Parlays are among the favorite bets of recreational gamblers. There is a natural attraction in being able to win more than you've put at risk, in being able to make a big score on a ticket costing only a few dollars. And sportsbooks play along, giving the punter fair odds on parlays, at least for those of 3 teams or less. In general, you can think of a parlay as letting your bet "ride". If you bet $10 on a 2-team parlay, it is exactly the same as betting $10 on the first team, collecting your $9.10 in winnings, and then letting the original bet plus winnings ($19.10) ride on the second team. That would give you a total win of $26.48 on your $10 bet, or approximately 13-to-5 as offered by most sportsbooks on 2-team parlays.
As is, the odds on parlays are generally fair (although you'll usually pay a penalty for parlaying 4 or more teams), but they don't offer any advantage above beyond any advantage associated with the individual bets. However, there is a way to make parlays work to your advantage if the individual bets are correlated. Correlated means that if one leg of the parlay comes in, it becomes more likely that the second leg will also come in. In other words, the two bets are in some way related.
A great example of correlation can be found in the 2006 NCAA football season. In the second-to-last game of the regular season, the Temple Owls travelled to Happy Valley to challenge the Penn State Nittany Lions. The hapless Owls were installed as prohibitive 36-point underdogs, with the total for the game set at 46.5. The chances of a favorite/under parlay ticket cashing in this case is very slim - if Temple managed to score just a single touchdown, that parlay would have to lose. And if Temple just scored a field goal, Penn State would have to score exactly in the range of 39-43 points for a win (or at least for a push on the side and a win on the total). The bottom line is that particular combination of favorite and under requires a very narrow set of outcomes to end up a winner, and getting only 2.6-to-1 odds on that happening is not a very good bet. Conversely, if Penn State cover the 36-point spread, we can be fairly certain that the game will go over the posted total for the same reasons.
A normal parlay assumes each event is independent, and that the odds of either side winning are 50%. This predicts that, for any given 2-team parlay, each possible combination has a 25% chance of winning. Betting 25% shots at 2.6-to-1 is a slightly losing proposition. However, in the example of the Penn State vs. Temple game, the real distribution might be something like this.
Penn State/Under: 10% Penn State/Over: 40% Temple/Under: 40% Temple/Over: 10%
Notice that Temple and Penn State both are 50/50 propositions and over/under are likewise 50/50 props. So while the individual lines are "correct", the combinations are skewed, making some parlays advantageous and others terrible bets. In this hypothetical example, being on the right side of the correlation gives you a 40% chance of winning a 2.6-to-1 bet, or a profit of 0.6 for every 3 bets you make. It also should be obvious that playing both sides of the correlation has merit. The Temple/under and Penn State/over bets are both advantage wagers. Playing both means you are risking 2 to win 2.6, but you have an 80% chance of winning.
Football is one of the best sports for correlated side/total parlays due to the relatively low totals and occasional high pointspreads. The consensus seems to be that favorite/over and underdog/under parlays start to become valuable when the point spread is about 40% of the total. These situations are rare in the pros, but fairly common in NCAA football. Correlations also hold for bets on halves and quarters, and sportsbooks that allow side/total parlays on quarter and half lines should hold plenty of value for sharp bettors.
Correlations can also be found in plenty of other sports. Any situation with low totals and high spreads is a possibility. This includes runline/total parlays in baseball and puckline/total parlays in hockey. Another obvious example of correlation is betting the same team to cover the half and to cover the game. Unfortunately, most sportsbooks have caught on to the idea of correlated parlays and restrict their clients from betting them. The fear of correlated parlays is so great that some sportsbooks have disallowed same-game parlays altogether. This means players have to be a little more creative in coming up with parlay candidates.
One example of a unique correlation often occurs during week 17 of the NFL season. Imagine Washington and Minnesota are in contention for the last playoff spot, but Washington has to play at 1 PM while Minnesota has a 4 PM game. If Washington loses, Minnesota is in the playoffs. However, if Washington wins, Minnesota must also win to keep their playoff spot. Parlaying Washington with Minnesota may be correlated in this case - if Washington wins, Minnesota will have more incentive to play hard, and if Washington loses, Minnesota may end up not even playing their starters at all. This is a perfect example of correlation, as the results of the first game will have a significant impact on the results of the second game.
Numerous other examples of correlation are possible. Of course, those that aren't so obvious are the best kind, as many sportsbooks may still allow wagers on more creative correlations that haven't been recognized by the general public yet. The best way to evaluate correlated parlays is to pose the question, "If this one bet wins, does it increase the chance of another bet winning?" If so, then there may be a correlation there to be exploited.
"The Complete Square's Guide to Sports Wagering" is a recurring series aimed at educating novice sports bettors. The next article will examine contrarian betting strategies.
If you would like to make or read comments about this article, you may do so by visiting the Mess Hall forum at MajorWager where a thread has been started. Please click HERE