An interesting and albeit natural debate among Gator fans has risen regarding the QB situation for Florida. Kyle Trask gets a (deserving) amount of praise for his performance this year stepping in for an injured Feleipe Franks, who is (undeservedly) widely scorned. It seems to me that they are being graded in the court of public opinion on different curves. Don’t get me wrong- Kyle has been great and Franks certainly had some of those head-scratching moments. However, like most things sports, rationality isn’t the soup du jour.
As such, I decided to take a look at both QB’s performances across 8 games (since that is how many Trask has started now). I didn’t include the Kentucky game that was split between the two and just looked at the last 8 starts for Franks and did a side-by-side comparison. I looked at their performances along a few dimensions: Passer rating, completion percentage, yards, touchdowns, interceptions, longest pass plays. The results were interesting. The point here is not to determine who is ‘better’, just to look at the most recent comparable sample size and see if there is any difference in consistency as a passer (in terms of statistical variance). To determine who is better would require a much deeper dive, which I am not interested in- I’m glad they’re both Gators.
The Stat Lines
The table below shows each player’s overall performance for the 8-game sample:
This is a metric taken from ESPN and isn’t the overall QB rating. I’m not sure how ESPN calculates this, but since I’m only analyzing these two as passers, I decided to include this statistic.
This box and whisker plot shows that Trask is much lower variance than Franks. If you don’t already know, a box and whisker plot shows quite a few things. The top edge and bottom edge of the box are the hinges. These represent the 1st quartile and 3rd quartile. The area between them is the interquartile range (IQR). The line through the middle of the box is the median. The ‘X’ in roughly the middle is the arithmetic means (average). The whiskers represent the range of the data up to what is sometimes referred to as the upper and lower fences. These are the horizontal lines at the end of the whiskers. Any score outside of these fences is considered an outlier. From the graph above, we can see Trask has an outlier score (passer rating of 202 vs Towson).
This graph shows the same data but has some different information. This shows how much difference it can make to include outliers in analyses. The arithmetic mean, or average, can be highly susceptible to extreme scores, especially in smaller sample sizes. When comparing Trask and Franks, it may not be best practice to do so by average passer rating because of the impact the outlier has on Trask’s score. The median may be a better statistic to use here, and they are virtually dead-even there (150.65 for Trask, 151.55 for Franks). Nonetheless, including Trask’s outlier, he is still much lower variance than Franks.
Outliers: Trask vs Towson (90%), Franks vs UT Martin (92.6%)
The outlier is Franks’ 2-interception game vs Miami.
Longest Pass of Game
My two cents
Trask is playing better. Not because he is throwing for more yards and more touchdowns, which is boosted by the fact that he is throwing more (32 attempts per game to Franks’ 24), but he is throwing more passes and is lower variance in every category except interceptions. He is the more consistent passer, and he is consistently good. Caveat- Running stats are not included and should certainly be considered, as Franks had 8 touchdown runs over that span and a median long run of 14.5 yards per game. Yes, he had a better line for 6 out of those 8 games, but it is what it is. He was a threat in the run game. I think both of them are good QBs, but the degree of separation is ever so slight. Performance-wise, I appreciate consistency and Trask has definitely been that while still being highly effective.
Here is how FF 1st 8 games under Mullen would look when plotted against the above sample:
FF first 8 games under Mullen vs Trask: