One of the more interesting aspects to NFL games is that it can often seem when you look at the box score afterwards that the wrong team won. This can happen to some extent in other sports - in baseball for instance one team may out-hit another by a considerable margin but fail to score runs and lose to a club that has fewer opportunities but gets the clutch hits. In hockey a team occasionally has a striking advantage in shots on goal but loses because too few find the net. With basketball it’s rarer still, but you’ll come across a team with a much higher field goal percentage losing.

In football though, it is commonplace to see a team out-gained in yards by a significant amount and still manage to win. Likewise a team may have an edge in first downs, turnovers, time of possession or any number of key statistics but still fall to the “statistically inferior” side.

As handicappers, these types of games can be frustrating if you are on the “right side” only to see your play lose despite what would seem to be stats good enough to get the job done. More important to us though is not the day after second-guessing, but what effects, if any, such a game may have on a team’s chances the following week.

To answer this question we’ll head back to the databases and do some research. We’ll track not only teams that “should have won” according to the stats but didn’t, but also those teams that “should have lost” but somehow, some way, managed to escape with a victory.

The following data represents the past ten years of NFL action, for games played in week three and beyond.

The “Wrong Team Won” column represents the next game results against the spread for the team that according to the statistic in question should have lost but was actually the winner. The “Wrong Team Lost” column represents the next game results for the team that should have won according to the statistic, but was actually the loser of the game.

In general there do not seem to be too many pronounced aftershocks for a team playing in one of these games, whether it’s the lucky or unlucky squad involved. There are however some exceptions and indeed if we were to break it out in more detail (maybe use only the second half of the season, or use more extreme differences) we can find some places of historical predictive value for the next game.

The most promising area in the above table is with teams that had a large time of possession advantage (58%+ or 34:48 of a non-overtime event) but lost have been lousy in their next appearance -- playing against such a team has been a 67% win history. Another time to consider stepping up against a “wrong team lost” side is when the team lost despite an 8+ first down advantage - these are 61% as go-against plays. A last warning sign for a team that “should have won” is when they outgained their opponent by 150+ yards passing but lost. This one is easier to understand since a losing team frequently has more passing yards in its desperate attempt to get back in a game by abandoning the run. These teams are 61% go against next up as well.

Surprisingly though it’s hard to find much evidence that winning a game that by statistical measures should have been a loss leads to trouble next out. In fact, it’s almost the opposite. With turnovers, a team that won despite a 2+ deficit (committing two more turnovers than their foes) is a strong bet next out as an away dog: 28-12 for 70%.

Our conclusion on all this is that the predictive value of a “messed up box score” is not all that great, more a reflection of the nature of the NFL where key plays can mean so much and have such a dramatic effect on the scoreboard. The “wrong team winning” is seldom a dire clarion for the upcoming game. The best avenue for locating a few good bets is to pour over the boxes and look for the teams that can’t get it done even when key stats (time of possession, turnovers, passing yards) are in their favor…and then take a stand against them!