Do Thursday Night Games Negatively Impact Fantasy Football Performance?
The NFL expanding its offering to include Thursday night football games has done nothing but make NFL fans, bar owners, and fantasy football players happy - who doesn't want to watch football during a new day of the week?
However, it doesn't take an avid film-watcher to notice that the quality of play for these games sometimes seems to be a little bit off, most likely due to little preparation, practice time and compressed game plans. What you tend to get are sloppy, turnover-ridden games with low scores.
This got me thinking about whether or not there was a decline in production for fantasy football players playing on Thursday.
With that in mind, I decided to embark on a data gathering process to include fantasy football points scored from 2011 to 2013 for all Thursday games played. My data compilation utilized historical points scored numbers from Fantasy Football Toolbox for the quarterback, running back, top two wide receiver and tight end positions.
I collected standard-scoring fantasy points from both Thursday games played as well as calculated the average points for all games a player played outside of Thursday games. After collecting that data, analysis and testing was performed with some help from friend and Miami University Adjunct Professor Matt Booher. The goal, again, was to see if there truly is a fantasy points decline on Thursday vs. the traditional Sunday and Monday night games where there is typically a full week of preparation and practice.
The following assumptions were used in compiling and analyzing the data:
1. The initial Thursday night game that opens each season is not part of the analysis. The thought process is that there isn't a short week for the teams involved, as teams could, in theory, have months to prepare. Additionally, there'd be no short period of time to recuperate from a full-speed game the Sunday prior.
2. The initial Thursday night game would be part of the season averages outside of the Thursday game that a player played.
3. The season averages for players excludes the Thursday night game being analyzed and incorporates only the number of games that particular player played. So if a player missed time for an injury, the average was only taken for games played. If a quarterback played a few snaps and then was injured during a game, that game was not included as part of the average either to be conservative for seasonal fantasy points average.
4. If a major player was out due to injury on a Thursday, the statistics of his replacement were used against the replacement's season statistics.
5. Where practicable, the players used were the top fantasy point scorers at their position for their respective NFL team for the season. If a star player got hurt outside of their Thursday night game and it was known that they would've been the highest scorer at their position (think a player like Percy Harvin in 2012 who got hurt the game after his Thursday night game), the injured player was the player that was used.
6. If a player is no longer in the NFL, their fantasy points were not listed or available for their game in 2011 through 2013 and could not be used as a data point.
The Null Hypothesis: There is no statistical difference between Thursday games and Sunday and Monday games for each position.
|Overall 3-Year Positional Analysis||Thurs. AVG||Thurs. STDEV||Seasonal AVG||Seasonal STDEV||Difference||P-Value|
|TE||5.27||4.07||5.82||2.3||-0.55||0.265||2013 Positional Analysis||Thurs. AVG||Thurs. STDEV||Seasonal AVG||Seasonal STDEV||Difference||P-Value|
|TE||6.56||4.82||6.4||2.51||0.16||0.86||2012 Positional Analysis||Thurs. AVG||Thurs. STDEV||Seasonal AVG||Seasonal STDEV||Difference||P-Value|
|TE||5.09||3.32||5.45||1.96||-0.36||0.61||2011 Positional Analysis||Thurs. AVG||Thurs. STDEV||Seasonal AVG||Seasonal STDEV||Difference||P-Value|
What does the above data exactly tell us? First of all, it says that, based on the p-value being less than .01,we can reject the null hypothesis regarding the quarterback position - there is a statistical difference between Thursday night games and other contests. As a result, we can accept that quarterbacks as a group perform worse on Thursday games compared to their yearly average by a total of 1.77 fantasy points over the 2011-2013 period. (For a more detailed explanation on p-values, Wikipedia has you covered)
Would this alone be enough for you to move the needle and not start a particular quarterback you own? Possibly not. But if all things were equal, you should consider it, especially if that quarterback has an unfavorable matchup.
What About the Other Positions?
With the other positions, the p-values are high enough to validate the null hypothesis in that there is no statistical difference between Thursday games and games played on other days or nights.
It should be noted, however, that running back play actually has improved on Thursday games relative to Sunday and Monday night games by approximately one fantasy point. As running plays are simpler in nature to block for, this makes sense. There have also been some monster running back performances on Thursday nights in recent memory, including Doug Martin torching the Minnesota Vikings in 2012 for 33.4 fantasy points and DeMarco Murray getting his turkey trot going on Thanksgiving 2013 versus the Oakland Raiders for 28.2 fantasy points. This clearly brings up the Thursday average.
The top-ranked receiver on each team tends to score one less point on Thursday nights versus Sunday and Monday contests over the last three seasons. This variance probably wouldn't cause you any stress, as not many fantasy owners are benching guys like Brandon Marshall anyways. Coincidentally, Marshall was the only WR1 to score more than 20 standard fantasy points on a Thursday on short rest in 2013. However, the drop-off on Thursday nights combined with worse quarterback play can make it hard for a wide receiver to "go off" on Thursday, especially the top option who the defenses can more easily game plan around on a short week.
Does Familiarity Breed Productivity on Thursday?
Earlier in the preseason, I was watching a game on CBS in which Jim Nantz and Phil Simms were talking about there being more divisional games this year than in years past on Thursday nights (televised on CBS this year), and that this should bring better quality play, as it's easier to plan for a "familiar foe." After hearing that and deciding to do this analysis, I decided to take it a step further and test the null hypothesis again to see how Thursday games versus a division foe fared by position, again, comparing the results to Sunday and Monday games.
Here are the results:
|Divisional Analysis (All Seasons)||Thursday AVG||Thursday STDEV||Seasonal AVG||Seasonal STDEV||Differnce||P-Value|
|2013 Divisional Analysis||Thurs. AVG||Thurs. STDEV||Seasonal AVG||Seasonal STDEV||Differnce||P-Value|
|2012 Divisional Analysis||Thurs. AVG||Thurs. STDEV||Seasonal AVG||Seasonal STDEV||Differnce||P-Value|
|2011 Divisional Analysis||Thurs. AVG||Thurs. STDEV||Seasonal AVG||Seasonal STDEV||Differnce||P-Value|
Surprisingly, quarterbacks fared worse on Thursdays against their division than they did against the total pool, to the tune of 2.05 fantasy points against Sunday and Monday games. The quarterback p-value of .03 highlights a strong presumption to reject the null hypothesis, leading to the likely conclusion that there is a statistical and not random relationship to quarterback performance versus their division on Thursday nights.
This p-value is slightly worse than the overall one, which makes sense as there are more samples for all games versus division games only. However, most of the other positions didn't demonstrate a strong enough relationship to be relied upon to reject the null hypothesis, and you can say with certainty - based on a 0.91 p-value - that running back production is not dependent on day of the week the game is played. With that said, the CBS crew may not see better offense in the divisional games, as fantasy scoring was down across the board for them on a short week, even if the results were random.
Of course, there are other factors that go into all of this data for 41 Thursday games played from 2011 through 2013, including whether backups were playing, the quality of the defensive opponent, the weather, etc.
As a fantasy owner, you're always looking for an edge, so hopefully the next time you're contemplating starting a mediocre quarterback on a Thursday night, you'll think twice about it and remember the statistical analysis put forth here. For the other positions, you probably shouldn't lose any sleep, as three years of games didn't produce earth-shattering results, rejecting the null hypothesis.