The greatest day of the year for sports bettors - Superbowl Sunday - is nearly upon us. Sportsbooks look to capitalize on the gambling frenzy by giving clients as many betting options as possible, including an enormous array of proposition bets. One area where bettors can really see an expanded selection is in "alternate" lines. Most games are offered to bettors with one line in the vicinity of -110 odds each way - for the Superbowl, Indianapolis -6.5 with a total of 48. For marquee games, however, sportsbooks will also post alternate point spreads and totals at varying odds. For example, 5Dimes offers spreads all the way from Indianapolis +42.5 through Chicago +42.5, and totals from 31.5 to 60.5.
More posted lines means more opportunity to gain an edge for wise bettors, and the more lines that are posted, the more likely it is for some of those lines to be "off". For handicappers with access to a database of past results, examining the point distribution of prior games can give an idea of what the odds should be on various point spreads.
I took a look at all NFL games lined 6, 6.5, 7, and 7.5 since 1994. This gave me a sample size of 668 games, with an average line of 6.71. This sample seems to approximate the upcoming Superbowl, where the line has hovered between 6.5 and 7. As a testament to the accuracy of NFL wide-market lines, the favorite in these games won by an average of 6.72 points, almost exactly equal to the average line. However, the median margin of victory was only 5. This means that the "true line" has historically been about 5.5 for these games (favorites went 331-337 versus a -5.5 spread), a result of the "favorite bias" in the NFL. Overall, the favorite in these games went 294-354-20 (45.4%) against the actual game spread.
The underdog has won outright 202 times of the 668 games (30%), making for a fair moneyline of around +230. Most sportsbooks are currently straddling that number. In the first half, the favorite has won by an average of 4 points, and the favorites have won the first half outright 65% of the time for a fair moneyline of around +185. Sportsbooks in general are posting lines right where expected for the first half.
To evaluate alternate spreads, I'll compare them to the historical expectation against the primary line. Indianapolis -6.5 is available, and based on database results that number is about negative 10 cents in expectation per dollar wagered at -110. Likewise, Chicago +7 -110 can be found by line shopping, and that wager has a positive expectation of about 6.5 cents. I'm not saying that these numbers are any measure of the actual expectation on this particular game, but they are the historical expectation of the data sample and we can use that expectation as a baseline to compare against the historical results of alternate spreads.
I looked through some alternate lines and compared them to the historical results to see if any of them offered a better value than the "main" spread of Indianapolis by a touchdown. I'll use prices that can be found at TheGreek.com, but it's important to do some hard-core line shopping to find the best number out there and boost your expectation by a few cents.
For Indianapolis backers, the moneyline at -250 has produced a historical loss of 2 cents per dollar wagered, better than the spread at -6.5 -110. In alternate spreads, the Greek has Indianapolis +10.5 -850. Based on the database, this is about a 92% cover rate, with a fair line of around -1100. The expectation on the Greek's line is only 2.6 cents, but that compares pretty favorably with the negative expectation on the -6.5 line. This line let's you pick up a ton of key numbers if you don't mind laying the heavy chalk.
For those favoring the Bears, buying the extra half to +7.5 -120 doesn't do much in terms of expectation, with a push rate of about 5% on the 7-point spread. Chicago backers may want to consider the Bears +3.5 +140, currently at The Greek, which would have covered over 45% of the time and returned 9.2 cents in expectation.
Based on the historical data, here are the "fair lines" I would assign to some commonly offered alternate game lines:
Moneyline: +230 Indianapolis -6.5 +112 Indianapolis -7 +126
Indianapolis -21.5 +595 Indianapolis -17.5 +415 Indianapolis -14.5 +280 Indianapolis -10.5 +175 Indianapolis -3.5 -120 Indianapolis +3.5 -385 Indianapolis +7.5 -715 Indianapolis +10.5 -1115
Of course, it is not quite that simple. First, this assumes that the current line is "correct". Second, there are other factors to be considered rather than just raw number crunching. One of the other factors you may want to consider is the game total. For our database sample, the average lined total of those games was 41, compared to the 48 offered in the current Superbowl. So you may expect the average value of a point to drop slightly due to the higher total.
Now let's take a look at the total. The line is a fairly solid 48 with some books moving towards 48.5, so I used a database of games with totals within 1.5 points on either side - all NFL games since 1994 with totals from 46.5 to 50. This gave me 268 games with an average line of 47.7. The average combined score in these games was 47.1, and the median was 46. As expected, we can see the slight bias towards the "over". The best "true line" was 46.5, with 50.7% of games going under that number. Overall, playing the under against this data set was an impressive 147-114-7, for a record of over 56% against the closing line. Not surprisingly, the first half in these games produced an average of 24 points, exactly half of the posted full-game total.
Over bettors have had an historical expectation of about negative 19 cents per dollar wagered (for betting into a 44% subset). However, you may be able to improve that by taking a look at some alternate totals. For those not scared to lay the chalk, over 31.5 at -500 looks playable, with an expectation of about 4.3 cents per dollar risked. If you think over 48 -110 is a decent bet, then over 31.5 -500 is even better.
Players siding with the under earned about 10 cents per dollar wagered in this historical sample. The primary line of under 48 was generally better than any of the alternate lines offered for this game. For under bettors, under 42.5 +200 and under 54.4 -210 are in the same ballpark if you want to take a shot with some alternates. For the bridge-jumpers, under 60.5 -360 is another option - of the 268 games in this sample, only 43 have gone over 60.5 points, for true odds of around -525.
Here are my "fair lines" for some commonly offered alternate totals based on the 268-game sample of totals "around" 48:
Under 48 -135
Under 31.5 +665 Under 37.5 +330 Under 42.5 +175 Under 54.5 -260 Under 60.5 -525
Overall, we can notice a couple of things about the pricing of these alternate lines. First, it would seem that the moneyline is a better bet than the spread for Indianapolis backers, while taking the points is superior for Chicago backers. It is probably no coincidence that this flies in the face of "common knowledge" of taking the underdog on the moneyline and laying the points with the favorite. Secondly, the numbers generally support the value being on the "heavy chalk" side of the bet, for example, over 31.5 -500. Since many bettors avoid risking large amounts to win small amounts, it makes sense that the "favorite" side of these alternate lines is underbet.
The numbers presented here can be used as a guideline for evaluating "rogue lines" offered at some sportsbooks. Books like 5Dimes will also offer many more lines than illustrated here. However, as our friends in the mutual fund industry like to remind us, "past performance is not a guarantee of future results". Remember to stay within your normal money management strategy, and be careful of choosing too many correlated "alternate lines" that can set you up for a big loss. This Superbowl Sunday, shopping around for alternate lines may help put a few extra dollars in your pocket.