NFL Playoffs & Seeding Probability

Briz

Member
SoSH Member
Jan 23, 2011
1,509
NH
Updated odds from Football outsiders:  http://www.footballoutsiders.com/stats/playoffodds
 
Our beloved Pats are 99.9% to make the playoffs, 99% to win the division, most likely to get the #2 seed (43%) but about 1 in 5 for the #1 seed (18%).  They would get a bye 61% of the time upon simulating the remaining season 20,000 times.  52% they appear the AFCCG.  They make the SB 26% of the time and win it 12% of the time. 
 
 
Conference Summary (Team, % to get seed;)
 
AFC
1)      Den, 76%; NE, 18%
2)      NE, 43%; Cinci, 38%; Den 17%
3)      Cinci, 55%; NE, 34%; Indy, 6%
4)      Indy, 94%; NE, 4%; Cinci, 2%
5)      KC, 95%; Den , 5%; Everyone else 0.3% with Mia @ 0.2%
6)      Mia, 52%; Balt, 34%; NYJ, 4%; SD, 3%;
 
NFC
1)      Sea, 96%; NO, 3%
2)      NO, 70%; Car 23%
3)      Phi, 5%; Det, 18%; Chi, 15%; Dal, 11%
4)      Det, 49%; Dal, 25%; Chi, 11%; Phi, 10%
5)      SF, 44%; Car, 39%; NO, 12%; Ari, 2%;
6)      SF, 44%; Car 31%; NO, 12%; Ari 8%
 
Playoff odds (did not include teams with < 1% to win SB)
 
Team       Conf App          Conf Win        SB Win
SEA             65%                   40%            24%
DEN             70%                   42%            21%
NO               57%                   27%            15%
NE                52%                   26%            11%
CIN               42%                   19%             8%
CAR             32%                   15%             8%
SF                17%                    7%              4%
KC                21%                     8%             3%
PHI               12%                     4%              2%
DET              8%                      3%              1%
 

DJnVa

Dorito Dawg
SoSH Member
Dec 16, 2010
53,837
FL4WL3SS said:
How are those playoff odds calculated?
 
 
The playoff odds report plays out the season 20,000 times. A random draw assigns each team a win or loss for each game. The probability that a team will be given a win is based on an equation which considers the current Weighted DVOA ratings of the two teams as well as home-field advantage. (DVOA ratings are explained here.) Mean Wins is the average number of wins for this team over the 20,000 simulations. Odds are given for winning each playoff seed, as well as for winning the division (DIV), earning a bye week (BYE), or winning a wild card (WC). TOT represents odds of making the playoffs in any fashion.
 
Please note that the playoff odds report does not adjust for the future impact of recent injuries. Also, this week's report is only 20,000 runs instead of 50,000 due to a technical issue we hope to clear up soon.
 

DanoooME

above replacement level
SoSH Member
Mar 16, 2008
19,831
Henderson, NV
Updating the NFC this week
 
Seattle 11-2
New Orleans 10-3
Philadelphia 8-5 
Detroit 7-6 (wins H2H tiebreaker over Chicago)
Carolina 9-4 
San Francisco 9-4
 
Arizona 8-5
Chicago 7-6
Dallas 7-6
Green Bay 6-6-1
 
Everybody else is officially eliminated.
 
Games remaining for potential top seeds (with total opposition winning percentage in parentheses):
 
Seattle - @NYG, Ariz, StL (.462)
New Orleans - @StL, @Car, vs TB (.436)
Carolina - NYJ, NO, @Atl (.487)
San Francisco - @TB, Atl, @Ariz (.385)
 
Games remaining for the fascinating NFC North race (with total opposition winning percentage in parentheses):
 
Detroit -  Bal, NYG, @Min (.397)
Chicago - @Cle, @Phi, GB (.474)
Green Bay - @Dal, Pit, @Chi (.487)
 
Games remaining for the NFC East race  (with total opposition winning percentage in parentheses):
 
Philly - @Min, Chi, @Dal (.449)
Dallas - GB, @Wash, Phi (.449)
 
And Arizona's schedule, just to round out the field: @ Ten, @ Sea, vs. SF (.641 WPct - Ouch)
 
Seattle's magic number to clinch the division (SF) as well as the overall #1 seed (NO) is 2.
 

normstalls

Member
SoSH Member
Mar 15, 2004
4,486
RedOctober3829 said:
Next week could be pivotal for the Patriots. If they can win at Miami, they can watch Cincinnati play a tough road game at Pittsburgh on SNF. If the Steelers win, which isn't a stretch, they could go up 2 games with 2 to play for #2 seed.
 
