2016 ALEX: Midseason Report
by Scott Kacsmar
Our first look at ALEX this season showed that Ben Roethlisberger and Aaron Rodgers were as aggressive as ever at attacking the sticks on third down, while Sam Bradford and Blaine Gabbert (since benched) were still coming up short in new offensive systems. We also saw three passive rookies rank in the bottom eight in Dak Prescott, Carson Wentz, and Cody Kessler (dead last). However, Prescott still ranked first in conversion rate, and Kessler was fourth for a struggling Cleveland offense.
How have things changed over the last month? Let this serve as our midseason ALEX report. For those new to this stat, ALEX stands for Air Less EXpected, and measures the average difference between how far a quarterback threw a pass (air yards) and how many yards he needed for a first down. If a quarterback throws a pass for four yards behind the line of scrimmage on third-and-14, then that would be minus-18 ALEX. A pass thrown beyond the first-down marker would earn a positive ALEX figure. Almost all quarterbacks are wise enough to attack the sticks on fourth down, which is why ALEX is best applied on third downs to gauge the aggressiveness of a quarterback. As you will see, we have identified some quarterbacks with a strong year-to-year tendency for aggressive or ultra-conservative play.
Football Outsiders now features ALEX data on the quarterback stats page, which can be seen at any time. As always, the numbers are subject to change once corrections are made in the NFL's play-by-play, as well as after the season when we remove certain incompletions such as passes thrown away or batted at the line. There may be some minor differences in the tables below compared to the above link due to these factors, as well as the exclusion of defensive pass interference plays. It's not that those penalty plays are not important examples of decision making, but our data on ALEX from 2006 to 2015 does not include penalties, so we are ignoring them in contemporary data so we can make valid historical comparisons.
The following table looks at every quarterback this season with at least 30 third-down passes through Week 10. The quarterbacks are ranked from highest to lowest ALEX. Also included are their conversion rate, DVOA, and average need yards for a first down. Finally, Short% is the percentage of third-down attempts that were thrown short of the sticks (negative ALEX). ALEX and Short% generally have strong correlation (near minus-0.85), but Short% is a good way to account for the outliers.
| Rk | Quarterback | Team | ALEX | CONV% | Rk | DVOA | Rk | Passes | Avg. Need | Rk | Short% | Rk |
| 1 | Ben Roethlisberger | PIT | 5.0 | 44.0% | 11 | 58.3% | 5 | 75 | 6.9 | 31 | 29.3% | 3 |
| 2 | Aaron Rodgers | GB | 4.9 | 47.8% | 4 | 30.9% | 15 | 92 | 7.1 | 27 | 20.7% | 1 |
