Section 1: Sports Betting as an Investment
Making Money by Betting on Sports
Most people think that sports betting at is about finding ‘sure things,’ but in reality such ‘locks’ are nothing more than gamblers’ fancy. Just as in real estate, currency, stocks, or any other speculative market, ‘sure things’ simply do not exist. As a professional sports bettor, my goal is to find and exploit many small edges over a long period of time to earn a compounding return. Winning 55% of games is very significant, and with very conservative bet sizing, you can grow your return very quickly. Investing $10,000 into the stock market for a year and earning a 10% return is considered a great investment – but your return winning a modest 54% of your sports bets would trounce that return.
My picks have yielded a much higher risk adjusted return than the stock market. Obviously, the variance from season to season is formidable, but as anyone who had a significant amount invested in stocks or real estate in 2008 can tell you, such swings aren’t limited to sports. In the long run, my edge in what I do is far greater than the edge that you could hope to gain in any other speculative market.
Juice, and the power of 55%
Even though most sports bettors are losers in their own right (as a whole, bettors actually win an average of only 48% of their bets – less than they would expect to win if they just flipped a coin for every game), their losses are compounded by the fact that the house takes a cut of winnings, also known as the ‘juice’ or ‘vig.’ Most sports books charge a 10% commission on wins, which means that a bettor must actually win 52.4% of his games just to break even. (Wagering $100 per game, a bettor loses $100 with a loss and wins $90.91 with a win, so he must go 11-10 (11/21 = 52.38%) to break even).
In order to beat the juice and win in sports betting, a bettor must employ a disciplined approach in their analysis of each game using methods that have proven to be successful in the long run. I discuss my math models and analytical metrics in my Handicapping Methods essay, but you must realize that only the best and most knowledgeable handicappers can win more than 52.4% of their games. In their 2007 two page article about my handicapping success, the Wall Street Journal wrote, “…fewer than 100 people in the world can sustain win rates of 55% over time. Most of them belong to professional betting syndicates that hire teams of statisticians, wager millions every week and keep their operations secret.”
Touts often claim to be able to hit 60% or higher, but as I explain in my essay on Bayesian Probability, anyone who tells you that their long term expected winning percentage is higher than 60% is deluding themselves. Fifteen to twenty years ago a sharp handicapper could win about 60% long term but those days are over, as odds makers have become more savvy in the past decade or so. For a bettor to claim a greater than 60% long term expected win percentage, that would be mean that Vegas would have to consistently release lines with egregious errors, and that simply just does not happen often enough nowadays for claims of a greater than 60% long term expected win percentage. Any short term win rates of around 60% or higher are simply due to blind, short-term luck. For instance, in 2016, the first season using advanced metrics from play-by-play data, Dr. Bob Sports’ NFL Best Bet sides were an incredible 66-26 (71.7%), but that record was enhanced by winning a very large percentage of close games (31-12 on Best Bets decided by 7 points or less) rather than splitting the close ones. It still would have been a great season on NFL Best Bet sides (62%) if the close games had been 50% but I couldn’t expect the new model to win 60%-plus on sides based on that one season – although the play-by-play model back-tested at a very profitable 56% winners and the NFL Best Bets have been 57.0% since that season.
I often hear amateur gamblers erroneously claim that winning 55% of games isn’t even enough to beat juice. As demonstrated above, a bettor only needs to win 52.4% to break even, and a 55% bettor will be very profitable in the long run if they pursue an optimal money management strategy.
Of course, as in any game of chance, there is variability in the actual results and just because you have won 55% in the past and expect to win 55% in the future doesn’t mean that you’re going to win 55% this upcoming season. There is variance in sports betting, as there is in most investments, and I calculate the standard deviation to figure out how much of my bankroll I can safely wager on each game during the season to accommodate potential negative swings while having very little chance of exhausting my bankroll. I have extensively quantified the variance that exists in sports betting, and use mathematical formulas to dictate the exact optimal amount to invest so as to maximize the ratio of profits to variance.
My long term percentage on College Football Best Bets is 55% (53.9% on Strong Opinions) and NFL Best Bets are 57.6% (368-271-6) with on Strong Opinions at 55.4% (263-212-11) since the NFL play-by play model was introduced in 2016. I will continue to use 55% winners as a realistic goal going forward. If I expect 55% on 250 Football plays (I had 258 last year at 57.8%) then the expected profit at -110 odds would be 250 x (0.55 – (1.1 x 0.45), which is +13.75 units. The Kelly Criterion recommends a wager of 5.5% of your bankroll for a wager with a 55% chance of winning and odds of -110. However, the Kelly formula assumes sequential betting and sports betting usually involves simultaneous betting, which is part of the reason behind using some fraction of full Kelly to reduce risk. If I play 3.0% of my initial bankroll per bet (i.e. flat betting) then my expected return during football season (5 months) is 41.3%. Adjusting your bankroll after each week rather than flat betting will increase your expected return, as explained in the KC simulation section of my money management section.
A basketball season with 54.5% winners (the College Basketball plays are 1459-1217-45 in 3 seasons using the current advanced metrics model) on 1000 bets would on average yield +44.5 units ( (1000*.545) – (1000*.455)*1.1 ). Using a conservative 1% of bankroll per bet (full Kelly at 54.5% at -110 odds is 4.45% of bankroll) results in an expected return of 44.5%. So, despite a lower overall winning percentage and smaller average wager size, a season’s worth of basketball wagers is a bit better than a season of football because there are so many more bets in basketball season.
