Writers of Pro Football Prospectus 2008

Most Recent FO Features

ProctorSpe15.jpg

» VN: Hits, Misses, and the Year Zero Effect

Bill Connelly looks at the college offenses, defenses, and overall teams that have improved (Air Force!) or regressed (North Texas!) the most in 2014. Year Zero is a real thing (sometimes).

11 Oct 2012

Varsity Numbers vs. Third Downs

by Bill Connelly

It is still one of my favorite Varsity Numbers columns. In October 2010, I wrote about Four Truths of college football statistics. The third of the four was this:

Truth No. 3: Leverage Rate gives us what we think we get from third-down conversion rate.

This is probably a less obvious truth. Announcers and coaches talk endlessly about how games are won and lost on third downs, and it is absolutely true. But it is also without context. Some defenses, particularly young ones, give up far too many demoralizing third-and-8 conversions. But most of the time, a team that allows a high third-down conversion rate is doing so because they're giving up too many yards on first and second down, and their opponent is converting third-and-3 instead of third-and-8.

The difference in the level of success on standard downs and passing downs is staggering. If you have a significant talent advantage -- always possible in college football -- you might be able to get away with falling into passing downs. But the team that wins is the team that better avoids passing downs. That is why Leverage Rate is included atop the Varsity Numbers box scores I analyze. If your Leverage Rate is too far below the national average of 68 percent, then your quarterback better be Colt McCoy (who was truly a magician at converting third-and-7 and, it appears, masked some serious, developing deficiencies for Texas on the offensive side of the ball) or you are probably going to struggle to win.

We have long accepted that third downs win games. If your third-down conversion rate is low and you lose a game, that is one of the first statistics both coaches and analysts will break out to explain the loss. But a failed series has often failed far in advance of third down. Even a bad offense has a better chance of converting third-and-2 instead of third-and-7, so it probably goes without saying that the more yards you gain on first and second down, the more likely you are to convert on third down.

I thought it would be interesting, then, to break down third downs a little further. I wanted to create an adjusted third down conversion rate of sorts that would tell us who is actually particularly competent on third downs and who just sets the table with good first and second downs.

Below, you will find a table with the following three measures:

  • Third-Down Conversion Rate. This is exactly what you think: a team's success rate on third downs.
  • Adj. Third Down Rate: This is an adjusted rate based on the comparison of a team's ability to convert on third-and-1, third-and-2, ..., third-and-10, et cetera, as compared to the national average. The idea is to isolate a team's true third-down abilities from the work it typically does on first and second downs.
  • Leverage Rate: Referenced above, Leverage Rate is the percentage of a team's snaps that took place on standard downs (first downs, second-and-6 or fewer, third-and-4 or fewer). This basically measures your ability to stay on schedule.

