There’s a lot of noise regarding QB production. Not a lot of stats get down to the heart of what matters for helping you to judge if a players is a worthy starter for your fantasy team. There’s probably also more variation in opinions on QB than about any position. Take for example Blake Bortles. He finished third in fantasy points among QB’s, and a healthy portion of the fantasy community will tell you he stinks. I’m not a Bortles apologist, but I don’t think any such argument exists about whether Brandon Marshall or even Doug Martin are #goodatfootball. So let’s cut through some of the noise by discarding these 3 stats/ ideas that really don’t tell you much about the QB in question.
Completion Percentage
It seems intuitive that this would be a very good thing. But when we look at the fantasy landscape from top to bottom there is a very narrow range of completion percentages. Every starting QB in the NFL is completing over 50% of their passes. With such a narrow range of outcomes we can’t glean a lot of information about fantasy value.
Here we’ve got the top 5 QB’s in terms of points and then QB’s 26-30 with their completion percentages. *Yellow values have been adjusted to reflect what that player could have been expected to score had they played a full 16 games. This was done so as not to skew statistical correlation with injury.
The worst 3 numbers here belong to the QB1 overall on the year Cam Newton, Blake Bortles, and Andrew Luck. The connection between those 3 are deep ball attempts. In 2015 Bortles was #1 (6.9 dba’s per game), Cam was #4 (5.1 per game) and Luck threw 6.3 deep balls per game(player profiler.com). The Deep Ball Attempts stat was one of the most highly correlated to fantasy points. Clearly, taking those kinds of shots down field is going to simultaneously rack up the fantasy points and destroy your completion percentage. The best deep ball in the NFL in 2015 was completed only 51.7% of the time (by Russell Wilson)
Taking another look at the chart the bottom 5 in fantasy points on average had better completion percentages than the top 5. Completion percentage in this case has a lot more to do with scheme than the accuracy, or talent, of said quarterback. It’s a reflection of the types of passes they’re completing, and those passes aren’t particularly valuable to fantasy owners. The system Teddy Bridgewater plays in is one of the most risk averse, run heavy, systems in existence. While he may run that offense very efficiently, his deep ball completion % is only 37.8 and his air yards per attempt were only 3.165. He’s dreadful in both categories, so while he has a high rate of completed attempts they just aren’t worth much.
At any rate, these completion numbers aren’t telling us a lot about fantasy value. Also, even if this was a good measure of accuracy this situation wouldn’t be allowed to play out. Serious steps are taken to keep these numbers tight. No QB is going to be allowed to stay in a game and complete 30% of their passes. If that was even close to happening either another QB would enter, the scheme would change, or both. Scheme can do a lot to hide inaccuracy.
Here we have the best five in terms of Comp %, and the worst 5 in the top 30
While the top 5 doesn’t look so bad, in the bottom 5 we find the #1,3, and 5 QB’s in terms of fantasy points on the season. You’re missing out on a lot if you’re looking through this lense to pick a QB. I’m not saying it doesn’t matter to being a good real life quarterback, but for fantasy purposes I’m looking for guys who throw down-field.
Interceptions
Interceptions are the cost of doing business in an increasingly pass happy NFL. A strong (0.67) correlation exists between attempts and interceptions across the top 30 QB’s. Take for example, Exhibit A: The most intercepted QB’s
This is basically the 600 attempt club. The only players in the club that aren’t pictured here are Brees, Eli Manning, and Rivers, which is a feather in their hat. However, none of us would make the argument that the best fantasy quarterbacks are those with the lowest attempts. If you’re throwing it, sooner or later you’re throwing it to the wrong team. Second of all, interceptions are highly correlated with deep ball attempts. Again, this is a behavior we want. There are risks associated, but few stats have a closer connection to fantasy points scored than the number of shots a QB is taking down field.
I think it’s worth mentioning that I understand having a QB that turns the ball over is bad for the offense. Clearly there comes a point where this behavior is a real problem, especially if those picks are coming in the red-zone (looking at you Matt Ryan). Interceptions have a moderately negative correlation to attempts while leading, and a QB with a attempts while leading is more valuable. But, the argument that we would want a QB who doesn’t turn the ball over, and therefore probably also doesn’t throw it much leads us to Teddy Bridgewater. Anytime an argument leads you thinking the Vikings pass attack is good, it’s time to go back to square one.