Pardon my ignorance...why is there no consideration for KC as the #2 seed?  Is there a stipulation that doesn't allow both 'bye teams' (1 and 2 seeds) to be from the same division?
 

Briz

Member
SoSH Member
Jan 23, 2011
1,509
NH
normstalls said:
 
Pardon my ignorance...why is there no consideration for KC as the #2 seed?  Is there a stipulation that doesn't allow both 'bye teams' (1 and 2 seeds) to be from the same division?
 
Based on FO's sims, KC gets the #1 seed 3.5% of the time and #2 seed 1.4% of the time.  The issue is (as mpx pointed out) leapfrogging Den for the division.  Due to Den sweeping them home and away, KC has to make up 2 games with 3 to play.  Honestly, I'm actually surprised the % is as high as it is considering Den finishes SD, @Hou and @Oak.  KC has @Oak, Indy, @SD.  
 

epraz

Member
SoSH Member
Oct 15, 2002
6,175
I'm also surprised that FO gives KC a better chance to get the #1 than #2 seed, given that the Pats have a better conference record than KC.
 

lambeau

Member
SoSH Member
Feb 7, 2010
1,175
Connecticut
Tonight realistically is the last chance for the Broncos to lose the #1 seed given that they finish with Oakland and Houston.
 

H78

Fists of Millennial Fury!
SoSH Member
Jul 22, 2009
4,613
This is completely random, but you know what dawned on me recently? I've seen a lot of people mention that this team "should" or "could" have only one loss right now if the Panthers and Jets games were called properly.
 
But couldn't this team easily be 4-9 right now as well? It took last second plays, Gostowski's insane season, and/or miracles for them to beat the Bills, Falcons, Saints, Broncos, Texans and Browns. I realize that's the nature of the NFL and every team in the league has close wins and losses that could have gone either way, but I guess my point is as much as we've been talking about the tough luck this team has had, they've had some pretty good luck when it comes to wins and losses:
 
-A first-game rookie quarterback and two fourth quarter field goals from Gost in Buffalo
-Talib breaking up the pass in the end zone against the Falcons
-The miracle comeback against the Saints
-The miracle comeback against the Broncos
-The Texans pissing down their leg and Gost making long field goals
-The miracle comeback against the Browns
 
Let's say none of those work out for the 2013 Patriots (just look at the 2013 Texans as an example of a team with horrible close-game luck/ability). They'd be 4-9 right now. Hell, even if half of those didn't go their way they'd be 7-6 and tied with the Phins going into Sunday.
 
I do believe that on some level you earn your luck, but as much as we have complained about injuries and bad calls in a couple of losses this year, it really is a miracle this team hasn't lost at least a couple more games. Not trying to be a downer; I'm actually trying to point out how good luck is better than no (or bad) luck, and the Patriots have had a lot of good luck.
 

dynomite

Member
SoSH Member
Absolutely.  
 
The true issue in the NFL -- especially in 2013 -- is the stunning parity of the league.  The difference between "good" and "bad" teams is vanishingly small.  So many games are decided by one play or one call by an official.
 
Think about the Chargers.  They're 6-7 and on the outside of the NFL playoffs looking in.  This is a a team that blew a 21-point 2nd half lead Week 1 to the Texans, then lost on a Hail Mary against the Titans in Week 3, then in Week 9 couldn't score from the 1 yard line at the end of regulation and lost in OT to the Redskins.
 
Change two plays during the entire season and they're 8-5 and a dangerous playoff team.
 