| 3 | Tom Brady | NE | 4.1 | 66.7% | 1 | 185.7% | 1 | 42 | 7.8 | 17 | 31.0% | 4 |
| 4 | Tyrod Taylor | BUF | 3.5 | 42.5% | 13 | 52.7% | 8 | 73 | 8.1 | 9 | 32.9% | 7 |
| 5 | Eli Manning | NYG | 3.3 | 33.3% | 29 | -2.7% | 30 | 93 | 7.1 | 26 | 35.5% | 10 |
| 6 | Cam Newton | CAR | 2.4 | 40.7% | 16 | 17.8% | 19 | 81 | 8.0 | 14 | 27.2% | 2 |
| 7 | Blake Bortles | JAC | 1.9 | 36.8% | 22 | -26.5% | 32 | 95 | 7.1 | 30 | 31.6% | 5 |
| 8 | Jameis Winston | TB | 1.9 | 46.7% | 7 | 44.0% | 10 | 92 | 7.8 | 18 | 33.7% | 9 |
| 9 | Andrew Luck | IND | 1.8 | 42.1% | 15 | 3.4% | 28 | 76 | 8.0 | 12 | 36.8% | 13 |
| 10 | Drew Brees | NO | 1.7 | 53.8% | 2 | 75.6% | 3 | 91 | 6.6 | 34 | 45.1% | 23 |
| 11 | Alex Smith | KC | 1.7 | 36.4% | 23 | 13.2% | 21 | 66 | 7.3 | 23 | 39.4% | 16 |
| 12 | Marcus Mariota | TEN | 1.6 | 46.9% | 6 | 59.9% | 4 | 96 | 7.9 | 16 | 33.3% | 8 |
| 13 | Kirk Cousins | WAS | 1.4 | 42.2% | 14 | 9.2% | 24 | 90 | 7.3 | 25 | 35.6% | 11 |
| 14 | Jay Cutler | CHI | 1.4 | 33.3% | 30 | 9.3% | 23 | 33 | 7.3 | 24 | 36.4% | 12 |
| 15 | Brian Hoyer | CHI | 1.2 | 34.7% | 27 | 4.8% | 27 | 49 | 8.0 | 10 | 32.7% | 6 |
| 16 | Brock Osweiler | HOU | 1.2 | 37.4% | 20 | -9.2% | 31 | 91 | 7.1 | 29 | 42.9% | 19 |
| 17 | Joe Flacco | BAL | 1.0 | 34.6% | 28 | -1.6% | 29 | 107 | 8.3 | 8 | 46.7% | 25 |
| Rk | Quarterback | Team | ALEX | CONV% | Rk | DVOA | Rk | Passes | Avg. Need | Rk | Short% | Rk |
| 18 | Ryan Tannehill | MIA | 1.0 | 37.0% | 21 | 18.8% | 18 | 81 | 7.1 | 28 | 39.5% | 17 |
| 19 | Dak Prescott | DAL | 0.5 | 48.7% | 3 | 53.1% | 7 | 78 | 7.4 | 20 | 37.2% | 15 |
| 20 | Philip Rivers | SD | 0.5 | 47.4% | 5 | 53.5% | 6 | 95 | 6.7 | 33 | 43.2% | 20 |
| 21 | Ryan Fitzpatrick | NYJ | 0.4 | 32.9% | 31 | -31.3% | 33 | 82 | 7.3 | 22 | 50.0% | 31 |
| 22 | Matthew Stafford | DET | 0.4 | 45.7% | 9 | 41.1% | 12 | 81 | 8.0 | 13 | 43.2% | 21 |
| 23 | Colin Kaepernick | SF | 0.3 | 28.6% | 33 | 5.8% | 26 | 35 | 9.2 | 1 | 37.1% | 14 |
| 24 | Trevor Siemian | DEN | 0.2 | 39.5% | 17 | 21.7% | 17 | 76 | 8.0 | 11 | 47.4% | 26 |
| 25 | Russell Wilson | SEA | 0.2 | 38.5% | 19 | 35.6% | 14 | 78 | 8.3 | 6 | 42.3% | 18 |
| 26 | Case Keenum | LARM | -0.1 | 38.9% | 18 | 16.3% | 20 | 90 | 6.8 | 32 | 53.3% | 34 |
| 27 | Derek Carr | OAK | -0.1 | 36.1% | 24 | 41.8% | 11 | 97 | 8.6 | 5 | 47.4% | 27 |
| 28 | Carson Wentz | PHI | -0.2 | 35.0% | 26 | 6.7% | 25 | 80 | 8.6 | 4 | 48.8% | 29 |
| 29 | Matt Ryan | ATL | -0.2 | 42.5% | 12 | 76.4% | 2 | 80 | 8.3 | 7 | 46.3% | 24 |
| 30 | Sam Bradford | MIN | -0.4 | 44.3% | 10 | 36.5% | 13 | 79 | 7.4 | 21 | 50.6% | 32 |
| 31 | Carson Palmer | ARI | -0.4 | 45.8% | 8 | 52.2% | 9 | 83 | 8.8 | 3 | 44.6% | 22 |
| 32 | Andy Dalton | CIN | -0.6 | 32.6% | 32 | 28.7% | 16 | 86 | 8.9 | 2 | 52.3% | 33 |
| 33 | Blaine Gabbert | SF | -1.2 | 25.5% | 34 | -61.3% | 34 | 47 | 7.6 | 19 | 48.9% | 30 |
| 34 | Cody Kessler | CLE | -1.9 | 36.0% | 25 | 12.7% | 22 | 50 | 7.9 | 15 | 48.0% | 28 |