Here are your least intercepted passers:
Mariota, Taylor, and Smith all play in fairly anemic passing offenses, and Tyrod is the only one even close to the top 12, QB1 range in points. If you’re drafting Brady or Wilson it’s not because they don’t turn it over. I’m really impressed with Brady’s nearly 90 attempts per INT. I think that’s a product of an offense where he’s throwing a lot of timing routes. It’s the sort of situation where knowing what the throw will be could be confusing, especially with all the option routes, but the throw itself isn’t all that tough. He clearly knows what he’s doing in that system.
Attempts While Trailing, The Pursuit of Garbage Time
It’s not that this is a stat that gets cited frequently, but it’s a well beaten path. “They’ll be facing a negative game script, and throwing the ball a lot” goes pretty much unquestioned. To be fair, I think it works for Wide Receivers. But to quote JJ Zachariason “Getting garbage time takes a garbage performance”. For the signal caller, this just isn’t the situation you want. I know, I know, Bortles. Here are the top 5 in scoring with their attempts while trailing (according to pro-football reference) and the last 5 in the top 30 for comparison.
I think this paints a pretty clear picture. It might actually over-sell the very weak but slightly negative relationship to fantasy points. In the top 5, 3 had less than 200 attempts while trailing. The exceptions were Bortles and Luck (Luck’s numbers were extrapolated out to 16 games). On the other hand in the bottom set, 3 of 5 were on pace to top 300 attempts while trailing.
This should make us reconsider our ideas about negative game scripts, at least for the QB. It also makes me think about my evaluation of Bortles. He still has messy footwork, which leads to inaccurate throws, but if we’re going to attribute his production to situation and game script, then we have to answer what it is that makes Bortles productive under those conditions when it doesn’t have that effect for virtually anyone else. Phillip Rivers and Ryan Tannehill saw a very similar volume and attempts while trailing. They finished 13 and 23rd in fantasy points. My opinion is that the two things that make Bortles so different are named Hurns and Robinson. Add to that a coach who’s willing to throw it in the red-zone, and an attractive fantasy QB you make.
Here are your top and bottom 5 (in the top 30) in terms of attempts while trailing:
The 5 QB’s here who are taking the fewest attempts while trailing are outstanding fantasy QB’s. This does not require elaboration. It does, in my opinion, put this to bed.
As an alternative, attempts while leading is a much, much better predictor of fantasy points:
I think what we’re really measuring here is “on a good passing offense”. Clearly that’s a desirable trait. Another conclusion I’m drawing is that attempts while trailing may serve as a sort of “futility” measure, while attempts while leading is the opposite. Effective offenses create valuable players. Effective offenses spend more time in the lead.
The “So What”
The lense you use to evaluate a player is very important. Don’t shy away from players, especially in a week to week streaming situation, who take chances and might turn it over more. The very safe dinks and dunks only create value if they reach Patriot level death by a thousand cuts. Ryan Fitzpatrick, Tyrod Taylor, and Jameis Winston have all shown they are not afraid to air it out, and take those valuable shots down-field.
Another somewhat predictable thing to look for is how an offense approaches their red zone attempts. While Brady may not be throwing deep as often as Cam Newton, he is throwing it to Gronk and Edelman in the end-zone. The Jaguars are another team who like to throw it across the goal line. They may have gone out and gotten Ivory, but I don’t think you put receivers like the pair they have on a shelf once you’re inside the 20. The threat of being able to run may be all they really need from Ivory, although he has proven to be quite capable in short yardage. Flacco, Cousins, and Cutler get a healthy share of red zone attempts, and may be unowned in your league.
Last but not least, I’ll take a QB with a good game script over the hope of a 4th quarter flurry every time. These guys usually have a good defense supporting them. On the cheap you can find Alex Smith, Brock Osweiler, and Andy Dalton. I really like Osweiler in Week 1 if you’re streaming.
When you’re looking for the right things, what you find may be opportunity. Thanks for reading. You can find me on Twitter @jccrocker.