Team 3rd Down
Rate
Rk Adj. 3rd
Down Rate
Rk Diff. Leverage
Rate
Rk
Marshall 57.5% 3 62.5% 1 2 69.4% 63
Oklahoma State 59.6% 1 59.2% 2 -1 76.3% 4
Texas Tech 57.5% 4 57.6% 3 1 73.0% 18
Louisville 55.9% 6 57.1% 4 2 72.6% 23
Texas 58.1% 2 55.7% 5 -3 75.1% 11
Purdue 51.9% 13 55.6% 6 7 69.6% 58
UL-Monroe 52.2% 12 55.6% 7 5 68.6% 76
Kansas State 54.0% 7 54.5% 8 -1 75.2% 9
Clemson 54.0% 7 54.3% 9 -2 70.7% 41
Oklahoma 48.0% 27 53.1% 10 17 71.0% 37
San Jose State 50.7% 17 52.4% 11 6 70.4% 49
Pittsburgh 45.3% 42 52.2% 12 30 68.3% 79
Toledo 46.0% 34 52.1% 13 21 70.0% 53
Nevada 51.7% 14 52.1% 14 0 74.3% 14
Texas A&M 49.3% 22 51.8% 15 7 69.3% 67
Oregon 49.5% 21 51.3% 16 5 72.9% 19
Georgia 46.2% 33 50.6% 17 16 71.7% 32
Michigan 51.5% 15 50.5% 18 -3 69.6% 57
Air Force 56.2% 5 50.4% 19 -14 76.2% 5
Western Michigan 47.9% 29 49.9% 20 9 70.7% 42
Team 3rd Down
Rate
Rk Adj. 3rd
Down Rate
Rk Diff. Leverage
Rate
Rk
Arizona 45.7% 39 49.4% 21 18 71.8% 31
Middle Tennessee 52.5% 11 48.6% 22 -11 75.1% 12
Florida State 43.7% 48 48.5% 23 25 73.3% 17
Northwestern 51.1% 16 48.0% 24 -8 72.6% 22
Utah State 41.1% 64 47.2% 25 39 68.4% 77
Maryland 36.1% 94 47.1% 26 68 60.2% 122
West Virginia 53.4% 9 47.0% 27 -18 76.4% 3
Ohio 48.0% 27 46.6% 28 -1 70.5% 46
Baylor 46.0% 34 46.6% 29 5 75.5% 6
Syracuse 47.5% 31 46.6% 30 1 72.8% 21
Tennessee 42.9% 50 46.3% 31 19 69.4% 61
San Diego State 45.8% 38 46.3% 32 6 70.7% 43
Washington State 36.3% 92 46.0% 33 59 64.6% 114
Miami-FL 45.1% 43 45.8% 34 9 69.1% 69
South Carolina 45.5% 40 45.7% 35 5 68.2% 80
Alabama 49.2% 23 45.4% 36 -13 72.2% 27
Rutgers 39.2% 82 45.3% 37 45 61.9% 120
Boise State 42.2% 56 45.3% 38 18 67.9% 85
Northern Illinois 48.8% 25 45.2% 39 -14 71.9% 29
North Carolina 46.0% 36 45.1% 40 -4 68.8% 75
Team 3rd Down
Rate
Rk Adj. 3rd
Down Rate
Rk Diff. Leverage
Rate
Rk
Central Michigan 41.3% 62 44.8% 41 21 69.6% 55
Troy 50.6% 18 44.7% 42 -24 75.5% 7
Miami-OH 41.9% 58 44.6% 43 15 67.9% 83
Ole Miss 50.0% 19 44.4% 44 -25 74.0% 15
Central Florida 46.4% 32 44.2% 45 -13 71.5% 34
Massachusetts 39.8% 75 44.1% 46 29 59.7% 123
New Mexico State 40.0% 71 43.7% 47 24 66.9% 96
Cincinnati 48.2% 26 43.6% 48 -22 69.3% 65
UL-Lafayette 40.3% 69 43.1% 49 20 69.3% 66
Virginia Tech 40.0% 71 42.7% 50 21 67.3% 92
BYU 44.2% 46 42.7% 51 -5 69.3% 64
Kentucky 42.4% 55 42.5% 52 3 65.3% 108
UNLV 47.7% 30 42.4% 53 -23 67.5% 89
Texas State 39.7% 78 42.4% 54 24 64.4% 115
Washington 39.5% 80 42.3% 55 25 66.1% 103
South Florida 46.0% 36 42.3% 56 -20 70.0% 54
Ohio State 44.9% 44 42.2% 57 -13 72.4% 25
NC State 39.8% 76 42.1% 58 18 67.6% 87
Arizona State 41.2% 63 42.1% 59 4 72.2% 26
Houston 42.5% 53 42.1% 60 -7 69.6% 59
Team 3rd Down
Rate
Rk Adj. 3rd
Down Rate
Rk Diff. Leverage
Rate
Rk
Penn State 45.5% 40 41.7% 61 -21 70.8% 40