Money Management is as critical to a sports investor as picking winners. I have devoted many hours of careful analysis and math to optimal money management systems, which I have painstakingly outlined in my Money Management articles. Sports betting is more high risk (higher volatility and standard deviation of return) than stocks, but also results in a higher return if you follow a proven long term winning handicapper (of which there are very few).
My Money Management articles outline how to adjust your bet sizing based on your goals (expected return vs. probability of positive returns), your investment length (one season or many), your growth preference (flat or compounding), your risk tolerance (high or low) and the proportion of your overall bankroll.
It is always better to set conservative expectations to avoid over betting.
Factoring in the Cost of my Service
The cost of my College football service is $1295, the cost of the NFL service is $1495 ($2,595 for both services), my Basketball service is $2,495 ($4,695 for all Football and Dr. Bob’s Basketball service). You must factor in that cost when calculating your expected return on investment (ROI). As explained above, winning 55% on the Football plays and 54.5% on my Basketball plays would yield an expected profit of 85.8% on flat-betting using your initial bankroll. Using an optimal betting strategy, as explained in the advanced money management section, would yield even higher long term returns while protecting the downside risk in the inevitable negative variance seasons that plague even the best long term handicappers.
If you had $20,000 that you could comfortably afford to risk as your sports wagering bankroll and $4,695 went to pay for the all Football and all Basketball service, then you would have $15,305 left for wagering. As explained above the expected return on the combined Dr Bob Football and Basketball services is +85.8% per year (using a less optimal flat betting approach), which would result in a return wagering profit of +$13,132 on the $15,305 initial bankroll. The overall profit, after factoring in the cost of the services, would be $8,437 (($15,305 x 0.858) – $4,695 = +$8,437), which is a very good 42.2% expected return on your $20,000.
That percentage return is higher for higher bankrolls and lower for lower bankrolls since the cost of service becomes a smaller percentage of higher bankrolls and a higher percentage of smaller bankrolls. If you want to subscribe to the all Football and all Basketball package you would need a total of at least $10,200 to invest to expect a positive return after factoring in the cost of the service. The calculations above are based on expected results based on long term records and some years are better and some years are worse.
What is a Point spread?
Before I delve into rigorous explanations of how a bettor can gain an advantage against the point spread, it is important to understand what the spread actually represents. Point spreads were invented to keep bettors interested in games between teams of different talent levels – if a perennial powerhouse like Alabama plays a mid level team such as South Alabama, you’ll find very few people willing to bet on which team will win the game since Bama would be such a prohibitive favorite. However, most are willing to bet on whether Alabama will ‘cover the point spread’ and win by a certain number of points. If the point spread is 21.5, then the Crimson Tide must win by 22 or more points for their side of the bet to cover, while South Alabama must either win outright or lose by 21 points or less to cover their side. Point spreads are designed so that the probability of each outcome is roughly equal, and are generally set so as to approximate the median score differential between the two teams at the given site of the game.
However, skewed public perception, results-oriented analysis, and unsound metrics result in point spreads that are often biased one way or another. While the casual bettor does not possess the capacity to exploit these advantages, I have used mathematical models, situational analysis, significant trends and quantitative player analysis that are far more complex and accurate than anything else on the market to gain an advantage, which is why I have won 53.5% to 56% of my Best Bets (depending on the sport) over the last 34 years.
How are the lines set?
While the odds makers do to try approximate the median margin of victory between two teams, they also try to reduce their exposure to risk by setting lines such that money will fall evenly on both sides of a game, so that they can offset the bets against each other and earn a profit on the juice (cut of winnings taken by the house, explained below) without exposing themselves to large potential losses. Thus, odds makers are often in the business of gauging public perception rather than team performance, and therefore the betting public actually sets the line. In more recent years, the betting public has had less influence on the odds than professional betting syndicates or sharp money has had, but there is still value to be found – although in different ways than in previous decades. If Georgia is 4 points better than Georgia Tech according to my advanced metrics and analysis, but the aggregate public perception is that Georgia is 7 points better than Georgia Tech, then the posted point spread is likely to be closer to 6 or 7 points (public perception) than it will be to 4 points (the realistic difference between the teams). This makes my job as a professional handicapper much less daunting; not only can I exploit lines where the odds makers are off, but I can also exploit the uniformed opinions of the general betting public, and more recently take advantage of betting syndicates and ‘quants’ that rely more and more on algorithms but can overlook some of the hidden value in changes in team personnel or lineups and in the particular match up between two teams.
Isn’t gambling risky?
I don’t believe that the term ‘gambling’ applies to what I do. I sell information to subscribers, with which they can take positive expectation positions in uncertain markets. With correct financial optimization and bankroll management, long term risks are nominal compared to the risks of investing in other, more conventional markets. Just as a single stock may go up or down in a day, any one team may win or lose a given game. But as long as the investor maintains a long-term perspective, understands variance, and doesn’t over-extend themselves or bet more than they can easily handle, risk can be highly mitigated, and they can earn a very attractive risk adjusted return.