clock menu more-arrow no yes

Filed under:

What can we learn from Drew Lock’s first five starts? Part 1: the data

New, comments

Some stats, some history and some good vibes for the future of the Denver Broncos.

Oakland Raiders v Denver Broncos Photo by Dustin Bradford/Getty Images

Drew Lock made five starts for the Broncos in 2019. There have been 229 different NFL QBs who have made their first five starts at quarterback over the past 30 seasons. Drew Lock’s passer rating in his first five starts ranks 40th among those 229 - just behind Brett Favre and just ahead of Chase Daniel.

Those who are high on Drew Luck and going to latch onto to the fact that his numbers in his first five were similar to Brett Favre, while those who are not sold yet will point to his similarity to Chase Daniel or Case Keenum or Dan Orlovsky. Below is a subsection of the full 229 QB table showing Drew Lock and the five QBs above and below him. For space reasons we will not show the full table.

Rank Quarterback Cmp Att Cmp% Yds TD INT Rating
35 Ty Detmer 104 172 60.47 1270 8 3 91.5
36 Jake Locker 76 127 59.84 903 6 2 90.8
37 Steve McNair 67 125 53.6 1041 6 2 90.8
38 Scott Mitchell 77 137 56.2 1136 8 4 90.8
39 Brett Favre 109 169 64.5 1135 6 2 90.7
40 Drew Lock 100 156 64.1 1020 7 3 89.7
41 Chase Daniel 112 163 68.71 1103 6 4 89.6
42 Case Keenum 88 160 55 1161 8 2 89.6
43 Brad Johnson 100 162 61.73 1272 10 7 88.8
44 Dan Orlovsky 96 162 59.26 1163 6 2 88.6
45 Jay Cutler 81 137 59.12 1001 9 5 88.5

Is there any predictive value here?

While it is interesting to see how his performance compares with other recent QBs in their first five starts, it is more informative to look at how different QBs have gone on to perform in their next five starts. Note that I have not corrected for era - these are the raw numbers as the players compiled them. The average passer rating in the league in 1992 when Favre made his first five starts was 75.3. In 2019 it was 88.0. In other words, Favre’s performance in 1992 in his first five starts was much more impressive than Lock’s in 2019 because of how it compared to the rest of the league. Lock’s passer rating was a little above average. In 1992 future Hall-of-Fame QBs Troy Aikman and Jim Kelly made the Pro-Bowl with passer ratings of 89.5 and 81.7 respectively.

This is not an article about era adjustment of passing data, though. That might be another topic for the off-season. So let’s focus on the comparison of individual QB changes in performance from the first five starts to their next five starts. That does not need to be adjusted for era.

First let’s see how Drew Lock compared to his peers, fellow rookie quarterbacks who made at least five starts. Comparing their performance in their first five starts we see this (note that I counted Gardner Minshew’s game where he didn’t start but played 90% of the offensive snaps as a start).

Quarterback Cmp Att Cmp% Yds TD INT Rating Sacks Sk Yds Y/A AY/A
Gardner Minshew 102 169 60.36 1167 7 1 92.5 13 90 6.91 7.47
Drew Lock 100 156 64.1 1020 7 3 89.7 5 26 6.17 6.2
Kyler Murray 126 201 62.69 1324 4 4 80.1 21 139 6.59 6.09
Daniel Jones 104 171 60.82 1127 6 7 74.9 18 126 6.59 5.45
Dwayne Haskins 60 111 54.05 661 2 2 70.4 18 116 5.95 5.5

In his first five starts, Gardner Minshew played the best of the rookie QBs this season, despite being drafted the lowest of the rookie Qbs who made at least five starts. Lock had the second best performance in his first five.

Drew Lock has already had his “humbling” start as I chronicled last month. Lock’s humbling came in his third starts against the Chiefs in KC in the snow. Kyler Murray’s was in his 13th start (passer rating of 56.4). Minshew’s was in his fifth start (passer rating of 51.4). Jones had his worst start in his 4th start (rating of 35.2). Haskins was humbled in his third start (rating of 47.5).

How do Quarterbacks do in the next five starts?

