Those of you who know me know that I loate the quarterback rating statistic. I don't believe it really tells you much about a quarterback. I don't think rating has any bearing on how good a quarterback actually did or how good they actually are. I don't think it translates at all to wins or losses. For the life of me, I can't remember when it became such as commonly used statistic, but it's been propped up the last couple of years when people talk quarterbacks.
Do you even know how to calculate quarterback rating without a script or a program? Well, here's how you do it:
Divide a quarterback's completed passes by pass attempts. Subtract 0.3. Divide by 0.2 and record the total. The sum cannot be greater than 2.375 or less than zero. Divide passing yards by pass attempts. Subtract 3. Divide by 4 and record the total. The sum cannot be greater than 2.375 or less than zero. Divide touchdown passes by pass attempts. Divide by 0.05 and record the total. The sum cannot be greater than 2.375 or less than zero. Divide interceptions by pass attempts. Subtract that number from 0.095. Divide that product by 0.04 and record the total. The sum cannot be greater than 2.375 or less than zero. Add the four totals you recorded. Multiply that total by 100. Divide by 6. The final number is your quarterback rating.
Yep. That easy.
Part of the reason I hate quarterback rating is because it rewards carefulness above all else. While it's great for a quarterback to not turn the ball over, it's not necessarily great if they're always looking for the safe pass. It's hard to make game changing or big plays always looking for what's safe.
- Quarterback A throws for 15/21 for 167 yards and a touchdown. His rating is 110.6.
- Quarterback B throws for 22/35 for 280 yards and one touchdown. His rating is 97.3.
That's a 13 point difference in rating, but who had the better game?
Part of the problem with quarterback rating is the weight put on certain things. If a quarterback doesn't throw a lot of passes, the likelihood of their rating to be higher is great. The more passes you throw, the more your quarterback rating can be damaged and the lower your completion percentage likely will be. For instance, if Quarterback A threw one interception, his rating would plummet to 86.8. That's a 20 point dump off for a single mistake, that could have happened early and had no effect on the game whatsoever.
The Big Lead did a write up on why quarterback rating stinks during the season. Jason Lisk argues that it's outdated, relies to heavily on completion percentage, and should include sack rate, which I agree with. His article focuses on David Garrard.
The issue I have occasionally mentioned, which is that sack rate should be part of how we evaluate a passer. Garrard has taken 17 sacks this year, and has been sacked on 8.6% of his dropbacks, which ranks him 34th out of 37 qualified passers. If he had managed to throw 6 of those passes away, his completion percentage would have dropped over 2%, affecting his rating, but helping his team.
To look at other quarterbacks who have had a better passer rating because of increased completion percentage (which is overweighted in the current formula) and because sack rate is not included, I found all quarterbacks since 1970 who threw 300 passes and were above the league average in passer rating. I then used the advanced statistics at pro-football-reference.com, and found everyone that had a difference of 10 points in their Rating+ score and their ANYA+ score (which does not include comp%, but does include sacks). I also looked at the points scored by those teams, and compared whether the points scored were more in line with their Passer Rating, or their ANYA.
As an example on his point about throwing a pass away instead of taking a sack, here is the difference in numbers:
- 236/366 (64.5%) - 2,734 - 23TD - 15INT - 90.8
- 229/366 (62.5%) - 2,734 - 23TD - 15INT - 89.2
It drops his rating only 1.6 points, but that's only accounting for 6 total throw aways. If you review each game and determine where he takes a sack when the ball could have been thrown away, it would drop it more. But, the fact that taking 6 sacks as opposed to throwing the ball away and avoid losing yards is better for quarterback rating seems self-defeating. As an extreme example, Garrard was sacked 33 times in 14 games. Let's pretend half of those sacks could have prevented with throw aways. So, he's sacked 17 times and has 16 more incompletions. This would dump his rating to 87.2, roughly 3 whole points.
If he'd thrown just one more interception? It drops by only 0.8.
So, why should 16 incompletions hurt a quarterback's rating 3 times more than a turnover?
I agree with his point on sacks should be incorporated, but then you've got to differentiate between sacks taken and sacks given up by the offensive line as a whole. Or, we could just you know... blow up QB rating, heh.
Is there a way to fix an already convoluted formula? Not without making it more convoluted. Why is quarterback rating held in such high regard in today's game, anyway?