in sports.

I'll put that percentage at 100%. None of us know: Gödel was right.

It's really cute how you and your boyfriend Wrecktangle have each other's back in every thread. What's not cute is how often you both embarrass yourself. LMAO at your "simulation." Why not post your results then? I'd certainly love to see how badly you two mucked it up.

[Flying Dutchman]

It's really cute how you and your boyfriend Wrecktangle have each other's back in every thread. What's not cute is how often you both embarrass yourself. LMAO at your "simulation." Why not post your results then? I'd certainly love to see how badly you two mucked it up.

Here's troll 1 having troll's 2 back...

BTW, bonehead: fractional Kelly IS NOT OPTIMAL by Kelly's def, so it's just another betting system that you blowhards are waxing on about.

Oh and do the math, Walters can't use Kelly, the market can't support it.

...have a nice life.

[Wrecktangle]

Hey Wrecktangle, I was going to reply to your PM but it says your inbox is full. Thanks

Fixed, my apologies.

[Wrecktangle]

What an absolutely ignorant statement. So Billy Walters doesn't use FRACTIONAL Kelly? It's unbelievable how much putrid shit you spew in this forum.

So you work for BW? Why make statements about something which YOU obviously know nothing?

I'll leave you and your fellow troll FD to fight on, as someone said: don't get into the muck with a pig, the pig likes getting dirty.

[Wrecktangle]

All well and good, but the OP didn't ask your opinion (or my opinion, or anyone else's opinion) about whether Kelly can "work" in the real world. He had a technical question about how to compute the Kelly stake, which you answered incorrectly.

Case closed.

My apologies to the OP and you. Moving on.

[MonkeyF0cker]

So you work for BW? Why make statements about something which YOU obviously know nothing?

I'll leave you and your fellow troll FD to fight on, as someone said: don't get into the muck with a pig, the pig likes getting dirty.

Yeah. It's obvious who knows nothing. Read Smart Money by Michael Konik once. I also know a few people who happen to work for him. More horseshit spewed from the mouth of Wrecktangle. Shocking.

[MonkeyF0cker]

Here's troll 1 having troll's 2 back...

BTW, bonehead: fractional Kelly IS NOT OPTIMAL by Kelly's def, so it's just another betting system that you blowhards are waxing on about.

Oh and do the math, Walters can't use Kelly, the market can't support it.

...have a nice life.

I just said he uses FRACTIONAL Kelly. What part about that don't you understand? When you and Wrecktangle get together to do your "simulations," does that include beating each other in the head with bricks? Good Lord. It's hilarious how the two of you insert yourself into every Kelly thread and fill it with absolute pig vomit. Not only do you try to trash a staking strategy that has been proven by professional handicappers in the past, but you actually think people are dumb enough to not take what you say at face value and actually believe you nimrods.

[Flying Dutchman]

I just said he uses FRACTIONAL Kelly. What part about that don't you understand? When you and Wrecktangle get together to do your "simulations," does that include beating each other in the head with bricks? Good Lord. It's hilarious how the two of you insert yourself into every Kelly thread and fill it with absolute pig vomit. Not only do you try to trash a staking strategy that has been proven by professional handicappers in the past, but you actually think people are dumb enough to not take what you say at face value and actually believe you nimrods.

OK, you win. Monkey-wacker and "Data" are so smart, they must be right.

EVERYBODY START BETTING KELLY!!!

...let's see your bankrolls grow. Since this works so well, you'll all soon be rich!!

...you know, Wreck actually gives a shit whether the noobs, lurkers, & assorted flotsam and jetsam on these forums lose their asses...

...I'd rather see you all fvks crash and burn...then we'll be entertained by your "bad beat" stories in PT!

[Maverick22]

You three should just kiss and make up

[durito]

OK, you win. Monkey-wacker and "Data" are so smart, they must be right.

EVERYBODY START BETTING KELLY!!!

...let's see your bankrolls grow. Since this works so well, you'll all soon be rich!!

...you know, Wreck actually gives a shit whether the noobs, lurkers, & assorted flotsam and jetsam on these forums lose their asses...

...I'd rather see you all fvks crash and burn...then we'll be entertained by your "bad beat" stories in PT!

scoreboard

[RickySteve]

Are the bets simultaneous in time during that day? If not, 1st goes off at Kelly, and the 2nd goes off at Kelly after the new bankroll has been re-calced with the result of the 1st bet.

