Quick Study: Do NFL Preseason Wins Mean Anything?

posted in NFL

The four preseason games each NFL team contests every season are always a bizarre dance of competitiveness. On the one hand, the league does keep score and the players on the field are usually going all out.  On the other hand, a team’s key players may only be on the field for a half, a series, or (if you’re the Vikings’ Adrian Peterson) not at all.

Despite the lack of playing time for starters, though, one can build a logical case that a team’s preseason results still matter. A team with a good preseason record often had strong play from youngsters and backups, and those players will become more valuable as injuries mount and experience builds during the regular season. In addition, other facets of game play, such as coaching strategy or special teams, more closely emulate the regular season during these warm-up games, which could lend more meaning to preseason results.

Then again, there are also numerous examples of times when the preseason clearly serves as a poor proxy for future potential. Tom Brady wasn’t in the building during the Patriots football game this week; unless Brady goes down injured this year, it’s hard to draw any conclusions from how the Pats’ pass-heavy offense will fare based on that game. More famously in recent times, the 2008 Lions finished the preseason 4-0, then lost all 16 games they played in the regular season. Ouch.

Even though preseason pundits love to recall the 2008 Lions, one case like that doesn’t categorically discredit any relevance of the NFL preseason. So what’s the real answer?  We looked at data from the past seven seasons and answered three burning questions about what preseason success means.

Do Preseason Wins Correlate With This Year’s Regular Season Wins?

In short — NO. Our data shows no significant correlation between wins in the preseason and the regular season.  Your team went 4-0?  Too bad, it won’t help.  They started 0-4?  Don’t worry, it’s wiped clean. This doesn’t mean you can’t get excited about a rookie playing well, because that still can mean something. Just don’t put any stock into the wins and losses when you’re assessing your season win totals bets or Week 1 picks.

Do Preseason Wins Correlate With Next Year’s Regular Season Wins?

Play it again, Sam — NO. We investigated this question because we were curious if the promising young talent that evidences itself in helping win preseason games also bodes well for a team’s success a little farther in the future. Again, however, the answer is that there is no correlation. Young talent in the preseason may mean a bright future, but if so, it’s a more distant one.

Do Preseason Wins Correlate With Next Year’s Preseason Wins?

Funny enough…YES!  It’s a weak correlation but it’s still significant, and gives some insight into what preseason wins indicate. Preseason wins don’t mean a team is great, and they don’t mean a team will get better. They just mean that one side tends to put more effort into the win. The Colts and Chiefs haven’t averaged only one preseason win since 2005 because they’re bad; they’ve done so because of how they treat the preseason.  So if you’re team’s doing poorly in the preseason that’s unfortunate, but only if you’re a big fan of preseason football.

And before you say, “Duh, already could have guessed that,” just keep in mind that a startling amount of conventional wisdom about sports is easily proved wrong by the numbers, once a skilled analyst (and not your run of the mill TV sportscaster) digs into the data. In this case, the analysis confirms what most NFL fans probably think. There’s no signifiant correlation between a team’s preseason and regular season wins.



  • evo34

    So do these findings (from 10 years ago) not hold up on more recent data?


  • Jonathan

    Great counter point.

  • http://www.teamrankings.com/ David Hess

    Interesting find! I asked Austin to take a look at his data, and hopefully he’ll have time to reply soon.

  • Austin Link

    Those are cool articles. It looks like we did a similar analysis to them, but just by chance our data sets do not overlap. They end with 2005, we started with 2006.
    Part of the non-relation of preseason and regular season wins may be due to Peyton Manning. His teams in our set averaged less than 1 preseason win but more than 12 regular season wins.
    Even after removing his teams there is still no significant relation between the preseason and regular season. We can’t guarantee that an effect doesn’t exist, but if it does it’s very small, and has gotten smaller since Two Minute Warning did their research.

  • AlaskaLady

    I read a correlation study where they broke all teams into three groups, those teams that won 10 or more games the previous year, those teams who won 7-9 games in the previous year, and those teams who won six or fewer games in the previous year. In this study, those teams who fit into the middle category (7-9 wins) did have some correlation between preseason wins and regular season wins. Most of these teams that win 3 or more preseason games improved their regular season record over the previous season with 44% improving to the point of winning 10 or more games. Most of the teams in this category who won 2 or fewer preseason games did not improve in the regular season. I would suggest that further evaluation of
    numbers tells might suggest that it would likely be the same with prior season good
    teams and prior season poor teams if you could somehow factor out the
    “nowhere to go but down” or “nowhere to go but up” issue. The fact that
    you can only lose or win a limited number of games makes it difficult
    for already good teams to improve or already poor teams to do worse.

    Another issue that could skew the results are those teams whose first string do not play to their ability during preseason in order to avoid injury.

  • AlaskaLady

    I just now checked out your link and it is the same one that I read and cited in my response below!

  • http://www.teamrankings.com/ David Hess

    For those who missed it (not you, AlaskaLady, I see you saw it above), Austin replied to a question about this article above.