Wednesday, June 20, 2018

How much to wager - Kelly Criterion


When it comes to wagering on a sporting event, you have several decisions to make. First, the obvious. What side to take? Then, the type of wager. Generally simple, a straight bet. Finally, and very important, how much to wager? This final decision seems simple but can be complicated and quite important long term. It’s this final issue I am going to deal with over the next several blog posts.

There are two types of approaches that I will compare. First is flat betting. Simply bet the same amount on every wager. Second is varying the amount wagered based on how confident you are about your first decision, who to take. This will be referred to as Kelly or progressive wagering. Kelly refers to a specific approach of computing a percentage of your bankroll to wager based on the probability that you will win the wager and the odds. Progressive is an approach of betting more on some wagers with higher expectations without going into heavy calculations.

This project will compare flat betting versus a Kelly approach. You can find several descriptions of the “Kelly Criterion” by doing a Google search. Following are two good examples.


When I first discovered Kelly several years ago, I was quite excited. I implemented it in some of the sports wagering systems that I was using. The most vigorous of these was in the 2012 and 2013 MLB seasons. The 2012 season was one of my better ones, I returned $1.089 for every $1.000 that I wagered. After the season, just out of curiosity I computed the results based on wagering the same amount of money on flat bets. To my surprise, the return would have been $1.094. When I did the same for 2013 (essentially a break-even season), the Kelly return was $0.993 versus $1.007 for flat betting.

In my back testing, I would periodically compare Kelly to flat and generally found similar results. I finally concluded that flat betting was probably better than Kelly when you look at return per dollar wagered. I subsequently have found anti-Kelly stuff on the web.

Add to these contradictions discussions on sportsbookreview.com in the Handicappers Think Tank sub-forum. There are some folks there who disagree with my conclusion. Since I highly respect some of their opinions, I have doubts about my own conclusions. I’ve decided to finally resolve the issue in my mind by taking several approaches. These will be discussed in future posts. Follow me on Twitter, @ole44bill, to know when these posts occur. Also feel free to comment or criticize on the blog, or better yet on the Handicapper Think Tank. I’ve started a topic there on this project called “Kelly Re-Visited”.

No comments:

Post a Comment

Run line analysis update

I looked back and had very slight profit on run line wagers in 2018. So, I decided to update my run line analysis from a year ago. I pos...