Virginia 40.2% 70 41.7% 62 8 68.9% 74
Ball State 43.6% 49 41.6% 63 -14 73.3% 16
Akron 38.1% 85 41.6% 64 21 65.9% 105
East Carolina 38.2% 84 41.3% 65 19 67.5% 88
Western Kentucky 49.2% 23 40.6% 66 -43 70.5% 47
Notre Dame 42.9% 50 40.5% 67 -17 70.9% 39
Mississippi State 41.4% 61 40.2% 68 -7 70.2% 50
Oregon State 39.7% 78 40.0% 69 9 69.0% 70
New Mexico 40.6% 65 39.4% 70 -5 70.6% 44
Florida Atlantic 40.5% 67 39.3% 71 -4 65.9% 104
Rice 39.8% 76 39.2% 72 4 66.6% 98
TCU 36.4% 91 39.0% 73 18 70.0% 52
Minnesota 42.5% 54 38.7% 74 -20 67.8% 86
Army 52.8% 10 38.5% 75 -65 77.1% 1
Georgia Tech 44.9% 44 38.4% 76 -32 72.5% 24
Florida 41.7% 59 38.1% 77 -18 72.1% 28
Louisiana Tech 44.1% 47 38.0% 78 -31 75.4% 8
Connecticut 37.9% 87 37.8% 79 8 69.4% 62
SMU 35.7% 97 37.5% 80 17 62.7% 119
Team 3rd Down
Rate
Rk Adj. 3rd
Down Rate
Rk Diff. Leverage
Rate
Rk
Hawaii 28.4% 119 36.9% 81 38 68.3% 78
UTSA 41.7% 59 36.7% 82 -23 67.2% 93
Tulane 30.1% 115 36.7% 83 32 57.2% 124
Arkansas State 40.0% 71 36.6% 84 -13 74.5% 13
Tulsa 40.0% 71 36.6% 85 -14 70.1% 51
California 33.3% 105 36.3% 86 19 64.8% 113
Michigan State 38.4% 83 36.3% 87 -4 67.4% 91
Iowa State 39.5% 80 36.3% 88 -8 69.0% 71
LSU 40.5% 68 36.2% 89 -21 75.2% 10
Temple 37.5% 88 36.1% 90 -2 67.1% 94
Colorado State 35.7% 97 36.0% 91 6 64.8% 112
USC 36.1% 95 36.0% 92 3 69.0% 72
UAB 42.9% 50 35.9% 93 -43 65.6% 107
Fresno State 37.0% 89 35.8% 94 -5 69.2% 68
Wake Forest 33.7% 104 35.2% 95 9 65.0% 110
Kent State 42.2% 57 34.8% 96 -39 68.9% 73
Indiana 38.1% 86 34.7% 97 -11 70.5% 48
Bowling Green 32.9% 108 34.5% 98 10 66.5% 101
Stanford 35.2% 99 34.2% 99 0 71.8% 30
Boston College 33.3% 105 34.2% 100 5 65.2% 109
Team 3rd Down
Rate
Rk Adj. 3rd
Down Rate
Rk Diff. Leverage
Rate
Rk
Florida International 31.9% 109 34.1% 101 8 67.9% 84
Utah 29.3% 117 33.0% 102 15 63.9% 117
UCLA 36.2% 93 32.2% 103 -10 72.8% 20
Illinois 35.9% 96 32.1% 104 -8 65.0% 111
Memphis 34.3% 102 31.8% 105 -3 70.6% 45
Iowa 40.6% 65 31.6% 106 -41 71.4% 35
Nebraska 50.0% 19 31.6% 107 -88 76.8% 2
Idaho 30.8% 114 31.1% 108 6 66.2% 102
Navy 33.3% 105 31.0% 109 -4 71.2% 36
Kansas 31.5% 111 30.9% 110 1 68.0% 82
Missouri 30.1% 116 30.8% 111 5 65.9% 106
Duke 34.4% 100 30.7% 112 -12 69.5% 60
Wyoming 31.8% 110 30.3% 113 -3 71.0% 38
Southern Miss 34.3% 101 29.9% 114 -13 67.5% 90
North Texas 31.4% 112 29.6% 115 -3 69.6% 56
Arkansas 31.3% 113 29.0% 116 -3 68.0% 81
South Alabama 27.8% 122 28.8% 117 5 66.6% 99
UTEP 28.1% 121 28.3% 118 3 61.9% 121
Buffalo 36.8% 90 27.8% 119 -29 66.8% 97
Vanderbilt 25.0% 123 27.1% 120 3 64.3% 116
Team 3rd Down
Rate
Rk Adj. 3rd
Down Rate
Rk Diff. Leverage
Rate
Rk
Colorado 33.8% 103 26.3% 121 -18 67.0% 95
Auburn 29.0% 118 25.5% 122 -4 66.6% 100
Wisconsin 28.2% 120 23.8% 123 -3 71.5% 33
Eastern Michigan 17.7% 124 16.2% 124 0 63.4% 118

