Guide to Gambling on the Super Bowl

Warning, this post will probably only appeal to degenerate gamblers, economists and nerds; i.e. family and friends. The bottom of the article has my advice on prop bets for the Super Bowl, like the length of the National Anthem and coin toss, for those that want to skip ahead.

I’ve read quite a few websites about the most popular gambling mechanism for the Super Bowl (squares) and I’m convinced that nobody has published a definitive article on the best squares or devised a system for valuing different squares. For those that aren’t familiar with the game, a 10 by 10 grid is made such as:

Each square (100 possible) is then purchased. The scoring is done according to the last digit of each team’s score at the end of each quarter. For example; the score at the end of the 1^st quarter is Pittsburgh 0 and Arizona 7, the person with the corresponding box (bottom row, 7^th column) would be paid the corresponding amount. For our example we will use a 10%, 25%, 10%, 55% payout model for the 1^st-4^th quarters, respectively, and assume a price of $2.00 per square. Additionally, Pittsburgh is always listed as the first number in this article as they are the favorite.

This game is popular due to the randomness of picking squares, anyone can win, and the large groups that can participate, i.e. offices.

The need for more thorough research on this topic comes from people wanting to sell or trade squares after they’ve been assigned. If you paid $2 and received the (7,0) square, how much would you sell it for, compared to the (8,8) square?

My complaints with the popular articles on this game are that they typically place the odds on combined squares, making (0,7) and (7,0) the same, and when they do separate the squares, they almost always do it according to home team vs. away team. Since each league switches being “home” each year, the randomness does not help predict the timing of each teams scoring and assign values to each box. I built my model according to favorite vs. underdog, which should provide more enlightening results, and weighted the expected value of each square according to the quarterly payout ((0,0) is more likely to pay in the 1^st quarter rather than 4^th quarter and should be weighted accordingly). I took the quarterly results of every Super Bowl to get the frequency that each square comes up. I admit that taking Super Bowl results is probably less statistically relevant than the last ten years of regular season play, but with 168 total results (42 games with four quarters) and the fact that teams typically play different in championships than regular season; I think the results are relevant.

Here is the “value” of each square according to my model:

So the A-hole who drew the (7,0) square should not sell his square for less than $11.00. FYI, I drew squares (2,5) and (9,6), which combined are worth about $2.19, not a great ROIC. Here are some little known facts: 49 of the 100 squares have never hit in a Super Bowl, 10 squares make up more than 48% of the potential payout in this example and 35 squares comprise more than 90% of the expected value (a pretty standard bell curve).

I have the Excel spreadsheet for anyone with an interest in checking my work or seeing how valuable your squares are. I accept PayPal, most credit cards and bacon for payment.

Prop Bets:

Over/Under - I think the best money this year is on the Under. Most sites have the Over/Under set at 47 points. Considering 70% of the last ten Super Bowls have had fewer points scored the 47, and the fact the Steelers only had one game all year that was more than 47 points scored, I think the Under is a relatively safe bet.

National Anthem - This year Jennifer Hudson is singing the National Anthem with the over under set at 114 seconds. This is slightly above the average Super Bowl time of 105 seconds (I have results from 18 Super Bowls), but I think further research needs to be done before taking the Under. I broke down National Anthem times in Super Bowls into a few different classifications, Male-Female and Black-White. Males average only 94 seconds (“that’s what she said”) while females average 112 seconds, but a bigger discrepancy comes from comparing race. Black artists average 114.5 seconds while white artists average 97 seconds. The longest was Natalie Cole in Super Bowl XXVIII at 153 seconds and the shortest was Neil Diamond in Super Bowl XXI at 61 seconds. Considering Ms. Hudson’s race, sex and flair for the dramatic, I would take the Over. Just listen to her sing “And I Am Telling You I’m Not Going” and you will realize you could be listening to her this weekend for well over two minutes.

Winner - I would go with Vegas on this one and pick Pittsburgh. The bookmakers have been right 69% of the time. The only downside to Pittsburgh winning is having to put up with their insufferable bandwagon fans for the next few months.

Coin Toss - Tails has come up 60% of the time, and 70% of the last ten. If you are dumb enough to take this bet than you deserve to lose your money. I’m going with Heads, but hoping for a repeat of the Pittsburgh/Lions Thanksgiving Day Coin Toss Debacle, http://tomferda.today.com/2008/11/27/the-infamous-thanksgiving-day-coin-toss/. The NFC team has won 11 straight coin tosses, so I predict Pittsburgh will call Tails, it will be Heads and Arizona will elect to receive.