Writers of Pro Football Prospectus 2008

13 Dec 2013

What Denver's Loss Did to Playoff Odds

I had Mike Harris run a new batch of playoff odds reflecting last night's big San Diego upset of Denver, to see what it did to the AFC playoff picture. We still have Denver as the favorite for the No. 1 seed, but not by very much. Meanwhile, San Diego's chances of making the playoffs more than quintupled.

Remember that these odds are based on DVOA that accounts for last night's game but does not account for the future effect of recent injuries such as Rob Gronkowski's torn ACL.

Odds of No. 1 Seed in AFC
Team Before SD-DEN After SD-DEN
DEN 75.6% 41.9%
NE 18.0% 37.1%
KC 3.5% 10.9%
CIN 2.9% 10.2%


Odds of No. 6 Seed in AFC
Team Before SD-DEN After SD-DEN
MIA 51.6% 49.5%
BAL 34.1% 25.8%
NYJ 4.4% 2.5%
SD 3.2% 17.3%


Odds of Winning Super Bowl
Team Before SD-DEN After SD-DEN
SEA 23.8% 24.8%
DEN 20.6% 15.9%
NO 15.0% 14.6%
NE 11.2% 12.4%
CIN 8.4% 8.4%
CAR 8.3% 8.5%
SF 3.7% 3.8%
KC 3.5% 5.4%
PHI 2.2% 2.4%
DET 1.0% 0.9%
IND 0.6% 0.9%
CHI 0.6% 0.6%
DAL 0.4% 0.4%
MIA 0.3% 0.4%
ARI 0.2% 0.3%
SD 0.0% 0.2%

Posted by: Aaron Schatz on 13 Dec 2013

51 comments, Last at 21 Jan 2014, 10:01pm by Cheap UGGS


by Danish Denver-Fan :: Fri, 12/13/2013 - 12:41pm

Thanks for this. Very interesting.

by MC2 :: Fri, 12/13/2013 - 12:44pm

It would be interesting to see what happens if you were to run the playoff odds using New England's DVOA from the games that Gronkowski missed earlier this year.

Obviously, it wouldn't be a perfect estimate of the Patriots going forward, but I think it would be more accurate than using their total DVOA, and certainly more accurate than their weighted DVOA, which is based primarily on the way they have performed with Gronkowski.

by Perfundle :: Fri, 12/13/2013 - 12:59pm

Even that would overrate the Patriots' chances, because their defense was a lot healthier during that time.

by dryheat :: Fri, 12/13/2013 - 1:19pm

It would overrate their defense, but probably underrate their offense, since Amendola and Vereen were out concurrently.

by Juvenal (not verified) :: Fri, 12/13/2013 - 8:35pm

There are other teams with injuries. How many adjustments you going to make?

by MC2 :: Fri, 12/13/2013 - 8:44pm

First, I'm not suggesting they replace the official simulation. I just thought it would be interesting if they gave us both. If I recall, they did something similar a few years ago when McNabb got hurt.

Second, there may be other teams with injuries, but none of them (with the possible exception of Aaron Rodgers) seems to be as significant as this one. Even by conventional statistics, the Patriots have been a Top 5 offense with Gronkowski, and not even a Top 20 offense without him, so let's not pretend that this is the same thing as somebody missing their slot receiver or their fullback.

by Lyford :: Fri, 12/13/2013 - 10:23pm

If you look at the Patriots with (32 ppg) and without (20.833 ppg) Gronkowski, it looks like a devastating injury. And it certainly hurts - he's a great player, an impact player. But there are confounding factors that make the difference look bigger than it is. As already noted, in five of the six non-Gronk games, they were also missing Shane Vereen. In three of those, they were also missing Danny Amendola. One of those games was played in a monsoon in Foxborough (13 points in a win against the Jets); the first game that Amendola returned (without Gronk or Vereen) was played in a monsoon in Cincinnati (six points in a loss).

I would never argue that the loss of Gronkowski doesn't hurt - it clearly does. But if everyone else is healthy, I think they're still a better than average offensive team even without him.

by BigWoody (not verified) :: Fri, 12/13/2013 - 10:54pm

No, no, no I just can't do sarcasm font on this.

"I think they're still a better than average offensive team even without him."

Yes, I agree.