(You can see the full list of teams' third-and-X success rates at Football Study Hall.)

All told, the variance between a team's third-down rate and adjusted third-down rate is not that high. A lot of teams at both ends of the spectrum are rather uniformly good or bad on most downs. Here is a look at the highest-ranked BCS offense in Adj. Third Down Rate (Oklahoma State) and the lowest (Wisconsin).

Wisconsin is oddly solid in third-and-9 and third-and-10, while Oklahoma State has a strange weak spot on third-and-5, but in general the two schools are better/worse than average no matter the distance.

For other teams, however, the story is a little messier. Here is a look at two messy teams, Nebraska and Washington State.

Nebraska is 19th in the country in third-down conversions because the Huskers avoid third-and-longs better than almost anybody. Their leverage rate is second in the country, and with a decent line blocking for quarterback Taylor Martinez and running backs Rex Burkhead and Ameer Abdullah, they are custom-built for converting on third-and-1 or -2. However, things fall apart on third-and-long. Taylor Martinez is a solid game manager, but drives end when Nebraska falls behind schedule. The No. 19 team in third-down conversion rate becomes the No. 107 team in Adjusted Third Down Rate.

On the flipside, Mike Leach clearly doesn't have the tools he needs to work his Air Raid offense at Washington State, and it shows dramatically on first and second downs. But his offense can still convert third-and-longs as well as almost anybody. The problem is that they face too damn many of them. The Cougars rank 92nd in third-down conversion rate but a healthier 33rd in Adjusted Third Down Rate.

Using Adjusted Third Down Rate and Leverage Rate, you get a much clearer picture of what truly ends up going into a team's third down rate. The raw rates are what end up mattering the most, but the gory details show you there are a lot of different ways to get to a particularly strong or poor third-down rate.

F/+ Rankings

Full rankings here.

F/+ Rk Team Record F/+ Last Week Change S&P+ Rk FEI Rk
1 Alabama 5-0 +50.2% 1 0 346.5 1 0.304 2
2 Oregon 6-0 +29.8% 2 0 284.0 2 0.204 11
3 Florida 5-0 +27.8% 8 +5 269.7 3 0.235 7
4 Notre Dame 5-0 +25.9% 5 +1 252.7 6 0.280 3
5 West Virginia 5-0 +24.6% 9 +4 242.2 13 0.308 1
6 South Carolina 6-0 +23.8% 12 +6 258.0 4 0.213 10
7 Oklahoma 3-1 +23.5% 13 +6 247.0 8 0.260 5
8 Texas 4-1 +23.0% 6 -2 245.4 10 0.258 6
9 Florida State 5-1 +22.0% 3 -6 245.9 9 0.235 8
10 Michigan 3-2 +18.8% 11 +1 256.1 5 0.121 24
11 Texas Tech 4-1 +17.7% 4 -7 236.3 17 0.198 12
12 Ohio State 6-0 +17.7% 25 +13 237.7 15 0.190 13
13 BYU 4-2 +17.2% 27 +14 250.1 7 0.118 25
14 Texas A&M 4-1 +16.7% 26 +12 238.7 14 0.166 15
15 Kansas State 5-0 +16.6% 22 +7 215.6 34 0.278 4
F/+ Rk Team Record F/+ Last Week Change S&P+ Rk FEI Rk
16 USC 4-1 +15.4% 16 0 235.8 18 0.154 17
17 Oklahoma State 2-2 +14.5% 14 -3 235.4 19 0.136 22
18 Arizona State 4-1 +14.1% 19 +1 230.7 22 0.152 18
19 Stanford 4-1 +14.0% 15 -4 237.3 16 0.118 26
20 Georgia 5-1 +14.0% 10 -10 242.3 12 0.092 32
21 LSU 5-1 +13.7% 7 -14 243.5 11 0.080 38
22 Iowa State 4-1 +13.3% 39 +17 225.9 26 0.160 16
23 Michigan State 4-2 +12.7% 18 -5 226.5 25 0.145 20
24 Baylor 3-1 +12.2% 33 +9 209.8 40 0.219 9
25 Arizona 3-3 +10.9% 23 -2 232.0 21 0.082 36

This Week at SB Nation

Posted by: Bill Connelly on 11 Oct 2012

5 comments, Last at 13 Oct 2012, 8:20am by Bill Connelly

Comments

1
by Brian Fremeau :: Thu, 10/11/2012 - 8:59pm

Good stuff, Bill. Does the frequency of each 3rd-and-X scenario factor into the adjustment? Or could a 1-for-1 on third-and-six really boost a team's adjusted third down rate?

2
by Bill Connelly :: Fri, 10/12/2012 - 10:07am

I did NOT use frequency for this version (wanted to make sure that everything carried equal weight), but I removed the max and the min, a la figure skating, then averaged the rest. So if you're 1-for-1 (or 0-for-1) in TWO categories, it could skew things, but for the most part there were not many examples of teams with N's of fewer than 3.

3
by Joseph :: Fri, 10/12/2012 - 3:42pm

Bill, you might want to develop a formula to do just that. For what you posted here, that's great--most teams have only played about 1/2 their games, so you're dealing with a partial sample size. It wouldn't surprise me that if you averaged all teams' leverage rate & adj. 3rd D%, you'd prob. get an overall ranking similar to the top 25 of most polls/websites--as both stats correlate very well to winning. I can't imagine a team doing well in both categories without being pretty good nationally.

4
by Bill Connelly :: Sat, 10/13/2012 - 8:20am

Absolutely. There are a lot of interesting directions we could take this. The first time around, though, I try to keep things as basic as possible.

4
by Bill Connelly :: Sat, 10/13/2012 - 8:20am

Absolutely. There are a lot of interesting directions we could take this. The first time around, though, I try to keep things as basic as possible.