There were 197 Qbs have made starts 6-10 during this 30 year time-frame. For almost every one of the 32 Qbs who didn’t make five more starts, it is because they were terrible in their first five. One exception is former #1 overall draft pick Alex Smith. He was horrendous in first five starts, but the 49ers were committed to him so he continued to start. Smith had a passer rating of 23.1 in his first five starts. FWIW, Paxton Lynch only made four starts at QB before his NFL career ended. Of those 197 who made starts 6-10 during this time frame, 13 of them did not make starts 1-5 during this time frame. So there were 184 QBs who made both starts 1-5 and starts 6-10 in the period from 1990 to 2019.

Of these 184, in starts 6-10, the average QB improved his completion percentage by 0.5%, threw for only 23 more yards, threw 0.3 more touchdowns, threw the exact same number of interceptions and improved his passer rating by 2.0 relative to starts 1-5. If 30 years of NFL history has any predictive value, what that means for Drew Lock is that we should expect about the same performance in games 1-5 of 2020 that we saw from him in games 12-16 of 2019. Projecting Lock’s performance from his first five starts to a full sixteen games would produce this stat line: 320 completions on 499 attempts, 3264 yds, 22 TD, 10 INT, 16 sacks and a passer rating of roughly 90.

I will be doing a full study at one point during this off-season to look at how recent quarterbacks who started 5-8 games as rookies at the end of their rookie seasons, fared in the second season. That study, however, has not been done yet. I would Lock’s passing yards to be higher from a full season. Were he to compile the stat line above, that would only be 6.5 yards per passing attempt, which is fairly low. That would have ranked 31st among qualifying quarterbacks this season; only Mason Rudolph (6.2) and Mitch Trubisky (6.1) had lower YPA in 2019. League average was 7.2. So if Lock is average in YPA in 2020, he would throw for ~3600 yards

I find it informative to look at the guys who regressed the most in starts 6-10 and the guys who improved the most in starts 6-10.

Quarterbacks who regressed the most

Rank Quarterback Cmp% 1st 5 Yds 1st 5 TD 1st 5 Int 1st 5 Rate 1st 5 Cmp% 2nd 5 Yds 2nd 5 TD 2nd 5 Int 2nd 5 Rate 2nd 5 Delta Comp% Delta Yds Delta TD Delta INT Delta Rating
1 Tony Romo 71.2 1394 10 2 115.8 60.5 1247 6 8 77.1 -10.7 -147 -4 6 -38.7
2 Kyle Allen 64.4 1129 9 0 107.5 59.6 1278 6 9 69.8 -4.8 149 -3 9 -37.7
3 Terrelle Pryor 63.6 995 6 3 92.2 51.9 921 3 8 54.9 -11.7 -74 -3 5 -37.3
4 Austin Davis 65.4 1328 9 3 96.6 57.7 951 3 8 60.1 -7.7 -377 -6 5 -36.5
5 Kurt Warner 71.4 1328 15 3 131.4 63.4 1321 12 4 97.1 -8.0 -7 -3 1 -34.3
6 Zach Mettenberger 60.7 1270 8 5 91.8 58.4 755 2 6 59.3 -2.2 -515 -6 1 -32.5
7 Geno Smith 60.3 1289 7 8 80.3 51.8 811 1 8 48.0 -8.5 -478 -6 0 -32.3
8 Jake Locker 59.8 903 6 2 90.8 53.5 1121 4 9 59.8 -6.4 218 -2 7 -31.0
9 Shane Matthews 62.7 1043 8 3 86.9 52.4 832 1 5 56.0 -10.3 -211 -7 2 -30.9
10 Carson Wentz 65.0 1186 7 1 99.9 61.8 1153 4 6 71.9 -3.2 -33 -3 5 -28.0

This group breaks down in to three sub-groups:

  1. guys who played well then regressed
  2. guys who played ok then stunk it up
  3. Kurt Warner

Kurt Warner was so good in his first five starts that even with regression he was still well above average in his second five starts. Remember this was 20 years ago when the average NFL QB had a passer rating in the high 70s. The league completed 57.1 percent of all passes in 1999 when Kurt Warner was doing this. He completed 71.4 percent of his passes in his first five starts. That’s just crazy. Warner’s adjusted completion percentage was 130 for that full season - that is one a scale where 100 is average. Warner actually led the league in adjusted completion percentage during his first three years as a starter. FWIW, Drew Lock had an adjusted completion percentage of 100 this season.

