Double Down Method

[Dark Horse]

Show me one single pro gambler who hasn't lost ten straight. Could we, realistically, increase it to fifteen? Surely, over the years, fifteen straight losses must be within reach... Where does that leave the full Kelly bettor?! All but B-R-O-K-E.

The gaping hole in the Kelly approach is that it doesn't account for the fluctuations inherent within Time. Can these fluctuations be predicted? That, my friend, is the billion dollar question.

[TLD]

A key difference with Kelly. When I do chase my unit size goes down. This is because I don't want to be victimized by a long losing streak.

A loss, and all the more so a series of losses, results in a smaller bankroll. If you’re using Kelly, then all else being equal, your bet size decreases if your bankroll decreases. So Kelly does exactly what you are claiming is prudent—it has you betting less and less of your bankroll as your bankroll shrinks, to keep you in the game longer.

Compare that to a flat bettor, who is betting the same size every time no matter the previous results and the changed size of his bankroll.

But perhaps you could better spell out your alternative. What bet sizing method results in your still being able to make non-peanuts size wagers after the kind of extreme losing streak that Kelly allegedly doesn’t account for?

[Dark Horse]

Sorry. I merely participate in a chase or two to have some action during baseball season. No major theory behind my bankroll management for chases. It's more of a break between NBA and NFL, so no doubt that approach leaves much to be desired.

[donjuan]

Show me one single pro gambler who hasn't lost ten straight. Could we, realistically, increase it to fifteen? Surely, over the years, fifteen straight losses must be within reach... Where does that leave the full Kelly bettor?! All but B-R-O-K-E.

IIRC, if you go 17-15 it doesn't matter whether you win the first 17 and lose the last 15, lose the first 15 and win the last 17 or any other mixture of 17-15; you still end up with the same exact bankroll using Kelly.

[Dark Horse]

Oh, I forgot another reality.

Psychological, emotional, you know.

If you start out with 100K and lose 15 straight, betting full Kelly, where does that leave you? Not just financially. Where does that leave YOU?

Is that important? Yes. If only because it is highly questionable that you can make the same quality wagers once your mental clarity is affected.

[TLD]

Again, what method of flat betting—or other bet-sizing method—avoids those emotional, psychological, and financial consequences of losing fifteen consecutive bets?

[Ganchrow]

If you start out with 100K and lose 15 straight, betting full Kelly, where does that leave you? Not just financially. Where does that leave YOU?

Is that important? Yes. If only because it is highly questionable that you can make the same quality wagers once your mental clarity is affected.

Over 10,000 bets the probability of a 55% picker suffering at least one 15+ game losing streak would be about 3.393%. If this picker were a full Kelly bettor wagering at -110 he'd be staking 5.5% of bankroll per bet (wagering limits not withstanding).

After 15 consecutive losses his bankroll would shrink to (1+5.5%)¹⁵ ≈ 42.803% of its previous value, a monstrous drawdown of about 57.197%.

Yeah, that'd be a bad day. A very bad day. I know I'd be pretty upset.

But just to get a little perspective, however, let's also look at the flip side. This same 55% full-Kelly bettor would experience at least one 15+ game winning streak over a 10,000 bet series with probability of about 43.639%. After these 15 consecutive wins his bankroll would have grown to (1+¹⁰/₁₁*5.5%)¹⁵ ≈ 207.89% of its initial value. That'd be a good day.

[Dark Horse]

I imagine that, for most people, it would be harder to deal with the losing streak than it would be exhilarating to experience the winning streak. You can only get so high (from things in this world), but there is virtually no limit to how deep a man can sink. I don't really care to test the boundaries of those dark depths. Life is challenging enough as it is.

[Dark Horse]

I should add that my goal is to enjoy life, and have plenty of money to do so. It is not to make as much money as possible, to the point where I don't enjoy life. So I always try to translate theory into real life. Amazing how many things look great in theory, but lack in that transition department.

[donjuan]

DH,

That's fair enough if your utility curve is like that. It's still optimal for you to bet some fraction of Kelly in that case, though.

[curious]

anyone ever try out the double down method?.... every time you lose you double down and bet the total amount that you're down until you win, thus bringing you back to even. I'm thinking about trying this with Sebastian's 100 star insiders. I was thinking about starting at $100. By the time I got to my sixth game, I'd be putting up $1600. I can't see him going 0-6 on insiders.....

Hey Lunds, have you been following the insider plays for a while? Have you had any luck with them? I haven't looked at sebastian's plays lately. I followed them for a while when he first started putting them out. You can guess why I stopped.

[RageWizard]

Are you suggesting that streaks don't exist, or merely admitting to your ignorance of them?

You will need alot of money to get over a bad losing streak. Probably like at least 10 plays. I bet even that wouldn't be enough. I know no-ones lost 10 times in a row. Just try it and you will see how close you come to losing that many times in a row.

[Ganchrow]

I know no-ones lost 10 times in a row. Just try it and you will see how close you come to losing that many times in a row.

My Streak Calculator figures the probability of losing a given number of times or more in succession over a stated number of trials.

For example, given a win probability of 50% your probability of losing 10+ times in a row at least once would be:

• 4.412% over 100 trials
• 21.443% over 500 trials
• 38.530% over 1,000 trials
• 91.361% over 5,000 trials
• 99.257% over 10,000 trials