The misconception of Jameis Winston’s accuracy

Football Outsiders (FBO) just released their Passing Plus-Minus numbers for the 2017 season. What is passing plus-minus, you ask? Bucs Nation is so glad you asked. FBO defines it as:

...passing plus-minus estimates how many passes a quarterback completed above or below what an average quarterback would have completed, given the location of those passes. It does not consider passes listed as “Thrown Away,” “Tipped at Line,” or “Quarterback Hit in Motion.” Player performance is compared to a historical base­line of how often a pass is completed based on the pass distance, the distance required for a first down, and whether the ball was thrown to the left, middle, or right side of the field. Note that plus-minus is not scaled to a player’s total attempts.

In other words, it is a modern substitution for completion percentage and accuracy because completion percentage has become a meaningless stat in the modern NFL. As Scott Kacsmar writes, the single-season record for completion percentage was reset four times in the last decade (three times by Drew Brees alone since 2009), being reset again in 2017 with Brees’ 72.01 percent. It’s not just Brees either - the league-wide average completion percentage record was broken every year from 2013-2016, with 2017 marking a narrow regression back down to 62.1 percent from 2016’s record 63 percent.

Okay, so we understand why Football Outsiders created passing plus-minus and what it is, but what does it mean? Kacsmar wrote this in 2015:

Then there is the significance of how far the pass was thrown. If a quarterback wanted to complete 75 percent of his passes in a season, he could definitely hit that mark, but it would likely come at the expense of a productive offense. Since 2006, aimed passes thrown to receivers behind the line of scrimmage have been completed 87.3 percent of the time. Aimed passes that travel exactly 10 yards beyond the line of scrimmage are completed 59.2 percent of the time. At 20 yards, that rate dips to 44.7 percent.

It is also worth noting which field direction the pass was thrown: left, middle, or right. A 4-yard pass to a slot receiver coming over the middle is actually still a shorter throw than a 4-yard pass thrown to a curling wide receiver on the outside. On average, the further the ball has to travel, the less likely it is to get there accurately.

Basically, they charted every quarterback’s pass and created a metric that can measure accuracy. No system is perfect, but plus-minus does a pretty good job of it and providing context for every throw.

That context is important because Kacsmar’s assertion that a quarterback and team could trade productive offense for consistency is exactly what has happened to much of the NFL. They have also explored the concept of failed completions, which is defined as any completed pass that isn’t successful, or one that fails to gain at least 45 percent of needed yardage on 1st down, 60 percent of needed yardage on 2nd down, and 100 percent of needed yardage on 3rd and 4th downs. In the 2016 season Jameis Winston finished with the league’s third-lowest failed completion rate per attempt at 11.3 percent. Last season, Winston finished with the NFL’s lowest failed completion rate of 11.4 percent. While he’s stayed pretty consistent in pushing the ball down the field and has kept Tampa Bay’s offense on schedule more than arguably any other quarterback in the league, we can see the rest of the league changing around him almost in real time as they throw shorter and shorter and become increasingly reliant on yards after the catch.

Okay, so Winston throws farther down the field and more often on average than most quarterbacks, which means his passes are generally going to be more difficult to complete than the passes his peers are attempting. But you can’t compare his completion percentage to Alex Smith’s, or Brees’, or even Marcus Mariota’s because it’s apples to oranges. Every NFL system is different and asks different things from its quarterback. But passing plus-minus’ historical baseline allows us to compare Winston’s passes to how an average player would do throwing the same passes. Yes, even his deep ball. It does this for every quarterback, for all throws included in the study, and then compares each quarterback.

2017

Winston’s passing plus-minus in 2017 was a positive 18.7, meaning he completed almost 19 more passes than an average player would have been expected to complete given what he attempted. That’s good enough to rank him fifth in the NFL. To express this as a rate stat, like completion percentage is, Winston completed passes at a rate 4.5 percent better than average, which is tied for third-best in the NFL. Kacsmar actually expanded on Winston’s 2017 season, saying:

It was a weird season, but Jameis Winston (+18.7) and Josh McCown (+16.8) ranking in the top six is up there with Keenum on the weirdness scale. It also shows some of the value of this stat. Winston was only 12th in conventional completion percentage (63.8 percent), but that doesn’t account for the fact that he is one of the most vertical passers in the game. His average depth of target in this study came at 10.7 yards, second only to [Deshaun] Watson (10.9).

That’s pretty good! For comparison, Watson ranked 17th in plus-minus at +2.9.

But what about tangible improvement?

When he’s not having one of his multiple-turnover games Winston is arguably one of the most underrated quarterbacks in the league. But so what, his haters say. That was just one season where the Bucs went 5-11 and missed the playoffs yet again. Winston is one of the most inaccurate quarterbacks in the league and has led the league in turnovers since 2015. What evidence is there his accuracy is improving or that Winston can be a franchise quarterback?

Kacsmar covers that too:

In fact, each year Winston’s passes get a little deeper, but he has done a great job of improving his plus-minus. He has gone from -14.5 as a 2015 rookie to +3.2 in 2016 to a similarly big increase in 2017.

In 2015 Winston’s -14.5 plus-minus ranked him 32nd. In 2016 he ranked 14th. And last season he ranked 5th.