Tony Romo, Kyle Allen and Carson Wentz all fit into that first sub-group. Combined the three threw 26 TDs and 3 INTs in their first five starts. They combined to throw 16 TDs and 23 INTs in their next five starts. Interestingly enough, Romo and Allen were both undrafted out of college. Carson Wentz was taken with the second overall pick. Romo and Allen were only starting because of injuries to the expected starters, Drew Bledsoe and Cam Newton. Wentz was given the starting job in game one of his rookie season.

From a statistical perspective Lock’s numbers are not that much different from Terrelle Pryor’s. Both are mobile QBs. Hopefully, Lock’s play in his next five starts is better than Pryor’s.

Quarterbacks who improved the most

Rank Quarterback Cmp% 1st 5 Yds 1st 5 TD 1st 5 Int 1st 5 Rate 1st 5 Cmp% 2nd 5 Yds 2nd 5 TD 2nd 5 Int 2nd 5 Rate 2nd 5 Delta Comp% Delta Yds Delta TD Delta INT Delta Rating
1 Alex Smith 45.9 551 0 10 23.1 58.2 1104 4 1 86.6 12.3 553 4 -9 63.5
2 Nick Foles 57.8 1135 4 3 76.3 64.3 1355 14 1 119.9 6.5 220 10 -2 43.6
3 Kent Graham 40.5 535 2 3 50.0 59.3 937 8 2 89.8 18.8 402 6 -1 39.8
4 Eli Manning 44.4 632 3 7 44.8 55.3 1034 8 4 83.1 10.9 402 5 -3 38.3
5 Baker Mayfield 56.5 1270 8 6 76.5 73.2 1406 11 4 114.4 16.7 136 3 -2 37.9
6 Anthony Wright 43.0 557 3 8 37.8 51.1 946 8 6 75.6 8.2 389 5 -2 37.8
7 Tarvaris Jackson 51.9 728 3 8 50.6 64.0 639 3 2 87.2 12.2 -89 0 -6 36.6
8 Jeff Garcia 52.8 939 3 5 63.4 66.5 1441 8 3 98.3 13.7 502 5 -2 34.9
9 Joe Flacco 62.5 844 1 7 60.6 61.4 969 7 2 95.2 -1.1 125 6 -5 34.6
10 Trent Dilfer 47.7 823 2 9 44.5 54.6 802 1 0 77.9 6.9 -21 -1 -9 33.4

This is an interesting group as it contains four quarterbacks who “won” a Super Bowl: Nick Foles, Eli Manning, Joe Flacco and Trent Dilfer. By passer rating all of them were well below league average in their first five and all except Baker Mayfield threw for less than one TD per game. Smith, Flacco and Dilfer probably would have been benched if their teams had another viable option at QB. They combined to throw three touchdowns and 26 interceptions in their first five starts. That was flipped in their in the second five where they combined to throw 12 touchdowns and 3 interceptions.

Alex Smith had the greatest turn-around, but all of the “most improved” quarterbacks basically stopped throwing the ball to the other team. What’s really interesting is that Trent Dilfer was not really throwing the ball to anyone. He threw for one touchdown total in the second five starts while only completing 13 passes per game during that span. His 121 passing attempts was one of the lowest number of attempts among the 197 who made starts 6-10 since 1990. That being said Tim Tebow had the fewest with 82 passing attempts in starts 6-10. So while Trent Dilfer was only completing 13 per game, Tebow was only throwing 16 per game.

With the exceptions of Anthony Wright and Tavaris Jackson, the rest of this group were (or are) multi-year starting QBs in the NFL. So it’s a good sign for the future when a quarterback can learn from his mistakes and show dramatic improvement in his second five relative to his first five.

The Rest of the Story

So this is part 1. What will follow is a narrative that will take you through the history of four quarterbacks whose stats in their first five starts were similar to Drew Lock and who share similarities with Lock either by franchise, physical tools, mental acumen or all three: Matthew Stafford, Trevor Siemian, Jay Cutler and Brett Favre.