If simultaneous then the highest Kelly fraction bet goes off at fully Kelly, then your bankroll is degraded by the 1st bet Kelly amount, and then the 2nd bet goes off at it's Kelly fraction of the remaining bankroll.

BTW, this is an excellent way to burn down a bankroll in record time if you are not absolutely sure of your Kelly fraction, and few (99%+ of posters in this, or any forum) are.

Since you are asking Qs like this, most likely you fall into the 99%+ group.

You should have your drinking water tested for lead.

[Dark Horse]

lol. You woke up the sect Wreck?

If only I could get a parlay down for any random RS post that it would be 1) useless and 2) negative. No need for fractional Kelly at all.

[Wrecktangle]

Yes DH, they do seem to like to pack up.

About fifteen years ago I went into shock when we found out that Kelly appeared to not be optimal as the result of some rather involved simulations. Never mentioned it until now. Actually I'm sort of glad they don't believe it.

But the real lesson to me was: don't take anything as gospel from any source. Verify it thru simulation. This was hard for me as Ed Thorpe was really into it, and in the BJ world he is a demi-god.

As it turns out, Kelly, while probably not optimal is actually a very good management scheme and can rapidly ramp up a BR until you hit market limits. But it is amazingly dangerous. If you you don't take extreme care to make sure you are drawing from populations that you have high confidence that you know the variance, if you don't re-calc your short window bayesians at every interval, if you don't recognize that your sport can change at any portion of a season to where your win% has changed and the advantage you had drilled out is no longer in effect, you can be rapidly screwed.

And when you've just lost 1/2 of your bankroll or more due to a "short term fluctuation" (happens a lot with Kelly) all sorts of stuff runs thru your head. Do you hold the course as this may only be a bad run of draws? Or has your population changed? The NFL this year is an excellent example of this: week-to-week fluctuations in rules interpretations plays havoc.

Of course "the sect" can manage this, just ask them, they can do anything, but I fear the casual punter cannot.

I'm sure Monkey et.al. will have a field day with this. Rant on.

[durito]

lol. You woke up the sect Wreck?

If only I could get a parlay down for any random RS post that it would be 1) useless and 2) negative. No need for fractional Kelly at all.

His posts have tons of useful info if you are willing to do some work.

[Dark Horse]

I know what the sect is selling, durito, and I'm not buying. Posts #14 and 25 are most recent examples of their communication skills. Without exception, they're data crunchers. They fail to see that the scientific (mathematical) principles they use were discovered by others. And those 'others' were willing to think outside of the box, and go beyond accepted norms, or they couldn't have added anything new to science. That is what this clash is about, and that is also why they will never be more than average at what they do.

[durito]

Average by what measure?

Isn't gambling about making money?

I'd bet a large amount that RS+MF have won combined significantly more than the other two.

[Dark Horse]

Average in the sense of closed-minded and stuck, in a world where many dynamic minds are working hard on redefining boundaries. (which makes it so amusing that they think they should protect the think tank from unwanted ideas).

And, as you surely know, success is measured in more than just money, my friend. An obviously unhappy camper, such as RS, would not be successful in my book if he had a billion dollars to his name.

[durito]

We are talking about success in sports gambling not life.

[Dark Horse]

We are talking about two approaches. One 'fundamentalist', the other 'pioneering'. The former have a major problem with the latter (not just in sports betting; pretty much across the board). In my opinion this has had a suffocating effect on the think tank. They don't like their dogmas challenged, and are highly uncomfortable with the unknown; because they already know everything... And yet, those who enrich not just themselves, but the field in which they operate, are those who successfully explore the unknown.

Example. How much more interesting could this think tank be if we tried, collectively, to come up with a better bet sizing formula than Kelly? The dogmatists will insist this is not possible. But what if it is? Why not explore the possibility? It's not like Kelly was a sports bettor. A project such as that, to me, would be a good use of the think tank. Get those creative juices flowing. If it leads nowhere, so what? If it does produces something better, everybody benefits.

But that kind of creativity is hampered, because the sect members, in their safety of numbers, show up every time something looks out of order, to nip it in the bud, making sure to collectively chew out the heretic.

[tomcowley]

How much more interesting could this think tank be if we tried, collectively, to come up with an even prime number greater than 2? The dogmatists will insist this is not possible. But what if it is? Why not explore the possibility?

[Dark Horse]

Yes, you already know everything. You forget, however, that science is always progressing. For many centuries math had no zero. Then they worked for many centuries without a decimal point. But all you can see is a fixed science. And if that works for you, that's all you need.