Defensive Stats: Meaningless and Otherwise
Back in October, I analyzed the Seattle defense's 3rd down conversion rate, which was then just 12th-best in the league at 37%.
I concluded that it was largely irrelevant as both an indicator of past success and as a predictor of future success. The Seahawks were preventing big plays, forcing occasional turnovers, and had no problem letting an opponent gain 10 or 20 yards before forcing them to punt.
Since then, Seattle's defense has gone on to lead the league in points, yards, and takeaways. Their third-down conversion rate did not improve significantly, finishing at 35% (10th in the league). Proving me right.
Here's a big mess of defensive stats to strain your eyes (summary below):
|Team|| yards/ |
| Turnover |
| Plays/ |
| Yards/ |
| Yds |
| Average |
| Minutes/ |
| Time |
| Points/ |
| Pts |
| 1st Down/ |
|San Francisco 49ers||5||13.1||5.5||26.5||8||25.3||2.47||12||1.41||4||1.49||7|
|Kansas City Chiefs||5.5||15.7||5.4||28.8||18||23.2||2.32||3||1.49||5||1.49||6|
|New York Giants||4.9||13||5.6||26.6||9||31.7||2.50||14||1.59||8||1.64||15|
|New Orleans Saints||5.2||10.5||5.4||27||11||26.5||2.42||9||1.6||9||1.51||8|
|New England Patriots||5.3||13.3||5.8||30.3||22||24.1||2.53||16||1.62||10||1.72||19|
|St. Louis Rams||5.4||14.7||5.6||30||21||26.8||2.68||28||1.78||16||1.83||25|
|New York Jets||5.1||7.3||5.7||28||14||31.1||2.53||17||1.79||17||1.53||11|
|San Diego Chargers||6.1||9.6||6||35.2||31||23.8||2.65||25||1.98||23||1.97||30|
|Tampa Bay Buccaneers||5.5||15||5.6||29.8||20||30.9||2.63||23||1.99||24||1.73||21|
|Green Bay Packers||5.9||12.2||5.8||33.1||28||29||2.67||27||2.19||28||1.83||26|
The real gem here is Seattle's league-best 23.4 yards per drive (if you're curious about Buffalo's 2nd-place ranking in this stat, be aware that they have the league's 4th-best defense in total DVOA). That's a number that is in no way padded or compensated for by playing short games with fewer drives or having a good red zone defense. In fact, the quality of special teams (giving a 6th-best average starting field position at the 25.8 yard line) gives the opposition more open field, and should make that number even worse.
Don't underestimate the disparity between yards/drive and first downs/drive. It's HUGE. If we were to go way down the rabbit hole and calculate the ratio of the two, Seattle's average of 15.43 yards/first down would be 2.7 standard deviations below league average, and almost an entire standard deviation below the next-lowest team.
To put that in layman's terms, when the opposition is facing 2nd and 2, the Seahawk defense's attitude is "Go ahead and have your little 2-yard run. You aren't getting a fresh set of downs. We are getting a fresh 10 yards."
Offensive 3rd down conversions
On the other side of the coin, Seattle's offensive third down conversion rate was 37% on the regular season, 17th in the league. If "third down differential" were the king of predictive stats, the Seahawks' +2% would make them an 8-8 or, at best, 9-7 team.
On the other hand, 3rd-down conversion rate might be a bunch of hooey.
Consider the following consecutive possessions from the NFC Championship game:
Drive start: Own 38, 2:49 3rd quarter
1st and 10: Lynch for 5
2nd and 5: Lynch for 6
1st and 10: incomplete
2nd & 10: pass for 13 yards
1st & 10: Turbin for 4
2nd & 6: Intentional grounding for -16
3rd & 22: pass for 15 yards
4th & 7: pass for 35 yards, Touchdown
8 plays, 62 yards, 4:03, 7 points, 3 first downs
3rd down conversion rate: 0/1
San Francisco 49ers:
Drives start: Own 11, 13:37 4th quarter
1st & 10: Kaepernick pass incomplete
2nd & 10: pass for 5 yards
3rd & 5: pass for 9 yards
1st & 10: Hunter for 3 yards
2nd & 7: Kaepernick for 6 yards
[delay of game]
3rd & 6: Strip-sack, turnover
6 plays, -5 yards, 3:37, 0 points, 1 first down
3rd down conversion rate: 1/1
It's not unusual to find a statistic like turnovers or rushing yards that fails to correlate with scoring points and/or winning games in specific instances. But in those cases, a few outliers do not necessarily change the underlying value of the stat line as a predictor of future performance, nor its value in measuring a certain strategic advantage such as sustaining drives.