SMU_Sox

queer eye for the next pats guy
SoSH Member
Jul 20, 2009
8,878
Dallas
No - they couldn't easily be 4-9. They have a high level of efficiency. It is more likely that they are at 10 wins than 4. FO estimates their wins based on their strength of schedule and how efficient they have been at 8.6 wins. I don't see it as breaks. Because you could point to the other team getting breaks too. I like looking at efficiency. Teams that play at a high efficiency should win more. Saying a team that has played as efficiently as the Pats could easily be at 4 wins is laughable and reeks of if my aunt had balls she'd be my uncle type of analysis.
If you define parity as most teams are within 2 TDs + EP's of each other than yes, there is parity in the NFL. Unlike many years where there are 4-5 elite teams, 4-5 very good ones, etc, this year teams are more bunched up. But there are certainly many teams that are very below average.
 
You used the Chargers as an example. The reason they are 6-7 is because they have an almost historically bad defense. Yes, they have a really good offense, but when you combine a great offense with terrible defense and mediocre special teams you are probably going to be around .500 or below. They blew a 21 point lead because their D can't stop anything.
 

Stitch01

Member
SoSH Member
Jul 15, 2005
18,155
Boston
4-9 would take some really shitty luck.  9-4 or 8-5 seems about right, not surprised FO is showing 8.6 as an estimate.  10-3 is a fortunate outcome by a game or two, Id buy that.
 
None of this accounts for injury luck, which is a huge driver in how teams finish and been a real drag this year.
 
The Chargers example does illustrate there is a fair bit of variance around that expected win number because of the nature of football and playing so few games in a season.
 

dynomite

Member
SoSH Member
SMU_Sox said:
No - they couldn't easily be 4-9. They have a high level of efficiency. It is more likely that they are at 10 wins than 4. FO estimates their wins based on their strength of schedule and how efficient they have been at 8.6 wins. I don't see it as breaks. Because you could point to the other team getting breaks too. I like looking at efficiency. Teams that play at a high efficiency should win more. Saying a team that has played as efficiently as the Pats could easily be at 4 wins is laughable and reeks of if my aunt had balls she'd be my uncle type of analysis.
 
Sorry, I agree with this.  "Easily" being 4-9 is silly, but so is claiming that efficiency explains everything.  Hell, the Dolphins were a toe away from being 6-7 last weekend, regardless of everything else that happened in the game and season to that point. 
 
I suppose my saying "absolutely" was more about agreeing with H78's general point (as I took it) that so many NFL games are ultimately decided by one or two plays and can have a dramatic impact on the standings. 
 
And as for the Chargers, who cares about the defense?  Their defense didn't overturn Danny Woodhead's TD against the Redskins.  Sure the Chargers defense gave up lots of points and if they'd played better the Chargers wouldn't have needed that play to win.  But they did need that play, and there's an element of... something that prevented Woodhead from reaching the ball the extra inch they needed to win.  You can call it execution or efficiency if you want, but I feel like there's a bit of luck as well.
 

SMU_Sox

queer eye for the next pats guy
SoSH Member
Jul 20, 2009
8,878
Dallas
I never said efficiency explains everything. I said with their level of efficiency they are more likely to be at 10 wins than 4. That seems pretty reasonable. But if you wanted to do a regression model on DVOA% and actual wins and losses... my friend I think it has a very high correlation. You're looking at one or two plays but without a positive or negative efficiency those plays don't mean much. That fourth and inches is in garbage time and not when you are down by 5. You're focusing on individual plays instead of the context that got them there. Most close games are close because the team played at a certain level before the final drive. Your analysis looks to that final drive as descriptive of what happened in the game and assigns bad luck if they lost or good luck if they won based on that drive. I think you should look at the whole game instead and weigh each drive. Sure the final drives counts for more but it isn't the whole game.
 
Of course there is luck involved. There is randomness. But efficiency explains a lot and I guarantee you it correlates more with actual wins and losses than anything else we have to look at aside from maybe a pythag regression.
 

SMU_Sox

queer eye for the next pats guy
SoSH Member
Jul 20, 2009
8,878
Dallas
If you want to analyze the Chargers and start it off with "Who cares about the defense" we are worlds apart in what matters to us when looking at teams. Frankly I think that makes absolutely no sense.
 
