Pats Bounceback After Loss? A Statistical Analysis

[Nikolai]

I don't post a lot of threads like this, but I wanted to kinda throw out something out there in the lead up to one of the biggest regular season games we've had in a long while.

We've seen all week, and I've repeated often in the office the old mantra, "the Pats are gonna bounce back from this loss like we always do. I'm not worried. We're gonna murder the Steelers."

But is it true? How do the Pats bounce back after the kind of loss we had against the Dolphins?

I'm using stats from the 2010 onward and decided on that season for three reasons:

1. That was the season we drafted Gronk.
2. We had significant turnover after 2009
3. And it was Brady's first fully healthy season since his 2008 injury.

It also feels like the most natural dividing line between the "old dynasty" and the "new dynasty". On to the numbers:

I calculated how bad a loss was by comparing the Pats points for and points against, against the season averages at the time of the game. I added the deviation together and came up with a number. In the case of the Pats-Dolphins, the Pats came in scoring an average of 29 points a game and giving up 18.6 points a game. Take the game's score and compare:

20-29 = -9
18.6-27 = -8.4
TOTAL = -17.4

Since the start of the 2010 season, there have been 12 other games where our deviation was below -17.4. Given that the Pats have lost 26 games since 2010, that pretty much puts this past game right in the middle of the pack as far as our "worst losses" go. Just to satisfy some curiosity, the three worst losses according to this method are:

2014 Kansas City: -32.7
2010 New York Jets: -28.0
2010 Cleveland Browns: -27.3

But what has happened in the weeks following these losses?

Pats Overall Record: 9-3
Pats Road Record: 7-2
Average Score: 31-20
Average Road Score: 29-17
Average Points For Improvement: 10.8
Average Points Against Improvement: -11.6

Pats Record Against 0.500 or better: 4-1
Average Points For Improvement against 0.500 or better: 16.3
Average Points Against Improvement against 0.500 or better: -12.5

That's an average improvement of 22.4 (or even 28.8 against 0.500 or better!) deviation over the losing game, which would put the Pats at a projected +5 deviation for this next game. The Pats have never lost a game since 2010 when at a +5 or better deviation. As you can see, 9 of the 12 games following a tough loss have come on the road, and the Pats have responded very well with a 7-2 record, including a 4-1 against teams that are 0.500 or better.

Strictly going by this, it would portend a 36-14 victory. Of course, I think that's a bit crazy, even if I suspect the Pats could well curb stomp the Steelers this weekend. But, given the history, I wouldn't say now is the time to panic.

Then again, as BB would say, "History doesn't help us."

And I guess that's why they play the game on the field.

 

[oldrover]

Strictly going by this, it would portend a 36-14 victory.

https://imgflip.com/memegenerator

 

[Bruins29]

Great stuff Nikoli. Great to see the work and thought put in to this type of thing (As opposed to the garbage we get from trolls and fake fans). Pats typically do bounce back from a loss pretty well and seemingly even better from a bad loss.

 

[cmasspatsfan]

My memory was they bounce back well but my memory ain't what is used to be. Thanks for doing the leg work, I was hoping to hear something like this on the radio while I was working but didn' hear it brought up.

 

[Nikolai]

My memory was they bounce back well but my memory ain't what is used to be. Thanks for doing the leg work, I was hoping to hear something like this on the radio while I was working but didn' hear it brought up.

That was my memory as well, but I wondered if maybe it was one of those old tales that didn't pass the "fact test". I was surprised how many of the worst losses were followed by a road game, and then to find the Pats won anyway. Not bad.

 

[WaterfallJumper]

Very cool breakdown. I actually think your projected score sounds about right. Didn't we beat them 36 - 17 last year? I think I recall that correctly. I wouldn't be surprised if each team is within one score of that figure.