## Team Pass Identities Through 10 Weeks (2015)

I’ve published the Game Scripts data from every game this year at the 2015 Game Scripts page, available here. What would it look like if we plotted Game Script score (on the X-Axis) against Pass Ratio (on the Y-Axis) for every game this year? Something like this:

(Note that this looks pretty similar to how it was through seven weeks last year, although the constant (i.e., league-average pass ratio) has increased by nearly a full percentage point.)

As you move from a more negative Game Script to a more positive one, the expected Pass Ratio decreases. But the relationship is not purely linear: in extreme cases, the Pass Ratios tend to move a bit more towards league average, and I think that trend is probably even stronger than it might appear on this graph (see, for example, the Saints in week ten). In any event, you can derive a best-fit polynomial equation from that data, which could give us an expected Pass Ratio.

The Patriots have been pass-happy in every game this season. There’s no better example of that than what New England did against Buffalo in week 2, which grades out as the most pass-happy game of the year. The Patriots had a Game Script of +9.8; put that number into our formula1, and you would “expect” a team to pass on 50.2% of plays given that Game Script. New England actually passed on 80.3% of all plays, meaning the Patriots passed on 30% more plays than expected. That’s the strongest pass identity of any team this year, and New England’s game against the Jets (GS = +0.6; Exp PR = 59.1%; Act PR = 86.4%; Diff = 27.3%) was the only other time when a team passed even 17% more than expected.

If we perform that calculation for every game this year, we can derive season grades. Let’s look at the Patriots line in the table below. In 9 games this year, New England has an average Game Script of +8.3, which happens to be the highest in the NFL (the table is fully sortable). Based on how each game has unfolded, New England would be expected to pass on just 52.1% of all plays if it was a team with a neutral pass identity; however, the Patriots have passed on 64.6%2 of all plays. That means the Patriots have passed an incredible 12.4% above expectation, the highest rate — by far — in the NFL this year. The table below lists that data for each team through week 10:

RkTeamGAvg GSExpActDiff
1NWE98.352.1%64.6%12.4%
2DET9-4.662.5%69.1%6.6%
3OAK91.657.7%63.4%5.7%
4DEN90.358.6%64.1%5.5%
5SDG9-3.762%66%4.1%
6NYG102.557.1%61%3.9%
7MIA9-2.661.8%65.4%3.6%
8JAX9-2.461%64.3%3.4%
9NOR10-0.159.3%62.5%3.2%
10HOU9-6.661.2%63.3%2.1%
11GNB92.457%59.1%2.1%
12KAN93.956%58%2%
13BAL9-2.862.1%64.1%2%
14IND9-3.362.1%63.5%1.5%
15CLE10-3.963%64.4%1.4%
16ARI97.154.1%55.4%1.3%
17PHI90.858.7%59.9%1.1%
18ATL90.360.1%60.5%0.5%
19WAS9-1.260.1%60%-0.1%
20CIN95.154.9%54.3%-0.6%
21PIT102.157.7%56.6%-1.1%
22TEN9-1.761.5%59.7%-1.8%
23NYJ91.058.1%55.4%-2.7%
24SEA92.757%53.3%-3.7%
25DAL90.059.5%55.4%-4%
26MIN92.856.8%51%-5.8%
27CAR94.755%48.8%-6.2%
28TAM9-2.559.7%53.3%-6.4%
29STL9-1.360.4%53.4%-7%
30SFO9-6.262.3%55.1%-7.2%
31CHI9-3.562.4%54.5%-7.9%
32BUF90.758.9%50.2%-8.7%
• The Patriots are always pass-happy under Tom Brady, but this is really extreme even for New England. New England has, by far, the best Game Script in the NFL, yet still has the 4th highest pass ratio this year! Consider that the three teams that have higher pass ratios all have strong negative average Game Scripts.
• Detroit and Oakland also grade out as really pass-happy. That’s not all that surprising for the Lions, who have tended to grade out as pass-happy just about every year. But the Raiders are wisely putting the ball into the hands of Derek Carr, and it’s working. On the other hand, look at how pass-happy the Broncos have been! The Denver running game has been very inconsistent this year, but given the struggles Peyton Manning has experienced, isn’t it crazy how pass-happy the Broncos have been?

Finally, let’s look at the same numbers for defense. The Jets have had an average Game Script of +1.0, but from the perspective of the defense, that means the Jets average Game Script is -1.0. That means you would expect Jets opponents to pass a little bit more often than average — here, it is 60.2% based on the average of the Jets nine games. In reality, on average, opponents have passed on 64.2%3 of plays, which is 4.0% more often than we would expect. So that means teams have been very pass-happy against the Jets. And, in what is no surprise, the only defense that teams have been more pass-happy against than the Jets? That would be Ryan’s Bills.

