## Great tight ends vs. Great pass rushing teams

Posted by Chase Stuart on January 25, 2008

When a great pass rushing team -- like the 2007 Giants (New York led the NFL in sacks) -- plays a team with a great tight end -- like the 2007 Cowboys (Jason Witten led all TEs in fantasy points), does something unexpected happen? That is, assuming the teams are of equal strength, does one team win more than 50% of the time?

Here's the theory: when you face a great pass rushing team, you generally need to keep your tight end in to block more often than usual. When the Cowboys face the Giants and Michael Strahan and Osi Umenyiora (23 combined sacks this season), Jason Witten will be called on to help his tackles pass protect more often than when Dallas plays Buffalo (starting DEs had 9 sacks combined; the team ranked 30th with just 26 sacks on the season). So, perhaps when these teams play, the great pass rushing teams win a disproportionate number of times (relative to the differences in team strength of the two teams) because the team with a great tight end is forced to use one of its best weapons in a suboptimal way. This is just a theory, of course -- but it *is* testable.

How so? It's complicated, so let me break it down into steps.

I looked at all games from 1970-2006. Do you remember the Simpe Rating System? Using that, we have a rating for each team that we can use to determine team strengths. I used two inputs -- home/road/neutral (with home being a 1, road being a 0, and neutral being a 0.5 for Super Bowl games) and the difference in the SRS of the two teams. Recall that a team with an SRS of 7.0 is simply seven points better than average, and a team with an SRS of 3.0 is three points per game better than average. If those two teams played, the SRS difference would be four points. If we run a regression analysis with win/loss/tie (1, 0, 0.5) as the outputs, here's the formula that best fits the 37 years worth of data:

Winning % = 0.42 + 0.157 * (home/road variable) + 0.0275 * (SRS difference)

What does that mean? If the team with a +7.0 SRS rating plays at home against a team with a +3.0 rating, the home team should win about 69% of the time. If the game was played at the weaker opponent's home, the better team should win about 53% of the time. Neither of those numbers should surprise you, as both fit in very well with our perception of team strengths and home/road variables. (Of course, like any linear system, this regression analysis will have its drawbacks the more extreme the opponents).

So we have some numbers that jive well with our gut feeling. Now what? We need to rank teams by pass rushing ability and tight end strength. You can use SRS numbers to rank pass rushing teams, too; in fact, we've done so. Here are the top teams at sacking the QB after accounting for strength of schedule and era:

nyj 1981 2.07 clt 1975 1.98 sfo 1976 1.92 nwe 1977 1.88 chi 1987 1.81 min 1989 1.80 was 1973 1.63 mia 1982 1.61 chi 1984 1.60 rai 1984 1.52

That means the '81 Jets would sack the opposing quarterback about two more times than the average defense would, if the average defense played the '81 Jets' schedule. That +2.07 happens to be the highest ratio since the merger.

How about ranking the tight ends? That's a bit complicated, because we need to adjust by era. I divided each tight end's receiving yards on the season by a special variable: the total number of receiving yards by all tight ends in that season divided by the total number of team games played in the NFL that year. Here are the top ten tight ends when using that metric:

id yr tm recyd ydratio CoatBe00 1994 nwe 1174 36.61 WinsKe00 1980 sdg 1290 36.50 GonzTo00 2000 kan 1203 35.68 SharSh00 1996 den 1062 34.64 SharSh00 1997 den 1107 32.90 GonzTo00 2004 kan 1258 32.57 SharSh00 1994 den 1010 31.50 ChriTo00 1986 rai 1153 31.46 ChriTo00 1983 rai 1247 30.08 SharSh00 1993 den 995 29.58

Here's what I did next. I looked at the top 100 teams at sacking the QB (measured by SRS) and the top 100 individual tight end seasons (as measured by the ydratio described above) over the 37-year period. It turns out those two teams ended up playing each other 152 times. So when the '81 Jets (+2.07 sack SRS, rank #1) played the '81 Bengals (with Dan Ross, whose 24.99 yard ratio ranked 40th), that game made the list. It turns out the '81 Jets won that game, despite being an inferior team. So maybe the theory is correct?

There are a bunch of ways to test the theory, but here's the simplest. I used our previous regression formula to predict how many wins you would expect the great tight end teams to win of those 152 matchups. It turns out the answer is 62.5 out of 152, which is 41.1 percent of the time. The reason? The great tight end teams have an average SRS rating of +1.3, while the great pass rushing teams had an average SRS rating of +4.6 (the teams each played 76 games at home). Not surprisingly, teams that lead the league in sacks win more games than teams that have the most productive tight ends in the league.

In reality, though, the great tight end team won 38.8% of the games -- 59 out of 62.5 times. That's not a really big difference, but it does trend in line with our theory. We could argue that teams with great tight ends win about 2.23% less often than they would against a team of equal strength, but without a great pass rush.

There are other ways to test this. Instead of using wins and losses (which are somewhat arbitrary), let's use points scored and points allowed. The regression result from the big data produces the following equation:

-2.79 + 5.58 * (home/road variable) + 1 * (SRS difference)

This formula is a bit more intuitive than the last. It merely states that home field advantage is worth about 2.8 points (which we know), and that a point difference in SRS is worth...a point difference in the point totals in a game. Basic stuff here.