There are 3 parts to a football performance: offense, defense, and special teams. That they lost games because their #2 offense wasn't even better than it already is doesn't mean much to me. FO ranks their O as 2nd best, D as 32nd worst (#1 worst overall) and their special teams as 15th. If you take that formula you would expect, what a .500 or so record, right? FO has their estimated wins at 6.2. That seems right. If you look through the looking glass you're going to get a very distorted view. They let WAS rack up 500 yards of net offense. That's why they lost.
 
If they were 8-5 their DVOA% and efficiency would be a lot better. HOU racked up 449 yards of offense to their 260 some. That's a game they aren't going to win no matter if they had a 3 TD lead or not. If you allow teams to obliterate your defense consistently you're not going to be that damn good. Saying they blew it is right but misses the point. That's who they are. Their D sucks. No lead is safe. They'd be 8-5 with a lot of luck and/or if they were a fundamentally different club.
 

Rook05

Well-Known Member
Lifetime Member
SoSH Member
Jul 18, 2005
3,115
Boulder, CO
lambeau said:
So it's 35 in Denver now  and we can pray for turnovers.
It's actually not bad here now, and it certainly doesn't feel like the mid-30s in New England. Peyton shouldn't have any trouble throwing the ball.

But fuck him for fucking up my commute.
 

Euclis20

Member
SoSH Member
Aug 3, 2004
8,014
Imaginationland
RedOctober3829 said:
Pats win out and they are the 1 seed. Even if they finish in a 3-way tie at 13-3 with Denver and KC, they'll have the tiebreaker due to better conference record.
Yes and no.  If the three teams all finish with the same record, I'm pretty sure the AFC west tiebreakers would kick in first, giving that division to Denver.  Then it would be Denver vs NE, which NE would win on the head to head.  The only way you get 3-way ties is if the teams are from 3 different divisions.  Different path, same result.
 

Stitch01

Member
SoSH Member
Jul 15, 2005
18,155
Boston
dynomite said:
With road games in Miami and Baltimore (and no Gronk) I'll be thrilled (and surprised) if they win out.
Getting the #1 seed this season would rank with Belichick's most impressive feats, yes?
Yeah probably closer to 1/3 than 50/50, but legit in play. Would be up there with 2001 and 2008 imho.
 

wilked

Member
SoSH Member
Jul 17, 2005
4,044
@Dolphins (50%)
@Ravens (50%)
Bills (75%)
 
I see the books have the game this weekend as essentially even, and can't expect much different at Baltimore.  I figure at home vs Bills they will be a touchdown favorite or so.
 
0.5x0.5x0.75 = ~20%
 
If you think they have a better chance of winning all three games, you either think the books are wrong this weekend (maybe), think they have a better chance of winning next weekend than this weekend (maybe), think they will be better than a TD favorite against the Bills (possible), or you are overestimating the odds.
 
-------------
On the flip side, Denver finishes with @Texans @Raiders.  Last night they were 10 point favorites at home against a pretty good team...on the road against terrible talent they will be 10 point favorites as well.
 
0.85x0.85 = 72% they win out
 
-------------
80% of the time Pats drop 1 of the 3 games, and in that scenario only 28% of the time does Denver slip up.  In other words, still a tough battle ahead.  I am not exactly sure how to combine the two, but agree with Stitch, closer to 1/3 than 1/2 time in my opinion
 

m0ckduck

Member
SoSH Member
Jul 20, 2005
1,714
wilked said:
@Dolphins (50%)
@Ravens (50%)
 
I see the books have the game this weekend as essentially even, and can't expect much different at Baltimore. 
 
I would agree that this is a good line... except for the 45-6 December record under Belichick. In light of that, 50% seems too low. Which is to say, if I were a betting man, I'd be hitting the Pats this weekend and maybe next. As far as likelihood of the 1 seed, I'd split the difference and say 40% chance, personally. 
 

