Pace and Cap  - Sartin Methodology & The Match Up

Pace and Cap - Sartin Methodology & The Match Up (http://paceandcap.com/forums/index.php)
-   General Discussion (http://paceandcap.com/forums/forumdisplay.php?f=5)
-   -   Kelly Criterion (http://paceandcap.com/forums/showthread.php?t=11943)

mick 11-04-2018 09:41 AM

Kelly Criterion
 
1 Attachment(s)
I'm reading an excellent book titled "Fortune's Formula" by William Poundstone. It's about John Kelly, Ed Thorp, Claude Shannon, beating blackjack and the casinos, organized crime, the stock market, and a myriad of other interesting topics. It's a big, sprawling book, but there's a single thread that weaves itself throughout and that's the Kelly Criterion. Wikipedia defines it as "a formula to determine what proportion of wealth to risk in a sequence of positive expected value bets to maximize the rate of return." It's the creation of a Bell Labs scientist, John L. Kelly, who died of a stroke on a Manhattan sidewalk at the age of 41 in 1965 and apparently, never used his own criterion to make money. A lot of other people have used it though, including Warren Buffett, the managers of many hedge funds, most good poker players and some horseplayers, too.

In my more disciplined periods, I've used a crude version of the formula based on my own hit rate and not some odds line that I tried to create. (I never was any good at creating an odds line.) My method worked but then I'd get bored and decide I wanted to play superfectas or some other form of craziness. Anyway, inspired by the book, here's a spreadsheet I created that hopefully explains it in a simple form, betting one horse to win.

If you know your hit rate (and that's easy enough to determine) and the tote odds, you can determine how much of your bankroll you need to bet to maximize its growth and also avoid gambler's ruin. Obviously, you only need to be betting when you have a positive "edge" and the range of tote odds needs to be something reasonable. My chart wasn't designed for bombs. :rolleyes:

I'm mulling over a second version that will address two-horse hedge betting, which is what I do most of the time. But for now, one-horse betting is enough. Hope you find it interesting. (The highlighted lines represent break even, of course.)

mick 11-04-2018 10:43 AM

Clumsy phrasing on my part when I wrote, "The highlighted lines represent break even, of course." What I should have said is "The highlighted lines represent the hit rate and tote odds where you no longer have an edge and shouldn't bet."

noddub62 11-05-2018 03:14 PM

Interesting read I thought.

Budman

mick 11-05-2018 07:35 PM

Quote:

Originally Posted by noddub62 (Post 118324)
Interesting read I thought.

Budman

Thank you, Budman. I'm glad you found it interesting. My friend Richard D (rdiam) says it should be a "fractional" Kelly and with an MBA from Wharton, he's a lot smarter than me.
I've tried to make odds lines, wasn't any good at it, and through trial and error, backed into a percentage that worked for me. But I was only betting in the tote odds range of 5-2 to 5-1, knew my hit rate and that's why I suggested using your own win percentage instead.

The Kelly Criterion isn't part of the Sartin Orthodoxy but it is useful in scaling up or down your bet size as your bankroll grows and sinks.

Dorianmode 11-05-2018 09:38 PM

There is a long discussion of the Kelly in #2 of the (early Sartin) "Follow Ups" *magazine*, (available in the library). Doc brought in Huey Mahl (I think) to discuss wagering, and he presented the Kelly Criterion. You may find it interesting. A number of financial/investment outfits use it. It's interesting also because originally Bell Labs developed it and used it to calculate the most efficient way/rate to transmit electricity in power lines ... (as I recall).

If you find the discussion in Follow Up #2 useful, PM me ... I have worked out all those formulas in a spreadsheet, which can be adjusted to your personal wishes/numbers. I've shared it before here, and am happy to again. You're right. Most people who use it, use a fractional Kelly. If your win rate is medium to high, the Kelly will spit out a huge fraction of the bank to bet, and it's just not a comfortable or prudent thing to do.

mick 11-06-2018 01:21 PM

Quote:

Originally Posted by Dorianmode (Post 118327)
There is a long discussion of the Kelly in #2 of the (early Sartin) "Follow Ups" *magazine*, (available in the library). Doc brought in Huey Mahl (I think) to discuss wagering, and he presented the Kelly Criterion. You may find it interesting. A number of financial/investment outfits use it. It's interesting also because originally Bell Labs developed it and used it to calculate the most efficient way/rate to transmit electricity in power lines ... (as I recall).

If you find the discussion in Follow Up #2 useful, PM me ... I have worked out all those formulas in a spreadsheet, which can be adjusted to your personal wishes/numbers. I've shared it before here, and am happy to again. You're right. Most people who use it, use a fractional Kelly. If your win rate is medium to high, the Kelly will spit out a huge fraction of the bank to bet, and it's just not a comfortable or prudent thing to do.

Thank you for directing me to Follow Up #2. Yes, it was Huey Mahl and I found it very interesting.

mick 11-06-2018 03:31 PM

To say that Mr. Mahl's article is "interesting" really doesn't do it justice. It's exceptional. Thank you, Dorianmode.

I've read the more recent "Follow Ups" but not the early ones and didn't realize what I had missed. My favorite part deals with the Kelly Criterion and dutching two horses. No need for me to go there. It would be like me trying to re-invent a wheel that had been made by the better craftsman.

Lefty 11-06-2018 07:02 PM

I agree, Huey did exceptional work on the "Kelly"

mick 11-08-2018 10:34 AM

For anyone who's interested, Mr. Mahl wrote additional articles in Follow Ups 3, 4, and 7.

Brian Chamberlain 11-09-2018 12:04 PM

Mick, I agree, Fortune's Formula was a fascinating read. I am a big fan of Ed Thorpe, more so on his investing style as I am in the business.

Of course a game of blackjack now and then is more fun knowing at least basic strategy. :)

Cheers,
Brian


All times are GMT -4. The time now is 03:50 AM.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2024, vBulletin Solutions Inc.