The problem with 3rd down conversion percentage is that it doesn't even measure what it's supposed to measure. Teams which turn the ball over on 2nd and long end out with better third down percentages. Teams that earn a first down on 1st or 2nd down end out with worse conversion rates.
A child could do better. But as no child is available, I will step up to the
plate line of scrimmage and present you with the True Conversion Rate.
It's simple and intuitive. When a team has a first down (either by change of possession or by a previously-earned first down), that's one opportunity to get a first down (note that scoring a TD counts as first in this case).
Then they either succeed or fail.
|Team||Drives|| Non-TD |
| Total |
| First |
| True |
| TCR |
| Made |
| True CR |
| True CR
|San Diego Chargers||167||332||499||373||0.75||2||34||0.80||2|
|New Orleans Saints||181||310||491||359||0.73||3||24||0.77||3|
|Green Bay Packers||186||309||495||351||0.71||5||33||0.76||5|
|New England Patriots||201||334||535||378||0.71||7||38||0.76||4|
|Kansas City Chiefs||197||282||479||323||0.67||16||22||0.71||20|
|San Francisco 49ers||188||247||435||286||0.66||21||32||0.71||17|
|St. Louis Rams||184||250||434||282||0.65||22||26||0.69||22|
|Tampa Bay Buccaneers||188||241||429||269||0.63||27||23||0.66||30|
|New York Jets||196||255||451||280||0.62||29||33||0.67||27|
|New York Giants||204||251||455||280||0.62||31||23||0.65||31|
(Ranks are accurate despite apparent ties, as they include more decimal places than are displayed.)
So now we have a far, far more informative measure of a team's ability to sutsain drives by getting first downs. The Seahawks' True Conversion Rate ranks 12th, about 1.2% better than league average, and good enough by itself to qualify as a playoff-caliber (but not elite) offense. Just for kicks, I re-calculated the TCR excluding made field goals. Red zone efficiency is important, but it is arguably a different measure. It's worth noting that all four of the teams with an average starting field position better than their own 30 (Kansas City, San Francisco, Seattle, and New England) had better conversion rankings (True CR EFG) when excluding field goals. Evidence enough that they had fewer opportunities to rack up first downs before hitting the compressed red zone.
Of course, if we were trying to use 3rd down conversion rates as an indicator of a team's ability to stay on the field, advance the ball, and score points, why didn't we just measure those things?
You wear down an opposing defense by running more plays. You give your defense a rest by taking more time. I combined play count and possession time per drive into the Minutes X Plays Index.
I then combined that number with the points/drive (using a geometric mean) to measure the Total possession value. Results:
|Team|| Plays/ |
| Yards/ |
| Minutes/ |
| Minutes X |
| Minutes- |
| Points/ |
| Pts |
| Total possession |
| Turnover |
|San Diego Chargers||6.6||37.7||3.22||4.61||1||2.33||2||3.28||11.4||27.6|
|New Orleans Saints||6.1||35.3||2.88||4.19||3||2.29||3||3.10||8.3||27.8|
|Green Bay Packers||6||34.5||2.65||3.99||5||2.15||6||2.93||12.4||28.5|
|New England Patriots||5.9||30.6||2.42||3.78||14||2.1||9||2.82||10||30.7|
|San Francisco 49ers||5.3||27.5||2.60||3.71||17||1.96||12||2.70||9||31.9|
|Kansas City Chiefs||5.4||27.3||2.52||3.69||20||1.79||19||2.57||9.1||33.9|
|St. Louis Rams||5.4||26.5||2.53||3.70||19||1.64||21||2.46||11.4||26.3|
|Tampa Bay Buccaneers||5.4||23.4||2.52||3.69||21||1.41||27||2.28||9.6||26.8|
|New York Jets||5.4||25.8||2.47||3.65||24||1.4||29||2.26||14.8||26|
|New York Giants||5||23.9||2.30||3.39||31||1.33||30||2.12||20.1||27.3|
The index doesn't account for yards, turnovers, or starting field position (the latter of which would increase points, but make every other measurement comparatively worse), so those are included separately.