OK, no...the Pats are a way above average team without Gronk.

by MC2 :: Fri, 12/13/2013 - 11:19pm

You raise some good points, but even if your conclusion is correct, I would still argue that there's a huge difference between an above average offense and an elite offense, particularly given the Patriots' recent struggles on defense.

by MJK :: Sat, 12/14/2013 - 12:25am

Actually, I would expect some of the offensive improvement of late is related to the collapse of the defense, in addition to Gronk being back. In general, teams with strong defenses tend to play more conservatively on offense, whereas teams with week defenses tend to use more aggressive strategies. And because the NFL generally is more conservative than it should be, playing more aggressive is closer to optimal and offenses perform better.

The Patriots performance this season fits the pattern. Early in the year, when healthy, the Patriots defense was playing at a top 10 level, but the offense was only mediocre. Once Wilfork/Kelly/Mayo/Talib/Gregory/et al went down and teams started scoring up and down on them, it felt like McDaniels started calling plays more aggressively, and sure enough, scoring went up.

Of course, this also coincided with first Amendola, then Gronk, and then Vareen getting healthy, so it's obviously confounded. But that was the point of why correcting for injuries would be hard.

by MC2 :: Sat, 12/14/2013 - 1:44am

Of course, it's easier for a playcaller to be more aggressive when he has all of the weapons in his arsenal. It will be interesting to see if McDaniels continues to be so aggressive without his primary tool for creating mismatches at his disposal.

by RickD :: Sat, 12/14/2013 - 12:30am

The problem is that Gronkowski wasn't the only guy injured - Amendola and Vereen were both out, too. Collectively they constitute 3 of Brady's top 4 targets (along with Edelman). Now Gronk is gone but the other two are still around.

A prime example of why you have to be careful with statistics. It's important to isolate a variable before you start drawing conclusions about its importance.

by intel_chris (not verified) :: Fri, 12/13/2013 - 1:05pm

Thank you. It is cool that you can turn things like this around quickly. That's the nice part of having the program already written, you just run it with the new data. Still it is nice that y'all made the effort to do so.

On a related note, I'm sure I'm not the only one waiting for your analysis and insight on what this means on the teams relative strengths, why it happened, etc.

by dmstorm22 :: Fri, 12/13/2013 - 1:06pm

I could be wrong, but isn't San Diego's only hope of making the playoffs for them to finish with a better record than both Miami and Baltimore.

Miami has the head to head win, and both have multiple NFC loses (Carolina, Tampa Bay and New Orleans for Miami; Chicago and Green Bay - with Detroit still to come - for Baltimore), so the Chargers would lose conference record tiebreakers as well.

They'll rue those ridiculous losses to Tennessee and Washington, because that is a dangerous team the way Philip Rivers is playing.

by apk3000 :: Fri, 12/13/2013 - 1:41pm

Someone on a Chargers forum mentioned that Miami has the tiebreaker over SD and the tiebreaker if Miami, Baltimore and SD all finish with the same record. I don't know how a Baltimore/SD tie would work out.

by zlionsfan :: Fri, 12/13/2013 - 1:56pm

It would go to Baltimore based on better conference record: Baltimore is 6-4, San Diego is 4-6.

And I don't think that Chargers forum post is necessarily correct. The Dolphins have an extra conference game remaining. If they win that, then the three-way tiebreaker sorts out all three teams, with Miami a game ahead of Baltimore and three games ahead of San Diego. If they lose it, then the three-way tiebreaker only knocks out San Diego, and Baltimore's win over Miami gives the Ravens the final wild-card spot.

by jonnyblazin :: Fri, 12/13/2013 - 2:17pm

Hmmm..., I'm not an expert but I think there is one issue with this. If Miami goes 2-1 and the Charges go 2-0, we know their conference records would be 8-4 and 6-6. But the Ravens conference record if they go 2-1 could either be 8-4 or 7-5 depending on if the beat the the Lions. And I believe if they are 7-5 then the head-to-head matchup with the Dolphins won't matter because the tie breaker for 3 teams is conference record (since the Ravens didn't play the Chargers, head-to-head is not applicable when 3 teams are tied).

However if the Ravens are 8-4 in conference, then the 3-team tiebraker for conference record would eliminate SD, and then the tiebraker would revert back to head-to-head with 2 teams remaining and the Ravens would be in over Miami.

If Miami and the Ravens win out then the Ravens are obviously in, but I don't anticipate that happening.