Roethlisberger and Rodgers remain in the top two spots, and Rodgers is still on pace for the second-lowest Short% season on record. Tom Brady, who was not listed in our first ALEX report for obvious reasons, is at third. That is unusually high for Brady, albeit on a sample of five games. Brady's ALEX has not been above plus-3.0 since the 2006 season, when he was plus-3.03. Brady's Short% would also be a new low for him, as he ranked 28th in Short% last season (45.3 percent). The difference in third-down DVOA between Brady and the No. 2 quarterback (Matt Ryan) is larger than the difference between Ryan and the No. 33 quarterback (Ryan Fitzpatrick). Not surprisingly, Brady's conversion rate is a blistering (but unsustainable) 66.7 percent this season. The highest season (minimum 50 passes) since 2006 was Peyton Manning's 2006, when he converted 59.6 of his third-down passes on an offense that broke the NFL record for third-down conversion rate (56.1 percent).
No play exemplifies Brady's 2016 success better than his third-and-25 conversion against Seattle last week. On a third-and-25, Brady dropped a perfect pass 28 yards beyond the line of scrimmage to Julian Edelman for a 33-yard gain (and plus-3.0 ALEX) to end the third quarter. Out of the 23 passes on third-and-25 or longer around the league this season, Brady's was one of two to have a positive ALEX. Josh McCown had the other (plus-1.0) on a third-and-28 pass to Terrelle Pryor against Baltimore. Brady's completion was just the fourth with positive ALEX on third-and-25 or longer on 297 such attempts since 2006. As if that weren't unusual enough, Brady targeted Edelman, who had just 15 targets on passes thrown 28-plus yards in his career coming into this season. Eight of those were first-and-10 plays, while none of the others had a need yards greater than 8. Seeing how Brady fares the rest of the season on third down will be one of the more interesting ALEX-related stories to follow.
Blaine Gabbert brings up the rear in conversion rate and DVOA, while Case Keenum is dead last in Short%. Both quarterbacks have been benched in the NFC West, which just makes you wonder why Colin Kaepernick and Jared Goff weren't starting in that 28-0 Week 1 Monday night debacle in the first place.
Cody Kessler is still last in ALEX, and his high conversion rate was expected to regress, but his decline has been even more dramatic than we would have expected, falling from fourth in conversion rate a month ago to 25th. In fact, Kessler's decline in conversion rate is only matched by that of his opponent this week, Roethlisberger. The following table shows a comparison of ALEX and conversion rate for third down between Weeks 1 to 5 and Weeks 6 to 10. Only quarterbacks with at least 20 pass attempts in each split are included. The table is sorted by descending ALEX difference.