Bucknahs Bum Ankle

Member
SoSH Member
Oct 31, 2003
8,487
Taintopolis
wilked said:
@Dolphins (50%)
@Ravens (50%)
Bills (75%)
 
I see the books have the game this weekend as essentially even, and can't expect much different at Baltimore.  I figure at home vs Bills they will be a touchdown favorite or so.
 
0.5x0.5x0.75 = ~20%
 
If you think they have a better chance of winning all three games, you either think the books are wrong this weekend (maybe), think they have a better chance of winning next weekend than this weekend (maybe), think they will be better than a TD favorite against the Bills (possible), or you are overestimating the odds.
 
-------------
On the flip side, Denver finishes with @Texans @Raiders.  Last night they were 10 point favorites at home against a pretty good team...on the road against terrible talent they will be 10 point favorites as well.
 
0.85x0.85 = 72% they win out
 
-------------
80% of the time Pats drop 1 of the 3 games, and in that scenario only 28% of the time does Denver slip up.  In other words, still a tough battle ahead.  I am not exactly sure how to combine the two, but agree with Stitch, closer to 1/3 than 1/2 time in my opinion
 
Roughly, 0.2 (Pats win out) + (0.8 x 0.28) (Pats don't win out and Denver doesn't win out) = 42%.  
 

Bellhorn

Lumiere
SoSH Member
Aug 22, 2006
2,328
Brighton, MA
Bucknahs Bum Ankle said:
Roughly, 0.2 (Pats win out) + (0.8 x 0.28) (Pats don't win out and Denver doesn't win out) = 42%.
But you have to account for the possibility of losing two games. Leaving out the possibility of Denver losing out the rest of the season, it is something like 0.2 (Pats win out) + (0.56 x 0.28) (Pats lose one and Denver loses one) = 35%.
 
Edit: brainfart this morning, should be 0.44 x 0.28, making the final answer around 33%.  But as noted, these are made-up numbers in any case.
 

Stitch01

Member
SoSH Member
Jul 15, 2005
18,155
Boston
I did it out for every combination using 55/55/85 for the Pats and 90/90 for Denver and came out to 30%.  Maybe 90/90 is too high given Denver is on the road for both games, but betting markets have them -9.5/-12 right now for the Texans and Raiders games which is 85/95% by money line converters so I think its not crazy off.
 
Denver losing once/Pats go 2-1 being the path was more relatively likely than I would have thought, guess math isnt my strong suit. 
 

Bucknahs Bum Ankle

Member
SoSH Member
Oct 31, 2003
8,487
Taintopolis
Bellhorn said:
But you have to account for the possibility of losing two games. Leaving out the possibility of Denver losing out the rest of the season, it is something like 0.2 (Pats win out) + (0.56 x 0.28) (Pats lose one and Denver loses one) = 35%.
True.  And factoring in Pats lose 2, Broncos lose 2 adds about 1%.  I went with "roughly" because we're already using made up percentages for each game (albeit with a basis in odds making).  Somewhere around 35% - 40% sounds right, but I'm an optimist.
 

wilked

Member
SoSH Member
Jul 17, 2005
4,044
30 or 40%, whatever the case, Pats still have an uphill battle for top seed.  That said, things look a helluva lot better today than yesterday.  More than anything, they control their own destiny, and the frustration you hear in Peyton's postgame last night speaks to this more than anything I think.
 

westneat

New Member
Apr 16, 2013
30
Stitch01 said:
I did it out for every combination using 55/55/85 for the Pats and 90/90 for Denver and came out to 30%.  Maybe 90/90 is too high given Denver is on the road for both games, but betting markets have them -9.5/-12 right now for the Texans and Raiders games which is 85/95% by money line converters so I think its not crazy off.
 
Denver losing once/Pats go 2-1 being the path was more relatively likely than I would have thought, guess math isnt my strong suit. 
 
Keep in mind we would also need KC to lose a game, because if the Pats and Broncos both lose one and KC wins out they get the 1.
 