If three teams are tied at 8-8... ugh, I'm not going to bother with that one just yet.

by Travis :: Fri, 12/13/2013 - 3:04pm

If Miami, Baltimore, and San Diego tie ...

1) If Baltimore beat or tied Detroit, Miami wins the tiebreaker based on having the best conference record among the three.

2) If Baltimore lost to Detroit, San Diego is eliminated on conference record, then Baltimore wins the tiebreaker based on a head-to-head win over Miami.

Whether the tie is at 8-8 or 9-7 doesn't matter.

by CBPodge :: Fri, 12/13/2013 - 4:12pm

From playing with the playoff machine the other day, if as many teams as possible finish 7-9 the tiebreaker order would start with the Raiders, then the Jags. Go Jags!

by jebmak :: Fri, 12/13/2013 - 9:03pm

Quirky Research is where I go for my tie-breaker needs:


(I have no affiliation, I just think that they are great.)

by Noah Arkadia :: Fri, 12/13/2013 - 11:15pm

Beautifully explained.

The man with no sig

by Bjorn Nittmo :: Fri, 12/13/2013 - 5:59pm

Not quite -- there's a third scenario for SD to make the playoffs (at #6) at 9-7 even if Miami is also 9-7. If the Jets somehow win out and end up 9-7 (they'd have to beat Carolina and Miami on the road, but stay with me), NYJ would finish ahead of Miami in the AFC East, eliminating Miami. If Baltimore is not also 9-7, Chargers would beat out the Jets on conference record. A slim chance, yes, but if Chargers and Jets both win out it doesn't matter to the Chargers what Miami does.

So Chargers need 1) to go 2-0 and 2) Baltimore to lose 2 and 3) either Miami lose 2 or Jets win 3. I'm still not sure I believe that adds up to 17.3%, which still seems too high to me.

As another commenter said, Baltimore would win two-way tiebreaker vs. Miami (head-to-head), but could get messed up by Chargers joining them in a 3-way tie because Miami might have the best conference record of the 3.

by jonnyblazin :: Fri, 12/13/2013 - 1:32pm

One thing I think these odds don't take into account is the Bengals history. The Ravens final game of the season at Cincy seems difficult, but I believe Marvin Lewis has rested his starters for the past three times he's made the playoffs, even when seeding was an issue.

by CBPodge :: Fri, 12/13/2013 - 4:13pm

I don't think he'd do that if a first round bye, of HFA through the playoffs was at stake though.

by vrao (not verified) :: Fri, 12/13/2013 - 2:00pm

The only way the Chargers make the playoffs is if they go 2-0 and the Dolphins and Ravens both go 1-2. The early season losses to the Texans, Titans, and Raiders killed the Chargers.

by boltsfromtheblue :: Fri, 12/13/2013 - 2:10pm

San Diego can also make the playoffs in a 3-way tie with Miami and New York.

by Travis :: Fri, 12/13/2013 - 3:45pm

Only if the Jets beat Miami and beat or tie Carolina.

by Revenge of the NURBS (not verified) :: Fri, 12/13/2013 - 2:14pm

What it did is bring a Colts/Broncos 1st round matchup into play. If that happens, I'll have to keep ESPN turned off for a week before and a week after.

by Bobman :: Fri, 12/13/2013 - 3:33pm


by titansjoe (not verified) :: Fri, 12/13/2013 - 4:11pm

you occasionally turn on espn? good lord, even the self-flagellating monks from the middle ages wouldn't go that far.

by panthersnbraves :: Fri, 12/13/2013 - 4:33pm

But what if you turned it on with no sound, just to look up the score on the ticker?

by Theo :: Fri, 12/13/2013 - 8:05pm

That's nonsense, no one ever has used the ticker for any information whatsoever.
It's just there to let 2 guys have a job.

by my name is my password (not verified) :: Fri, 12/13/2013 - 2:20pm

Also: Jacksonville has the tie-break over all of the other possible 6th seeds.

by Travis :: Fri, 12/13/2013 - 3:05pm

Not Oakland. Oakland wins any four-way tiebreaker at 7-9.

by RickD :: Fri, 12/13/2013 - 2:59pm

I'm interested in how DVOA will react to the #32 pass defense handling the #1 passing offense.