Looking at the table, I think that a team which outranks another across the board (Seattle vs. New Englad, for example) almost certainly has the better offense. But I suspect that the "Total possession value index" exaggerates the value of time of possession. Incomplete passes stop the clock, but they still give your defense just as much rest time as running plays which keep it running. Time of possession probably overrates the Seahawks-- they're just 17th in plays/drive. On the other hand, they are 11th in yards/drive (again with the caveat about starting field position degrading that stat), 6th in fewest turnovers per drive, and 7th in points per drive. And points, really, are what matters.
Our own John Fraley observed that Seattle is converting just 25% of red zone opportunities into touchdowns in the playoffs. Worrisome? Yes. Good sample size? No.
According to TeamRankings.com, Seattle's regular-season red zone percentage is 53.23%, ranking 14th and below the league average of 55.03%. Good sample size? Yes. Worrisome? Well, now...
Unlike the supremely ridiculous third-down conversion rate, red zone percentage at least measures what it's supposed to measure. There is no doubt that the higher a team's percentage, the better the prospects for its offense.
On the other hand, there are edge factors which influence this. Seattle likes to score from outside the red zone, which doesn't excuse the poor red zone performance, but indicates a different type of offensive potential, and a possibility that red zone percentage is not a complete predictor of future performance.
And, as I have continually said throughout this article, Seattle frequently enjoys good field position. When your defense or punt returner gives you the ball on the opponent's 20-yard-line, you should be more likely to score a touchdown than if you took possession on the opposite side of the field. But you may be less likely to score a touchdown than if you'd taken possession on the opposite side of the field and then moved 60 yards, wearing down the opposition and getting into an offensive rhythm. Furthermore, these situations often accompany a lead or heavily defensive game, prompting the offense to play conservatively.
But narrative is just smoke and mirrors. Without the mirrors. What do the numbers say? Do such numbers even exist?
I used Pro-Football-Reference's Drive Finder to look at all drives which started inside a team's own 40-yard-line, thus excluding anything that might be considered "good" field position. Results:
|Team||Drives||Touchdowns||TD %||FG Att||FGA%|| TD:FGA |
| TD:FGA |
When facing 61+ yards of green, the Seahawks are 6th in the league (22.4%) in scoring touchdowns. Which is pretty darned good. The ratio of touchdowns to field goal attempts is 1.43, a far more mediocre 12th in the league, so there remains, certainly, some red zone issue. But looking at the big picture, we can say that Seattle's red zone problem is definitely not a lack of touchdowns. It's simply an excess of field goals.
Seattle's Offense in the Playoffs
24 non-TD first downs
6 field goals
28 total first downs
28/45 = True Conversion Rate 62.2% (comparable to 30th in the regular season, which is poor; league-average was 67.3%)
True Conversion Rate excluding field goals = 71.8% (league-average in regular season was 71.4%)
With league averages (regular season) given in brackets:
5.6 plays/drive [5.6]
25.8 yds/drive [29.0]
2:53 time/drive [2:31]
4.76% turnovers/drive [12.5%]
2.19 points/drive [1.81]
Average Start Own 36.3 [28.3]
The regular-season trend of good starting field position degrading the statistical measures of yardage, plays, and first downs, while simultaneously giving a boost to points, is even more pronounced in the post-season. With that, and the quality of opposition faced, I see no evidence of a downtrend.
And now I have some traveling to do, so you'll have to excuse lack of a clever closing statement.
- Super Bowl 48, Seahawks vs. Broncos: Controlling the Flats
- Super Bowl 2014: What to expect when you're expecting to win - a look at 'Enver.
- EXCLUSIVE: Top secret documents suggest Earl Thomas may not be human
- Super Bowl 48: Rain City Redemption Part 3.2
- Seahawks vs Broncos, Super Bowl XLVIII: 4 MORE stats that say Seattle is about to get their championship groove on