Stitch01

Member
SoSH Member
Jul 15, 2005
18,155
Boston
Didnt factor that in, would haircut it by a few percent.  KC isnt that likely to win out, they'll be dogs @ SD
 

loshjott

Member
SoSH Member
Dec 30, 2004
14,943
Silver Spring, MD
Bucknahs Bum Ankle said:
 
Roughly, 0.2 (Pats win out) + (0.8 x 0.28) (Pats don't win out and Denver doesn't win out) = 42%.  
 
Are the Pats favored to get the #1 seed?
 
How do you figure Denver's chances in this scenario?
 
80% chance Pats don't win out x 72% chance Denver does win out = 57.6% chance Denver gets the top seed?
 

Bucknahs Bum Ankle

Member
SoSH Member
Oct 31, 2003
8,487
Taintopolis
loshjott said:
 
Are the Pats favored to get the #1 seed?
 
How do you figure Denver's chances in this scenario?
 
80% chance Pats don't win out x 72% chance Denver does win out = 57.6% chance Denver gets the top seed?
 
No, Pats aren't favored.  The percentages were in the post I responded to.  Yes I came up with basically 42% Pats/58% Broncos. That has since been further fleshed out to something like 36% Pats/63.5% Broncos/0.5% Chiefs. 
 

TomTerrific

Member
SoSH Member
Jul 15, 2005
2,700
Wayland, MA
Bucknahs Bum Ankle said:
 
No, Pats aren't favored.  The percentages were in the post I responded to.  Yes I came up with basically 42% Pats/58% Broncos. That has since been further fleshed out to something like 36% Pats/63.5% Broncos/0.5% Chiefs. 
 
Well, Football Outsiders agrees reasonably closely with the numbers being tossed around here, somewhere around a 38% chance for the Pats (up from 18% before the Broncos' loss).
 

Bucknahs Bum Ankle

Member
SoSH Member
Oct 31, 2003
8,487
Taintopolis
TomTerrific said:
 
Well, Football Outsiders agrees reasonably closely with the numbers being tossed around here, somewhere around a 38% chance for the Pats (up from 18% before the Broncos' loss).
 
Actually they've got it as DEN 41.9%, NE 37.1%, KC 10.9%, CIN 10.2%.  I didn't expect KC or CIN had that much of a shot.  And they've got it much closer between DEN and NE.  A win on Sunday should put us in the lead.
 

Bergs

funky and cold
SoSH Member
Jul 22, 2005
21,612
While I understand using conditional probabilities to estimate the odds here, to me it is a simpler question:
 
Will the Pats play well on offense without Gronk? If they do, I think they win out. If not, they won't (and might even go 1-2).
 
I tend to believe they will play well, and they will win out, and they will have the 1 seed, and they will lose in the playoffs.
 

Euclis20

Member
SoSH Member
Aug 3, 2004
8,014
Imaginationland
TomTerrific said:
 
Well, Football Outsiders agrees reasonably closely with the numbers being tossed around here, somewhere around a 38% chance for the Pats (up from 18% before the Broncos' loss).
 
I [obviously] understand why last night's game lowered Denver's odds of winning the Super Bowl.  I also understand why last night's game increased all other AFC team's odds of winning the Super Bowl.  I think I understand why the Super Bowl odds of most NFC playoff teams (Seattle, Carolina, SF, Philly) increased slightly, presumably because Denver is the best team in the AFC, and the odds are now greater that a weaker AFC team will make it to the Super Bowl.  But I do not understand why New Orleans (from 15% to 14.6% and Detroit (from 1% to .9%) now are LESS likely to win the Super Bowl. 
 

dbn

Member
SoSH Member
Feb 10, 2007
7,785
La Mancha.
All your maths are wrong.  According to the Ravens fans I spoke with at the MIN@BAL game last Sunday, the Patriots have a 0% chance of winning in Baltimore. 
 

dcmissle

Deflatigator
Lifetime Member
SoSH Member
Aug 4, 2005
28,269
What would you expect them to say?
 
It's death match football.  I'll be there in 9 days, and Pats may have the 1 seed on the line -- or the 2 (gulp).  Ravens will be playing for a playoff spot and maybe the division.
 
