Intro to Bankroll Management (9/7/10)

Play Value Calculator and the Kelly Criterion

Tuesday, September 7 at 11:18 AM ET
This will probably serve as the first of many discussions about optimal bankroll management as it relates to sports betting and, in particular, our Play Value Calculator recommendations.

I received several great questions through Support Contacts from some of our subscribers about our success metrics in units and how to use the Play Value Calculator.

We don't mind talking about "units" or actual dollars as we have to have some kind of baseline metric for our success. For instance, for a normal $50 player, our college against-the-spread (ATS) and Over/Under (O/U) picks using our Play Value Calculator recommendations were +$322 in the first weekend of college football. And just the plays mentioned as part of Paul's Top Picks were +$166. However, everything is relative. Everyone has different bankrolls, levels of risk aversion and comfortableness in how many different plays can be in place at one time.

To address the last point first, the true answer is all about the value available. A very general rule of thumb that we like to use is to play games above 60%. Sometimes there are several of these at one time and sometimes there may be none. Diversification can definitely be good because we've never seen a guaranteed game, but it only makes sense to play the games that return the most value. And it always makes sense to follow the Play Value Calculator's suggestions relative to wagers on multiple games at one time.

With respect to our expectations, in three years of tracking his NFL regular and post-season picks, I have accurately picked 59.7% (429 - 289 - 172) of ALL games against-the-spread, including hitting 78.3% (49-11-2) in the top plays of the week each week during football.

That being said, our expectation would be that the game has the likelihood of covering that our pick says. I don't mean that to sound too obvious or facetious. And, for what it's worth, our picks have generally lived up to that and more. That's just the best way to look at our picks. Even if picks with 60%+ confidence have hit at a rate well above that, the next 60%+ play is only expected to be as accurate as it's cover percentage. The play value recommendations with our confidence are essentially our conservative take on the "Kelly Criterion."

The Kelly Criterion assumes a bettor knows how accurate he is. But, even if a bettor is normally 60% accurate, it doesn't necessarily mean that that accuracy applies to the next pick.

That's where our confidence through the cover percentages comes in. We know exactly how often the pick covers in our simulations.

We recommend normal plays of 1-3% of your bankroll. "Normal plays" start at about 57% and everything revolves around that with the Play Value Calculator. "Weak" plays would typically mean less than 1% of a bankroll. When played individually, 2X and 3X plays could be between 2-9% depending on what is a comfortable Normal play for the wagerer.

When multiple plays are in place at one time, that changes the equation a little bit because that eats into more of the bankroll at that time. We allow this to be considered in the Play Value Calculator.

We say it is a conservative take on the Kelly Criterion because we need to better account for the market and its take on each pick (not an original approach - as most would suggest that the Kelly Criterion is a little more aggressive than "optimal").

And, as usual, please be conscious of lines with every picks. Especially early in the season, lines can change significantly and frequently so it is important to be sure that plays are appropriate to the line. This is why we offer the Customizable Predictalator to run games using any line and see our updated picks.

I hope that helps and I'm sure we will revisit this topic again. Please do not hesitate to contact me with any questions through a Support Contact.

Comments

Search Blogs

Keyword:

Archive

Ad

The Predictalator

The Predictalator is the most advanced sports forecasting software available today because it has the ability to account for all of the relevant statistical interactions of the players (playing or not playing/injured), coaches, officials, fans (home field advantage) and weather in each game.

04/21/2014 Highlight: Using the ResultsFinder for the week of April 14th - 20th, one could find that all highlighted, MLB "normal" or better picks went 22-16 (58% ML, O/U, RL). A normal $50 player utilizing our play value recommendations on these picks returned a profit of +$425 for the week

The week in the NBA was even stronger, particularly with the start of postseason play. Overall, for the week, normal or better picks went 3-1 (75% ATS and O/U), including starting the Playoffs 2-0 with such plays. New this, we have added halftime picks which have proven capable of providing strong opportunities to exploit the market. Over just eight NBA Playoff games thus far, halftime normal or better (and there are "better" halftime picks) against-the-spread plays were especially stron,g going 5-0 (100% ATS).

As the NBA and NHL postseason began, all highlighted, "normal" or better picks on the site, including halftimes, went 40-23 (63% ATS, ML, O/U and PL) for the week.

The Predictalator plays every game 50,000 times before it's actually played. This provides us the ability to assign probabilities to the likelihood of just about any outcome occurring in any contest including straight-up, against-the-spread and over/under winners of each game.

Ad
Ad