| 2016 ALEX: Weeks 1-5 vs. Weeks 6-10 (Min. 20 Passes Per Split) | |||||||||||||
| Quarterback | Team | ALEX 1-5 | Rk | ALEX 6-10 | Rk | DIFF | Rk | CONV% 1-5 | Rk | CONV% 6-10 | Rk | DIFF | Rk |
| Andy Dalton | CIN | -2.1 | 30 | 1.4 | 15 | 3.5 | 1 | 26.0% | 30 | 41.7% | 15 | 15.7% | 3 |
| Matthew Stafford | DET | -0.8 | 26 | 2.1 | 10 | 2.9 | 2 | 48.9% | 3 | 41.7% | 14 | -7.2% | 22 |
| Cody Kessler | CLE | -3.3 | 31 | -0.4 | 24 | 2.9 | 3 | 48.0% | 4 | 24.0% | 29 | -24.0% | 28 |
| Alex Smith | KC | 0.5 | 16 | 3.3 | 3 | 2.8 | 4 | 37.5% | 19 | 34.6% | 23 | -2.9% | 15 |
| Ryan Tannehill | MIA | -0.4 | 23 | 2.4 | 7 | 2.8 | 5 | 31.7% | 28 | 42.5% | 12 | 10.8% | 8 |
| Eli Manning | NYG | 1.9 | 7 | 4.7 | 2 | 2.8 | 6 | 37.5% | 18 | 28.9% | 25 | -8.6% | 24 |
| Marcus Mariota | TEN | 0.2 | 21 | 2.9 | 4 | 2.7 | 7 | 37.0% | 21 | 56.0% | 2 | 19.0% | 2 |
| Dak Prescott | DAL | -0.5 | 24 | 1.9 | 11 | 2.4 | 8 | 52.3% | 1 | 44.1% | 11 | -8.2% | 23 |
| Blake Bortles | JAC | 1.1 | 11 | 2.5 | 6 | 1.4 | 9 | 26.2% | 29 | 45.3% | 8 | 19.1% | 1 |
| Derek Carr | OAK | -0.7 | 25 | 0.6 | 19 | 1.3 | 10 | 36.0% | 24 | 36.2% | 20 | 0.2% | 14 |
| Drew Brees | NO | 1.1 | 12 | 2.3 | 9 | 1.2 | 11 | 47.5% | 6 | 58.8% | 1 | 11.3% | 7 |
| Carson Wentz | PHI | -1.0 | 28 | 0.2 | 20 | 1.1 | 12 | 37.5% | 20 | 33.3% | 24 | -4.2% | 16 |
| Sam Bradford | MIN | -0.9 | 27 | 0.1 | 21 | 1.0 | 13 | 46.3% | 7 | 42.1% | 13 | -4.2% | 18 |
| Aaron Rodgers | GB | 4.5 | 2 | 5.2 | 1 | 0.7 | 14 | 44.2% | 11 | 51.0% | 4 | 6.8% | 9 |
| Quarterback | Team | ALEX 1-5 | Rk | ALEX 6-10 | Rk | DIFF | Rk | CONV% 1-5 | Rk | CONV% 6-10 | Rk | DIFF | Rk |
| Jameis Winston | TB | 1.7 | 8 | 2.3 | 8 | 0.6 | 15 | 40.4% | 15 | 55.0% | 3 | 14.6% | 4 |
| Joe Flacco | BAL | 0.9 | 14 | 1.2 | 17 | 0.3 | 16 | 38.7% | 17 | 28.9% | 26 | -9.8% | 26 |
| Kirk Cousins | WAS | 1.3 | 10 | 1.6 | 14 | 0.3 | 17 | 40.4% | 14 | 44.2% | 10 | 3.8% | 11 |
| Philip Rivers | SD | 0.5 | 17 | 0.6 | 18 | 0.1 | 18 | 45.5% | 8 | 49.0% | 6 | 3.6% | 13 |
| Case Keenum | LARM | -0.2 | 22 | -0.7 | 25 | -0.5 | 19 | 33.3% | 27 | 45.2% | 9 | 11.9% | 6 |
| Trevor Siemian | DEN | 0.4 | 18 | -0.1 | 22 | -0.5 | 20 | 47.8% | 5 | 35.8% | 21 | -12.0% | 27 |
| Andrew Luck | IND | 2.2 | 6 | 1.3 | 16 | -0.9 | 21 | 44.7% | 10 | 37.9% | 18 | -6.7% | 21 |
| Matt Ryan | ATL | 0.4 | 19 | -0.8 | 26 | -1.2 | 22 | 45.0% | 9 | 40.0% | 17 | -5.0% | 19 |
| Cam Newton | CAR | 3.1 | 4 | 1.8 | 12 | -1.3 | 23 | 43.9% | 12 | 37.5% | 19 | -6.4% | 20 |
| Russell Wilson | SEA | 0.9 | 13 | -0.3 | 23 | -1.3 | 24 | 36.4% | 23 | 40.0% | 16 | 3.6% | 12 |
| Tyrod Taylor | BUF | 4.4 | 3 | 2.5 | 5 | -1.9 | 25 | 35.9% | 25 | 50.0% | 5 | 14.1% | 5 |
| Carson Palmer | ARI | 0.6 | 15 | -1.6 | 29 | -2.2 | 26 | 43.2% | 13 | 47.8% | 7 | 4.6% | 10 |
| Ryan Fitzpatrick | NYJ | 1.4 | 9 | -1.0 | 27 | -2.4 | 27 | 36.7% | 22 | 27.3% | 28 | -9.5% | 25 |
| Brock Osweiler | HOU | 2.7 | 5 | -1.0 | 28 | -3.8 | 28 | 39.2% | 16 | 35.0% | 22 | -4.2% | 17 |
| Ben Roethlisberger | PIT | 6.8 | 1 | 1.8 | 13 | -5.0 | 29 | 52.0% | 2 | 28.0% | 27 | -24.0% | 28 |
Roethlisberger is tied for the biggest drop in ALEX, but it should be noted that Week 6 was the Miami contest in which he tried to play through a torn meniscus that he suffered during the game. Still, the huge fall in conversion rate is troubling for Pittsburgh, which must play great offense down the stretch just to have a shot at the playoffs.
