Steve Fezzik is one of the top sports handicappers anywhere. And his specialty is betting NFL football. He’s won several NFL handicapping contests. He’s the brains behind the LVASports.com site. I was a guest on his (now history) radio show a few times and once he was a guest on Gambling with an Edge.
In 2011 NFL football he lost big time. Historically he’s won at a 54% rate (quite high enough to become wealthy if you bet big enough) but his 2011 picks lost money. In the December 23, 2011 column on the Las Vegas Advisor Question of the Day, he admitted as much. Many handicappers tout their wins. Few admit their losses. Fezzik did. I admire this honesty.
His ass-whupping was a bit of a surprise to me — but not a big one. There is variance in gambling and sometimes the variance goes against you. As the saying goes, sometimes you bite the bear and sometimes the bear bites you.
Winning at a 54% rate means that you have an 8% edge, or in the way the number is expressed in video poker circles, you’re playing a 108% game. Often 100.4% is as big an edge as I can find for the stakes that I want to play. Occasionally only 100.25%. On the very few occasions I have found an 8% edge I had to keep pinching myself to verify I hadn’t died and gone to heaven.
My article today is about the following: If a guy having an 8% edge occasionally has a losing year, what chance do I have with only a 0.4% edge to avoid losing years some of the time? My sample of data from 1993 through 2011 shows seventeen winning years and one losing year, but this is too small a sample to reach a definite conclusion.
My belief (ego?) says that I have a better chance of ending up ahead in 2012 than Fezzik does. How can this be? His edge is 20 times as large as mine.
The critical difference between betting the NFL and betting on video poker that makes me think this way is the number of betting occasions. In the NFL, there are 256 games over the season prior to the playoffs. That’s the maximum number of games anyone can bet — and it’s fair to assume that Fezzik took a pass on many of those. Sometimes he’ll think the over/under line has good value — which is a separate bet — bringing the sample size up to a maximum of 512. Let me assume for now he made 200 bets for the entire season. (Even if the true number turns out to be 100 or 350, it doesn’t affect my analysis significantly.)
Betting 200 times is 15 minutes work for me. I don’t play nearly as much as I used to (teaching, writing, radio show, square dancing, etc.) but I still play 800,000 base hands per year. Sometimes these are Triple Play through Hundred Play, so the total number of hands is likely closer to 5-10 million. And I usually have a 0.4% or larger edge on each of them.
The variance in video poker bets is larger than the variance in football bets. In football the wins and losses for most bets are about the same size — although this isn’t true for parlays and proposition bets. I don’t know what percentage of Fezzik’s bets are “straight” (meaning he lays 110 or so to win 100) and how many have potential wins significantly different in size from potential losses. In video poker an occasional 4,000-coin win is offset by numerous 5-coin losses. In my world, I expect to collect $80,000 a year from about 30 drawings — some are really big long shots — and sometimes those wins are included to make the edge as much as 0.4%. I don’t think Fezzik has anything similar in his NFL score.
Neither of us has a constant bet size. I regularly bet $10 per play at one casino and at others I bet $125 or $250. Betting sides or totals I suspect Fezzik can easily get $10,000 or $20,000 down if he wants and some propositions are limited to $500 or less. For both of us, if we do good or bad on our “big bets,” it swamps the small bets. But still, the small bets add up.
I can’t come up with an accurate estimate (is that an oxymoron?) of what my chances will be in 2012 to come out ahead of the game. Or even know for sure what games I’ll be able to find. Perhaps 100.2% will be as high as I can regularly find. And I have even more uncertainty about Fezzik’s chances. But I’m firmly convinced that a small edge multiplied by hundreds of thousands of independent events adds up to more certainty-of-being-positive than an edge 20 or 40 times as large multiplied by only 200 independent events.
When I say I have a better chance at coming out ahead in 2012 I mean he has perhaps an 80% chance and I’ll have an 85% chance. I’m pulling both those numbers out of a hat. The smartest bet will be that both of us come out ahead next year — depending, of course, on the odds we’d have to lay to make that bet.
If you think this article is putting Fezzik down in the slightest, you’re reading it wrong. Fezzik is one of the very best at what he does. I wish I had the skills to regularly obtain the advantage he gets. The entire point of this article is that a small edge spread over millions of occurrences provides more of a certainty of being ahead than a much larger edge spread over a few hundred occurrences.