Thad Lewis was solid in his second career NFL game, completing 19 of his 32 pass attempts for 216 yards and rushing seven times for 17 yards. His first pass was a 47 yard completion to TJ Graham (the pass was caught 27 yards beyond the line of scrimmage), which led to the first touchdown of the game: a Thad Lewis scramble for three yards.
26 of Lewis’ passes were thrown ten or fewer yards beyond the line of scrimmage. He completed 16 (62%) of those throws for 85 yards. He did equally as well when throwing to his left or right (left or right of his throwing spot on the field, not to the left or right side of the field), which is a development (or just a better passing attack) from his Cleveland game last season. Lewis was 11/18 throwing to his left and 8/14 on passes to his right. Also unlike his first NFL game last year, Lewis stayed in the pocket when throwing passes. No pass attempts (just one sack on 4th and goal) happened on a rollout. That could be a result of the foot injury early in the game. Lewis’ pass chart against Cincinnati is below.
Through two games, Lewis now has 64 career pass attempts. It’s a really small sample, but we are starting to get an idea about his strengths and weaknesses. Lewis doesn’t throw long passes very often, as 88% of his passes have traveled thirty total yards or fewer. 64% of those passes were completed. The eight pass attempts that traveled further than thirty total yards have only been caught three times.
Using the 64 passes, I reconstructed Lewis’ estimated completion rate by total pass distance. First I wanted to compare Lewis’ performances from his first and second games. The graph below shows two logistic regressions based on the quarterback’s two NFL games. It looks like the Bills’ coaching staff was able to get a more consistent result from Lewis, since his completion rate drop off was not as severe as his first game.
I then combined the results from the two games to the regression already run for EJ Manuel’s 150 attempts. Lewis appears to be more accurate on passes longer than nine yards, but it’s important to keep in mind that Lewis’ regression has a bit of a sample size error. Every pass attempt he makes has a relatively large impact to the shape of his expected completion curve. Manuel’s curve also suffers from a small sample size, but it’s not as severe as the Lewis curve. The comparison is below, but please don’t start with the “LEWIS FOR STARTER” commentary yet.
Lewis has another test in Miami this weekend. The Dolphins pass defense currently allows big gains (twenty yards or more) 9.6% of the time, which is ninth most in the NFL. Furthermore, they have only accumulated thirteen sacks (just under 6% of their defensive pass plays, good for 23rd in the NFL). Hopefully Lewis’
Add The Sports Daily to your Google News Feed!