[ad placeholder 3]
Elsewhere in the AFC North, Andy Dalton was very low in ALEX to start the season, but has improved to his usual range in the last month, which has also led to the third-highest improvement in conversion rate for his offense. Dalton now has the biggest difference between his ranking in DVOA (16th) and conversion rate (32nd). He has faced the second-longest third-down situations on average this season.
Believe it or not, but Alex Smith has the third-highest ALEX (plus-3.3) since Week 6. He is on track for his highest ALEX season since 2007, though he has still struggled with converting these plays into first downs. Blake Bortles and Marcus Mariota are two young quarterbacks who have upped their ALEX since Week 6, and also seen the two biggest conversion rate increases.
Overall, the correlation between ALEX in the two splits is 0.35, while the correlation between early and late conversion rate is just minus-0.06, practically non-existent. This makes sense in that the quarterback has more control of where he throws the ball, than what can happen after the ball has been released. Rodgers has certainly remained consistent here, ranking in the top two in ALEX in both splits, and he has actually increased his conversion rate since Week 6.
The training wheels are slowly coming off of Dak Prescott in Dallas, and while he has taken more chances since Week 6, his conversion rate is still solid for a rookie at 44.1 percent. Carson Wentz, Derek Carr, and Ryan Fitzpatrick are the only quarterbacks to rank 20th or worse in conversion rate in each split. All three are fairly low in ALEX too, though Fitzpatrick did not start the season that way. However, he won't get a chance to throw shorter passes to avoid more pick parades, because he has lost his job in New York. Drew Brees and Philip Rivers are the only two quarterbacks to rank in the top 10 in conversion rate in both splits. They also have had the two shortest averages for need yards this season.
The distance to go is certainly a factor in how successful, and perhaps how aggressive, quarterbacks are going to be on third down. Just like last year, I have provided a "heat map" to show how the quarterbacks stack up in the different third-down splits: short (1 or 2 yards to go), medium (3 to 6 yards) and long (7-plus yards). The colors indicate where a player is well above average (darker green) versus below average (darker red) based on standard deviations. Those conversion rates are also shown with a ranking.
That is an impressive clean sweep by Brady in conversion rate for each distance, especially highlighted by his success on third-and-long. Again, he has only played five games, but we'll see how this plays out. Brees is the only quarterback to rank in the top 10 for each distance split. Derek Carr, Brock Osweiler, and Blaine Gabbert are the only quarterbacks to rank 20th or lower in all three distance splits.
[ad placeholder 4]
Eli Manning is having one of those "Good Intentions, Bad Results" (high ALEX, low conversion rate) seasons, much like Tyrod Taylor had a year ago, but Taylor has been hitting his passes more efficiently in the last month.
