Newsletter

Does Pre-Season Matter?

Football fanatics wait anxiously for the pre-season every year and fantasy football players often watch pre-season games in order to find a sleeper for their fantasy football league. However, there seems to be very little evidence that players who surprise in the pre-season will continue to impress in the regular season. Some of the top players in yards per carry average last pre-season were Ray Rice, Danny Ware, Darren McFadden and DeAngelo Williams while the leaders in receptions were DeSean Jackson, Dominique Zeigler, Dwayne Jarrett, Dominique Thompson and Kevin Walter. Even though a few players perform well in both the pre-season and regular season, it doesn’t seem like player statistics have much meaning since it really depends on the playing time that each player will get. However, the one thing that is consistent between pre-season and regular season is the coaching staff and the system that they are using. For that reason, team statistics should have much more meaning than individual player statistics. In the following article I will look at pre-season team statistics to try and determine which ones can help improve player projections for the upcoming season.

The two statistics that typically interest fantasy football players are yards and touchdowns. Looking simply at those two statistics to try and predict regular season statistics would however not be the best method. Yards are dependent on both the number of attempts and the performance on each attempt. During pre-season, one thing that is consistent is the coaching staff so a team that runs the ball more during pre-season should also run the ball more in the regular season. I therefore thought that by separately predicting attempts and yards per attempt I would obtain more accurate results. Here is the list of pre-season statistics I looked at to try and predict regular season statistics: % of rushing plays, plays per game, yards per carry, yards per pass attempt, % of rushing touchdowns and touchdowns per game. The following table shows the formula that best predicts each regular season statistics using data from the past four seasons:

Those formulas show that if, for example, a team averaged 55 offensive plays per game in the pre-season, you can expect them to average 51.9 + 0.140 x 55 = 59.6 offensive plays per game in the regular season. The coefficient by which the pre-season statistic is multiplied gives an indication of how much meaning each statistic has. As initially expected, the percentage of rushing plays has the most meaning since a coach will usually use the same system during the pre-season and the regular season. Those numbers mean very little without a point of comparison so I did the same thing but comparing regular season statistics with the previous year’s regular season statistics:

As you can see the coefficients by which the previous year’s statistics are multiplied are all higher than for the pre-season statistics but they are quite close for the percentage of rushing plays. This means that none of the pre-season statistics have more meaning than the previous year’s statistic. However, looking at all the data, I noticed that very often the team’s statistics for the upcoming season were between the pre-season statistics and the previous year’s regular season statistics. For example, Seattle ran the ball 42% of the time in 2007, 52% of the time in the 2008 pre-season and 47% of the time in the 2008 regular season. This led me to believe that by combining both the pre-season statistics and the previous year’s regular season statistics I could improve the projections for the upcoming season.