Other than a healthy roster, we can't ask for much more.
 

dbn

Member
SoSH Member
Feb 10, 2007
7,785
La Mancha.
dcmissle said:
What would you expect them to say?
 
It's death match football.  I'll be there in 9 days, and Pats may have the 1 seed on the line -- or the 2 (gulp).  Ravens will be playing for a playoff spot and maybe the division.
 
Other than a healthy roster, we can't ask for much more.
 
It wasn't the usual hyperbole that comes with being a fan, though.  It was as if they really held the sober matter-of-fact belief that a Patriots victory was a mathematical impossibility. This, in a league where the "any given Sunday" maxim is re-proven on an almost weekly basis. (In that very game, the Ravens needed 3 TDs in the final 2:05 to beat a Peterson-less Vikings squad at home.)
 
Edit to add that the game was a blast. It was the first snow-game in M&T Bank Stadium history, a deluge of snow and sleet. Since moving to SoCal years ago, I've really missed that kind of game. Also, despite their overconfidence, I have to admit that the Ravens fans were overall a pretty classy group.
 

Bellhorn

Lumiere
SoSH Member
Aug 22, 2006
2,328
Brighton, MA
Euclis20 said:
 
I [obviously] understand why last night's game lowered Denver's odds of winning the Super Bowl.  I also understand why last night's game increased all other AFC team's odds of winning the Super Bowl.  I think I understand why the Super Bowl odds of most NFC playoff teams (Seattle, Carolina, SF, Philly) increased slightly, presumably because Denver is the best team in the AFC, and the odds are now greater that a weaker AFC team will make it to the Super Bowl.  But I do not understand why New Orleans (from 15% to 14.6% and Detroit (from 1% to .9%) now are LESS likely to win the Super Bowl. 
Yeah, that is kind of weird.  The only thing that really makes sense (as far as I can see) is if FO somehow has New Orleans and Detroit matching up better against Denver than they do against the rest of the field in the AFC.
 

dbn

Member
SoSH Member
Feb 10, 2007
7,785
La Mancha.
Bellhorn said:
Yeah, that is kind of weird.  The only thing that really makes sense (as far as I can see) is if FO somehow has New Orleans and Detroit matching up better against Denver than they do against the rest of the field in the AFC.
 
The odds are based DVOA that accounts for the DEN v SD game. The DVOA for DEN and SD will have changed due to that game, which will have a ripple-effect.  For example, if DEN's DVOA went down, then it is less impressive that NE beat DEN so NE's goes down, which means that NO lost to a "weaker" NE team than DVOA thought before last night's game. Etc.
 

Bellhorn

Lumiere
SoSH Member
Aug 22, 2006
2,328
Brighton, MA
dbn said:
The odds are based DVOA that accounts for the DEN v SD game. The DVOA for DEN and SD will have changed due to that game, which will have a ripple-effect.  For example, if DEN's DVOA went down, then it is less impressive that NE beat DEN so NE's goes down, which means that NO lost to a "weaker" NE team than DVOA thought before last night's game. Etc.
Yeah, at first I thought that might be the explanation, but do they really update the DVOA ratings (even internally) on a continuous basis? Somehow I had the impression that DVOA calculation was only a once-per-week operation, and that the change in playoff odds would only reflect changes in W-L until then.
 

dbn

Member
SoSH Member
Feb 10, 2007
7,785
La Mancha.
Bellhorn said:
Yeah, at first I thought that might be the explanation, but do they really update the DVOA ratings (even internally) on a continuous basis? Somehow I had the impression that DVOA calculation was only a once-per-week operation, and that the change in playoff odds would only reflect changes in W-L until then.
 
I have no idea how they do their stuff, but once the code is set up I imagine it should simply be a matter of entering the new data and hitting "enter".  Anyhow, it's not that important and probably doesn't deserve more space in this thread, but I do like thinking about such things for some reason.
 

DJnVa

Dorito Dawg
SoSH Member
Dec 16, 2010
53,837
Well, it'sn't all based on thousands of seasons being simmed?
 
Even if you never changed anyone's ratings, those numbers will slightly change each time you run that thing. It could just be noise in the numbers.