Carson Palmer finished 2015 second in ALEX (plus-4.2) and first in conversion rate (53.2 percent), but this season has been very different for Arizona's passing game. He is still eighth in conversion rate despite the fourth-lowest ALEX, but that is skewed by a minus-44 ALEX pass when Palmer threw a screen on third-and-44 against Carolina. Without that play, Palmer's ALEX is plus-0.1 (which would rank 26th) instead of minus-0.4 (31st). Still, Palmer's rank in Short% (22nd) is far from his second-place finish of a year ago.
When we return with an update for the end of the regular season, we'll see if Brady has sustained his hot start, and hopefully will have learned something about Jared Goff in Los Angeles. The No. 1 overall pick makes his debut this week.
Comments
9 comments, Last at 21 Nov 2016, 4:27pm
#1 by brambo // Nov 18, 2016 - 5:29pm
Hmm... you kinda lost me with ALEX stats. Ok, I don't always read the entire article (and didn't again here), but looking at the numbers, any stat that has Blake Bortles in the top ten while Wilson, Carr, and Ryan are bringing up the rear strikes me as completely irrelevant.
Your top three are cool I guess, but after that it just shows that ALEX really has nothing to do with QB and team success. These charts are as good upside-down as right-side-up. Am I wrong, or should I read the article?
#3 by ramirez // Nov 18, 2016 - 9:38pm
You'll like ALEX if you buy into the notion that air yards are fundamentally more valuable to offense than YAC. A team that runs a lot of screens, for example, would be able to produce a high conversion rate on 3rd down with a low ALEX figure. When you consider that guys like Kurt Warner, Steve Young, Peyton, Brady, Rodgers, and Rivers have all produced standout seasons with a high rate of YAC, I don't see any reason to think that yards gained after the catch hurt offense. In fact, I suspect that the most unstoppable type of offense, like the 2007 Patriots or 2001 Rams, is one that produces a high rate of both YAC and air yards.
#5 by Scott Kacsmar // Nov 19, 2016 - 5:49pm
I always think an article should be read before being commented on, but my issue is with "any stat that has Blake Bortles in the top ten while Wilson, Carr, and Ryan are bringing up the rear strikes me as completely irrelevant." This line of thinking is not good for any stat, as stats should not be judged that way. Does Chad Pennington invalidate completion percentage? Do Joey Harrington, Steve Walsh and Doug Williams invalidate sack rate? Does Sam Bradford invalidate interception rate? Are you comfortable with where Philip Rivers and Tony Romo rank all time in passer rating compared to other all-time greats?
When a not-so-good QB ranks high in a stat, that is usually a great opportunity to learn something specific about their style of play that can explain why they're so good at that stat, but not other areas of the game. In the case of Bortles, for example, I'm not surprised that he has a high ALEX, because I know the organization has looked into such data on throwing beyond the sticks in these situations. The fact that he hasn't been very successful with it, which is why I present other stats to go along with ALEX for context, is also not surprising. Just as you shouldn't judge a player on one stat, don't judge a stat on one player.
#6 by Chris West // Nov 20, 2016 - 12:54pm
I was under the subjective impression that the Seahawks had been much better about attacking downfield the past two weeks vs the Pats and Bills. How much are Wilson's week 6-10 numbers skewed by the Arizona game, where he faced a ton 3rd and 20+ to go situations in the second half?
#8 by fongs2 // Nov 21, 2016 - 10:20am
Is there a reverse ALEX which shows 3rd down defenses? If the variance between the defenses is larger than the variance between QB's, it might show that the defense dictates air attack plans more than the offense. Or it might mean nothing.
#9 by Scott Kacsmar // Nov 21, 2016 - 4:27pm
The 2015 defensive numbers are here http://www.footballoutsiders.com/alex/2016/alex-season-review but I never went back to look at previous seasons yet. I'd expect a tighter range for the defenses than for individual QBs.