RkTeamGAvg GSExpActDiff
1BUF9-0.758.8%63.1%4.3%
2NYJ9-1.060.2%64.2%4%
3OAK9-1.660.1%64%3.9%
4PIT10-2.160.8%64.1%3.2%
5BAL92.856.9%60%3.1%
6ATL9-0.358.1%61%2.9%
7SFO96.254.9%57.3%2.4%
8IND93.356.4%57.9%1.5%
9CIN9-5.163.5%64.8%1.3%
10HOU96.655.6%56.8%1.2%
11KAN9-3.961.4%62.5%1.1%
12JAX92.457.5%58.4%0.9%
13NWE9-8.365.6%66.3%0.8%
14CHI93.556.1%56.7%0.6%
15SDG93.756.4%57%0.6%
16CAR9-4.763.6%63.7%0.1%
17NYG10-2.561.7%61.3%-0.4%
18PHI9-0.860%59.5%-0.4%
19DET94.655.6%55%-0.7%
20NOR100.159.1%58.1%-1.1%
21MIN9-2.861.8%60.6%-1.3%
22DEN9-0.359.9%58.5%-1.4%
23WAS91.258%56.5%-1.5%
24CLE103.955.4%53.8%-1.6%
25DAL90.059.5%57.7%-1.8%
26TAM92.558.1%56.2%-1.9%
27STL91.358.1%56.1%-2%
28MIA92.654.8%52.5%-2.3%
29ARI9-7.163.5%60.5%-3%
30TEN91.756.1%52.8%-3.3%
31GNB9-2.461.3%57.7%-3.6%
32SEA9-2.761.7%56%-5.7%

1. 0.000011 * GameScript^3 – 0.000077 GameScript^2 – 0.010004 * GameScript + 0.5965. []
2. Note that this is taking an average of the average pass ratio in each game, not the average of the season totals. As a result, this number will be different than the standard way of calculating a team’s pass ratio for the year. []
3. Again, taking the average of the averages in each game. []
• jtr

I’m skeptical of the cubic fit. Is there an actual strategic phenomenon that you think explains the reversal of the overall trend at each end? If not, then I think the cubic equation is less meaningful than a best-fit line, even if the correlation is a little better than the line.
I guess the curl downward at the bottom end is from a couple of roll-over-and-die games, like the Jimmy Clausen Bears mashing the punt button against Seattle and San Francisco deciding that getting blown out by the Cardinals is better than even trying to let Kaep win the game. It still doesn’t strike me as particularly meaningful to use the cubic fit just to account for those outliers, especially since the data looks like a line would fit it pretty well.

• I do think there is the “give up” effect with a largely negative Game Script. Look at the Saints game in week 10. On the other side, I wonder if teams choosing to switch from passing to running is more about time left than score. In other words, for teams to have a really large Game Script, they need to get a bit lead early, and it may take coaches longer to switch from running to passing if they get that early lead. For example, Arizona and Atlanta both still passed over 45% of the time in those two blowouts.

I’m less sure on the + side, but on the – side, there definitely has been a tendency to just stop passing in blowouts and give up. Not many, but a few of these games every year.

• jtr

My first thought on the + side was the Patriots, but they’ve only really had one of their signature run-up-the-score pass heavy blowouts (vs Washington). I’m sure that in 2007 the would own the top right of this graph and make the cubic effect stronger on the + side. It definitely looks like it’s give-up games driving the cubic effect more than running up the score this year.

• Josh Sanford

The teams with defenses in the top two positions (Bills and Jets) have each played the Patriots. Do their rankings change significantly if you remove the Patriot games (since we know from the top half of the article that the Patriots throw a lot, regardless of who they play)?

• Yeah, ideally, we would adjust these things for “strength of schedule”.

• Richie

” look at how pass-happy the Broncos have been!”

I have heard people say Peyton Manning is a selfish player. I don’t believe it’s true. But when I see things like this, it gives me pause.

I’m sure part of the issue is that QB’s just generally feel that they are better off audibling into a pass. And I wonder if Manning might even be trying to work out of his funk by passing more often.

But you have to wonder, if the Broncos would have just run the ball more heavily, if Manning might have avoided getting injured, and might have avoided 5 or 10 interceptions this year. And the Broncos might look like a team that dominates on defense, plays good enough on offense, and has a HOF QB just in case they ever get into trouble.

• Andrew Healy

Think it’s interesting that in addition to ground-and-pound Rex and the Panthers (who are driven by a QB who is their most important runner, too), the teams at the bottom have some of the worst (or least promising) strategic coaches in the league: Fox, Tomsula, Fisher, Lovie.