≡ Menu

Last off-season, I looked at passing performance on “third downs”, and I thought it would be fun to revisit that idea this summer. As before, I am putting that term in quotes because I’m including fourth down data in the analysis, but don’t want to write third and fourth down throughout this post.

To grade third down performance, I included sacks but discarded rushing data (again, just in the interest of time). The first step in evaluating third down performance is to calculate the league average conversion rate on third downs for each distance. Here were the conversion rates I calculated last year.

To GoPassesFirst DownsRateSmoothed Rate

In the interest of time, I’m going to just use those same rates again.1 Once we know the expected conversion rate for each distance, it’s easy to grade the quarterbacks. The next table is a bit complicated, so let me just walk you through the best third down quarterback in the league last year. Peyton Manning had 169 third down passing plays (either pass attempts or sacks); the average distance to be gained on those plays was 6.8 yards, and the expected conversion rate (based on the table above for each play) was 37.2%, that means Manning would be expected, if he was an average quarterback, to convert 62.9 first downs. In reality, he converted 81 first downs, a 47.9% conversion rate. This means Manning produced 18.1 more first downs than expected, and had a 1st down rate over expectation of 10.7%, the highest in the NFL (that’s the column by which the table is ranked). On third downs, Manning averaged 6.6 Net Yards per Attempt and 7.43 Adjusted Net Yards per Attempt.

RkPasserTm3DAvg To GoExp. RtExp ConvAct 1D1D Rate1D Ov Exp1D Rt Ov ExpNY/AANY/A
1Peyton ManningDEN1696.837.2%62.98147.9%18.110.7%6.67.43
2Philip RiversSDG1656.438%62.77947.9%16.39.9%8.138.41
3Josh McCownCHI597.336%21.22745.8%5.89.8%7.088.44
4Aaron RodgersGNB947.136.4%34.24244.7%7.88.3%7.987.82
5Nick FolesPHI948.732.7%30.73840.4%7.37.7%6.537.6
6Drew BreesNOR1837.735.2%64.47842.6%13.67.4%7.337.38
7Ryan FitzpatrickTEN115736.7%42.24942.6%6.85.9%7.347.95
8Colin KaepernickSFO1408.134.2%47.85539.3%7.25.1%7.467.81
9Andy DaltonCIN1767.635.4%62.37140.3%8.75%6.536.13
10Robert Griffin IIIWAS1517.535.5%53.66140.4%7.44.9%5.265.13
11Jake LockerTEN707.735.2%24.62840%3.44.8%6.277.41
12Jay CutlerCHI1117.735.1%38.94439.6%5.14.6%6.867.45
13Matt RyanATL1907.336.1%68.67740.5%8.44.4%6.626.14
14Geno SmithNYJ1568.732.7%515736.5%63.9%6.285.47
15Cam NewtonCAR1378.233.9%46.55137.2%4.53.3%5.645.64
16Matthew StaffordDET2007.336.2%72.37839%5.72.8%6.756.15
17Russell WilsonSEA1527.635.5%545737.5%32%5.765.34
18Thaddeus LewisBUF607.834.8%20.92236.7%1.11.8%6.726.63
19Ben RoethlisbergerPIT1747.435.8%62.36537.4%2.71.5%5.485.86
20Kellen ClemensSTL947.435.8%33.63537.2%1.41.5%6.887.95
21Matt SchaubHOU1028.233.7%34.43534.3%0.60.6%5.714.28
22Tony RomoDAL1457.934.5%505034.5%00%4.314.31
23Eli ManningNYG1828.633.1%60.36033%-0.3-0.2%5.434.25
24Tom BradyNWE189736.7%69.46936.5%-0.4-0.2%4.814.47
25Carson PalmerARI1827.535.6%64.86435.2%-0.8-0.4%6.265.14
26Terrelle PryorOAK1078.334.7%37.13633.6%-1.1-1.1%5.794.07
27Ryan TannehillMIA1937.934.9%67.36533.7%-2.3-1.2%4.553.16
28Josh FreemanTAM608.833.1%19.91931.7%-0.9-1.4%5.324.57
29Joe FlaccoBAL2237.934.7%77.57433.2%-3.5-1.6%5.223.25
30Sam BradfordSTL816.737.5%30.42935.8%-1.4-1.7%4.775.44
31Matt FlynnGNB617.236.3%22.12134.4%-1.1-1.8%4.524.03
32Alex SmithKAN154834.6%53.35032.5%-3.3-2.1%6.036.32
33Christian PonderMIN817.834.8%28.22632.1%-2.2-2.7%4.912.69
34Matt CasselMIN837.435.8%29.72631.3%-3.7-4.5%6.575.9
35Chad HenneJAX1717.835.1%605230.4%-8-4.7%4.824.99
36Matt McGloinOAK706.637.7%26.42332.9%-3.4-4.9%6.214.5
37Mike GlennonTAM1608.633.1%534528.1%-8-5%3.292.95
38Andrew LuckIND1586.637.7%59.65132.3%-8.6-5.5%5.135.32
39Case KeenumHOU897.635.3%31.52629.2%-5.5-6.1%4.554.44
40Brandon WeedenCLE918.732.7%29.72426.4%-5.7-6.3%4.192.65
41Jason CampbellCLE1007.136.8%36.82929%-7.8-7.8%4.664.81
42EJ ManuelBUF1097.136.5%39.73128.4%-8.7-8%4.173.49
  • Manning was really, really good on third downs. But I guess we’re past the point of being surprised by that. Philip Rivers was second in both 1st downs and 1st down rate over expectation (and first in NY/A), and his ability to keep the chains moving was one of the biggest reasons for San Diego’s success in 2013.
  • Josh McCown had a magical season — or third of a season — and it showed on third downs, too. Despite facing an average distance of over 7 yards, he converted on 46% of his third down opportunities.  He also produced the highest ANY/A on third downs of any passer in the above table.
  • Nick Foles was consistently throwing in unfavorable third down situations (on average, with 8.7 yards to do). Presumably this is because the Eagles were much more likely than other teams to run on third and short, but based on the average distance, Foles was expected to convert just 32.7% of his first downs. In reality, he had a 40.4% rate, which vaults him into the top five.
  • Andy Dalton was very good on third downs last year, while Russell Wilson was just average. Don’t shoot me, I’m just the messenger.
  • Andrew Luck really, really struggled on third downs. Now we know that Luck adds a ton of value on the ground, particularly on third downs, and he draws a lot of penalties for the Colts. But from a pure passing perspective, he was terrible on 3rd downs in 2013. Despite his reputation as a very “clutch” performer, a 32% conversion rate simply won’t cut it considering that Luck only passed on some of the most favorable distances.2 Sure, including his scrambles would help, but Luck was consistently throwing in favorable third down situations but ended up with the 2nd most negative number of passing first downs added relative to average. Luck was much better in 2012, so this could just be a one-year blip (Dalton was also terrible on third downs in 2012, which is more evidence that third down performance is more random than some would like to believe).
  • The only quarterback to provide fewer first downs relative to expectation was EJ Manuel. He was terrible on third downs with a 28.4% conversion rate. Of the other rookie quarterbacks, Mike Glennon was similarly bad, but Geno Smith was actually above average.