Over the course of the 152 games, we would expect the great pass rushing teams to outscore the strong TE teams by 489 points. Note: this matches up perfectly, of course, with our other data. The SRS difference between the two teams was, on average, 3.216. And since the teams split the home/road games, 3.216 x 152 is equal to 489 points.

What * actually* happened? The great tight end teams lost by 584 points, an average of 3.84 points per game. So we could argue that the great tight end team's offense is inhibited by about two-thirds of a point a game when playing a great pass rushing defense.

I think there's enough data involved to be reasonably confident in the results, despite them being relatively small. Teams can gameplan around great pass rushing ends, of course. They might throw short passes to the tight end to nullify the pass rush, and these teams (with great pass catching tight ends) are particularly adept to do that. But they do seem to be at an ever so slight disadvantage when facing strong pass rushing opponents.

I did one last test -- I used only teams whose pass rush SRS score was over 1.00 (the very, very best pass rushing teams) and whose TEs ratio was over 25.00 (the very, very best tight ends). Such teams faced off just 28 times, fewer than once a season. The great TE team was expected to win 13.1 of the 28 games, and it won 13 of the 28 games. However, it should have been outscored by "just" 0.52 points per game; it was actually outscored 3.43 points per game. That's a really small sample size, of course; but it's worth noting that the very best pass rushing teams have neutralized the top tight ends by about three more points per game than you'd predict. The best example of this would be the '83 Raiders game against the Cardinals. The Raider's Todd Christensen had a 92/1247/12 season, one of the best ever by a tight end. The Cardinals led the NFL with 59 sacks that season, boasting three lineman with double digit sacks. The Cardinals went 8-7-1 while Los Angeles won the Super Bowl. But St. Louis won in L.A. 34-24 when the two teams met head to head, the Raiders last loss of the year.

If the theory is that good TE's are forced to block more often against good sacking teams, then should you exclude games where the TE ended up catching 5+ passes and/or 50+ yards? In those games we could deduce that the team was NOT keeping the TE to block more often.

Separate topic, and statistically insignificant, but shouldn't games such as: SF vs ARI at Mexico City, MIA vs NYG at London and MIA vs SD at Phoenix be counted as "neutral field" games?

I'm not so sure the results are that significant. A 41.1% coin flipped 152 times has a standard deviation of 6.07 or so. A difference of 3.5 games isn't that big a deal.

On the point variance side, you can estimate the relevant SD (SD of result based on simple ranking difference) by taking the SD of the score results for the special games. Since it's a 152 game sample, the result is about 7.7/result_spread_SD SD's off expectation. I'd bet that this is also under 1 SD of significance.

Maybe you have different thresholds for being reasonable confident, but you'd need about 12 times as many games to make the difference in win % reach 2 SD.

Dan R,

Great points about the significance. I suspect it will be hard to measure how statistically significant the results are because I don't think the actual impact is very severe. I *do* think teams with great tight ends are at a slight disadvantage against teams with an SRS of 0.0 and a great pass rush vis-a-vis teams with an SRS of 0.0 and an average pass rush. But I don't think it's anything near a large disadvantage.

Hey Richie,

Part of the problem is that we don't have individual game data going back before '95. I should have mentioned that in the blog post. I think having the individual game data would help, but wouldn't be perfect. I'd suspect that some of these teams would still send their star TE out on pass patterns, but would play 2-TE, 2-WR sets more often than their base 1-TE, 3-WR sets. And merely having their backup TE instead of slot WR on the field would be a harmful effect that wouldn't necessarily show up in the main TE's stats.

The harmful effect of a 2 TE set sounds fairly minor, if you have a great catching TE. Since the great pass rushing teams can disrupt any passing offense, and the effect is so small, is there any reason to suspect that it's different from the overall effect of pass rusher specialization vs passing specialization?

The test would be to compare the results of the best pass rush teams against offenses that concentrate on the pass vs their results against run-centered offenses.

If anything, what might matter more is quarterback speed.

I think where all of this matters is a small subset of plays of an entire game: 3rd & 5+ yds to go (good TE team on O, good pass-rush team on D). Depending on each game, this may only occur 2-3 times per game. Why this subset? 1) Obvious passing situation. 2) The great TE (Witten, Gates, Gonzalez, Sharpe, etc.) is a real threat in the passing game--esp. finding a hole in the zone just past the 1st down marker, but may be called upon to block vs. running a pattern. The alternate mentioned above is a 2nd TE staying in to block while slot/3rd receiver stays on the bench.

It would be extremely difficult to go back through the games between these types of teams since '95 (where we have PbP data) and check what happened on these specific plays. (Unless Chase REALLY has nothing else to do)

Another possible way to look at this is to look at tight ends (with a lot of starts) at the bottom of the rankings, who are more likely to be in the lineup for their blocking than their receiving. How do those teams fare against strong pass rushes?

I think #7 post is right on. How do poor receiving TE teams do against good pass rushing, and also how do good TEs do against a poor pass rush?

Do good blocking TEs stop a pass rush? Do poor blocking TEs do better against a poor pass rush?

The study of good vs good is interesting. The result shows very little impact.