By Kurt Bullard
Every year, there’s a player in fantasy football that someone just stumbled upon who came out of nowhere to be not only a good starter, but a great one. This year, it’s been Tyrod Taylor, who threw 35 passes in his four-year stint as backup to a ELITE Quarterback before bursting onto the scene this year as a top-five quarterback in most scoring formats. As someone who has started Taylor every game of the year, I’m thankful that he fell into my lap, while at the same time constantly waiting for the shoe to drop and to have Tyrod Taylor start playing like, well, a career backup.
Are fantasy stars who came from the depths of the league less likely to be able to maintain a high level of performance, or is a quarter of the year enough of a sample to trust that this player will continue his production? To examine this question, I took quarterbacks who ranked in the top 15 in fantasy points after the first four weeks of the 2010-2014 seasons. I then calculated their weekly average in the first quarter of the season compared to their total for the last three quarters of the season. I then calculated the difference between the two periods of the year to use as my measurement of whether a quarterback dropped in performance.
I then added a dummy variable to these 75 quarterbacks to see if they had finished in the top 15 in fantasy points the season before. This way, I could see quarterbacks who were performing much better than they had the season before. There were a total of 27 quarterbacks who were Top 15 quarterbacks through the first four weeks of these seasons who hadn’t finished as a Top 15 quarterback the year before.
For these quarterbacks, I performed a Sign Test to see whether it was more likely than not that the performance would drop off after the first four weeks of the season. Only five quarterbacks managed to outperform their first quarter totals, while 22 fell short. Using a binomial distribution, one can find that the Test-Stat for the Sign Test is -3.27, yielding a p-value of .001, a significant value. Nominally, it looks like quarterbacks who are new to being startable tend to regress back to their mediocre ways.
However, this doesn’t tell the whole story. It could just be that all quarterbacks – perennial fantasy studs or not – regress back to the mean, since we are just looking at the top 15 quarterbacks. While quarterbacks who didn’t place in the top 15 saw their production drop off 2.87 points per game for the rest of the season, those who had been top 15 quarterbacks still saw their production drop off, just at a slightly less tune of 1.47 points per game. There’s a trend that these QBs who play well through four weeks tend to fall back to earth.
Using a Rank Sum Test since the normality assumption for a T-test may or may not be met, we find that the p-value for whether the median ranks are different for the two types of quarterbacks is .33, not close to significant. It appears that newcomers to the Top 15 are not more likely to leave it than those who have been there before.
So for those of you like me who are leaning on Tyrod Taylor to help lead you to the fantasy playoffs, fear not – while he and other proven quarterbacks may take a step back, there’s hope that he could still keep this viable starter thing up.