What stands out to you about third down performance?

  1. Third down conversion rates went up ever so slightly in 2013, but hey, I’m a good enougher. I’ve also been really busy on non-FP related activities, so cutting corners is a must. []
  2. Note: If you put these numbers into the PFR Play Finder, they match up, but it takes a minute to reconcile them. The Colts had 161 passing plays on 3rd or 4th downs in 2013, with three of them coming from Matt Hasselbeck. Indianapolis gained 54 first downs on those plays, but two of them came via penalty against Kansas City (horse collar, taunting) and an unnecessary roughness penalty in Houston. So for these purposes, Luck gets credit for only 51 out of 158 third down passing plays. []
  • The entry for Robert Griffith should be Robert Griffin.

    • Chase Stuart

      Fixed, thanks.

  • Dave

    I combined 2012 and 2013 for qbs that played both years as that gets our average sample size up to 303 passes. Year to year correlation on 1d over expected was 0.24

    Top 5 1d rate over expected

    1. Peyton 10.10%

    2. Brees 6.8%
    3. Ryan 6.1%
    4 Rodgers 5.4%
    5.Rivers 4.8

    Bottom 5
    1. Henne -6.4%
    2. Weeden -4.6%
    3. ASmith -2.5%
    4. Bradford -2.3%
    5. Tannehil -2.2%

    • Chase Stuart

      Interesting, thanks for doing the work on that.

  • What is your most reliable source for 1D data? Maybe I am too green, but I can’t seem to find the info on PFR.

    • Chase Stuart

      This was the rare instance where it was not pulled from PFR.

  • David

    Geno Smith sticks out. I think he’s a better QB than people give him credit for. You don’t go 8-8 as a rookie QB without doing something right. I would be a shame for the Jets to start Vick over Geno, he has a lot going for him.

    P.S. Where is Vick on this list? He started 6 or 7 games last year.

    • David

      “IT would be a shame…”

    • The Fanchize

      In 2009 the Jets had a rookie QB who went 8-7 and then won two playoff games on the road. The next year he went 11-5 and won two more road playoff games. A good portent for Smith.

  • Quinton

    I tend to dismiss third down as just fluky and odd results just happen cause you’re cutting your sample size by 2/3rds to 75%. I would be curious if anyone overperforms their own expectations. I imagine this would be a fair amount of work but for example, Peyton Manning is the best against average (he’s the best against everything. Except in Super Bowls. Boom!…Kidding. Don’t ban me) but is Manning better than what we’d expect based on what Manning does on first and second down?

    • Ty

      I think 3rd Down Conversions are not as consistent year to year, but I believe QBs have some type of control over them. If you look at the best QBs and the worst QBs, the best QBs will have better 3rd down conversion numbers (not by significant amounts, but noticeably better, as noted above by Dave).

      • Quinton

        Of course, because they are better quarterbacks. I’m after the question, are there QBs who consistently elevate their play? It’s actually relatively close to the issue of “clutchness” but third down is a more well-defined situation

    • Richie
  • Ajit

    Pretty great idea, though I do wonder if we can make it a bit better. Obviously, not all yards are created equal. If I remember correctly, FO showing that, roughly speaking, the 0-20 zones of the field were about 1.5 times the value of the zones from 21-21. Since Bryan has access to the game logs, he can probably query what percent of the different first downs occur in those regions and then an average can be taken to further enrich the stat. Just a thought.

  • Nick Bradley

    Okay that was ridiculously overcomplicated.

    How about “3rd down EPA/play”?

  • Bill

    The unfactored element is the game success of the running game and field position when attempting third down. I agree that the field position needs to have some weighting system to factor the broader options available to a QB while in his own territory. Average starting position intuitively seems to have an effect of both the frequency of possible third down play calls and the options available plays. The effectiveness of the running game on both first and second downs is critical on all third down options, as well as the frequency of first and second down run calls. I believe the lack of a running game was the primary factor in both Flacco’s and Eli’s drop in overall effectiveness. Romo’s anomaly of success was that he threw so often on first and second down, that there was no running threat and the DBs were more fatigued from constant coverage.