Opponent adjustments should be fascinating. I'm picturing a computer going up in smoke, like they used to do in old movies when the math got really hard.

by Bobman :: Fri, 12/13/2013 - 3:35pm

Awesome image. That's when computers filled entire rooms and needed their own powerplant to run them. At full power, the entire remainder of the city dimmed. All the flashing lights alone took 1,000 amps of 220 current.

Well... 220, 221, whatever it takes.

by CM (not verified) :: Fri, 12/13/2013 - 8:20pm

220, 221, whatever it takes.

One of my all-time favorite movies!

by CBPodge :: Fri, 12/13/2013 - 4:15pm

I'm picturing this as the moment computers become self aware, as the DVOA calculator just goes "nah, that can't be right."

by panthersnbraves :: Fri, 12/13/2013 - 5:01pm

hurray! the Saint SB odds went down and the Panthers went up!

That's a direct result of Carolina beating San Diego and that affects the adjust...wait, they didn't play? So it must be about the New Orleans/Denver game back in the beginning of the ... whuh? They didn't play either? So because DEN played NE who played both CAR and NO? really?

Direct - DEN/SD - definitely
Rest of division - sure
Common Opponents (NFCE/AFCS) - makes sense
Non-common Conf Opponents (AFCN/AFCE) - OK, but it would seem to not be very much
The rest of the NFL, because the NFCS played the AFCE, etc. - {pewwww} mind blown

by Perfundle :: Fri, 12/13/2013 - 6:17pm

Since there were no games between the two divisions, I'm sure the slight change in the odds is from random variations from the simulation itself.

by Sporran :: Fri, 12/13/2013 - 8:35pm

I think it's because Denver's odds of making the SB went down -- and therefore made it easier for an NFC team to win.

by RickD :: Sat, 12/14/2013 - 12:36am

Also, the strength estimate for the Broncos (i.e., their DVOA) will surely go down. After the loss to the Chargers, the Broncos look more beatable.

by panthersnbraves :: Sat, 12/14/2013 - 1:51am

OK - so the Panthers go up - why did the Saints go down?

by Alternator :: Sat, 12/14/2013 - 2:13am

If the simulation includes head-to-head factoring for specific matchups, it could be giving the Panthers better odds to beat the Patriots than it gives the Saints to beat the Patriots.

by dcaslin :: Fri, 12/13/2013 - 7:47pm

If Baltimore has a better shot at the playoffs than SD right now, where are Baltimore's SB odds on the last table?

by BigWoody (not verified) :: Fri, 12/13/2013 - 10:36pm

So...Denver must have taken a big hit in DVOA in last night's game. Carolina with an average game vs the Jets should take 2nd in weighted DVOA and the Saints could take 3rd with a good game vs the Rams, dropping the Broncos to 4th. What is THAT going to do to the playoff odds report?

Baited breath til Sunday and more baited breath Tuesday afternoon. Come on Aaron, give us a clue! Anybody???

by willybhu :: Sat, 12/14/2013 - 12:39am

Why did NO SB odds go down, while every other NFC team's went up (except for DET, but they only went down .8%). And what's the rationale behind CIN's SB odds remaining unchanged? Is it because they're pretty much locked into that wild card game?

by td (not verified) :: Sun, 12/15/2013 - 6:11pm

so I guess now there'll be a post about what the new England loss does to the playoff odds?

by Bjorn Nittmo :: Mon, 12/16/2013 - 4:00pm

Certainly sends San Diego's odds back where they started the week, if not below.

by Wholesale Metal Mulisha Hats (not verified) :: Tue, 01/07/2014 - 5:01pm

Through any amazing design connected to points a person safe the best A+ for the purpose of effort. Exactly the place precisely people baffled us all wound away becoming in the majority of the truth of the matter. Because consumers state, particulars cause or quite possibly split the very debate.. And this could not are more proper this particular. With in any event, enable myself give out just what exactly do complete the task. The material (parts connected to it) is usually convincing which happens to be actually perhaps for what reason I'm making so as to remark. I do not necessarily helps it be a common routine to carry out which. moment, despite the reality that I will easily enough spot any jumps within make a person cause, I'm possibly not assured connected to precisely how a person manage to link an individual's opinions the fact that inturn help the finalized outcome. For the time being I'll, without doubt register for your own personal concern nevertheless have faith in from now on you extremely link your own private details better.

by Cheap UGGS (not verified) :: Tue, 01/21/2014 - 10:01pm

Hmmm it seems like your first short review hasn