No. 1 seeds: Laying waste to championship games since 2005 (1 Viewer)

How's this for a crazy stat (disclaimer: not a gambler, and anyone who gambles based on these insights does so at their own risk) 

• Between the 1995-2005 seasons, No. 1 seeds in championship games were 9-7 (0.563) and non-No. 1 home teams were 2-2.
  
• Since 2005, non-No. 1 home teams have gone 5-5. But No. 1 seeds have gone 15-1 (0.938)! The only 1-seed to lose during this time period was Atlanta in 2012.
  
• Also, prior to 2013, there had been only eight Super Bowl matchups featuring two No. 1 seeds. Since then, it has happened in four out of the last five years.
  

So basically, prior to 2005 championship games were pretty much 50/50 affairs. Since that time, they have remained so for everyone else, but No. 1 seeds have thoroughly dominated the field.

I don't have any explanation for why this has happened, or what might have changed over the past 13 years. And obviously a lot of these games were super close and could have gone either way (Pats last year, Seattle in 2014, Pats in 2011, Saints in 2009). And there's also the fact that nearly 40% of the time, the No. 1 seed was eliminated in the divisional round. So maybe the weaker 1-seeds were weeded out, and the ones that made it through to the championship game were dominant?

What say you, statheads? Is it just noise? Too small of a sample size?

Not drawing any conclusions, but it definitely seems like there's something there. If an NFL team went 15-1 in a season it would be hard to argue that they weren't empirically better than a 9-7 team.

zftcg said:

How's this for a crazy stat (disclaimer: not a gambler, and anyone who gambles based on these insights does so at their own risk) 

• Between the 1995-2005 seasons, No. 1 seeds in championship games were 9-7 (0.563) and non-No. 1 home teams were 2-2.
  
• Since 2005, non-No. 1 home teams have gone 5-5. But No. 1 seeds have gone 15-1 (0.938)! The only 1-seed to lose during this time period was Atlanta in 2012.
  
• Also, prior to 2013, there had been only eight Super Bowl matchups featuring two No. 1 seeds. Since then, it has happened in four out of the last five years.
  

So basically, prior to 2005 championship games were pretty much 50/50 affairs. Since that time, they have remained so for everyone else, but No. 1 seeds have thoroughly dominated the field.

I don't have any explanation for why this has happened, or what might have changed over the past 13 years. And obviously a lot of these games were super close and could have gone either way (Pats last year, Seattle in 2014, Pats in 2011, Saints in 2009). And there's also the fact that nearly 40% of the time, the No. 1 seed was eliminated in the divisional round. So maybe the weaker 1-seeds were weeded out, and the ones that made it through to the championship game were dominant?

What say you, statheads? Is it just noise? Too small of a sample size?

Not drawing any conclusions, but it definitely seems like there's something there. If an NFL team went 15-1 in a season it would be hard to argue that they weren't empirically better than a 9-7 team.

Click to expand...

Not sure how to answer this, maybe just that 1 seeds are obviously better?  If there's one dominant team in a conference like there sometimes is, they're the top seed so SHOULD make it all the way to the end.  It's the same in basketball... 1 seeds have astronomically better success than 2 seeds in both nba and college basketball.

There's also the fact of who they are playing.  Let's first assume your 'non-No.1 home team" is the 2 seed in the divisional round.  These teams automatically play the strongest seed coming out of weekend one, so there's often a very strong 3 seed as well, which would of course be a much harder opponent than the 4 seed (or even a 5 or 6 who snuck through round 1 and has to play another road game in a row). 

The 2 seed is forced to play a team that just played a home game, while the 1 seed LIKELY is playing a team on back to back road games (unless there's no upsets in wild card round which is rare.)  Back to back road games the heading to a team that is well rested from a bye, and was the top team in the conference, is very difficult (See Indy, See Philly). 

As to why this is more prevalent now than pre 2005, I'm not sure... but I do see the huge advantage to being a 1 seed over a 2 seed besides just the home field in the conf. championship game.  The league seems to have more parity in the regular season lately, but less in playoffs. 

Deamon said:

Not sure how to answer this, maybe just that 1 seeds are obviously better?  If there's one dominant team in a conference like there sometimes is, they're the top seed so SHOULD make it all the way to the end.  It's the same in basketball... 1 seeds have astronomically better success than 2 seeds in both nba and college basketball.

There's also the fact of who they are playing.  Let's first assume your 'non-No.1 home team" is the 2 seed in the divisional round.  These teams automatically play the strongest seed coming out of weekend one, so there's often a very strong 3 seed as well, which would of course be a much harder opponent than the 4 seed (or even a 5 or 6 who snuck through round 1 and has to play another road game in a row). 

The 2 seed is forced to play a team that just played a home game, while the 1 seed LIKELY is playing a team on back to back road games (unless there's no upsets in wild card round which is rare.)  Back to back road games the heading to a team that is well rested from a bye, and was the top team in the conference, is very difficult (See Indy, See Philly). 

As to why this is more prevalent now than pre 2005, I'm not sure... but I do see the huge advantage to being a 1 seed over a 2 seed besides just the home field in the conf. championship game.  The league seems to have more parity in the regular season lately, but less in playoffs. 

Click to expand...

All makes sense, but as you say, it doesn't explain why it's changed so much in recent years. I mean, you could say that the .500 record pre-2005 is unusual as well, since you'd expect HFA to carry more weight, but 9-7 isn't that big of an outlier, certainly not compared to 15-1. Also, all the talk about byes and HFA doesn't explain why non-No. 1 home teams (which yes, are almost all No. 2s) have continued to be a coin flip. Atlanta in 2016, SF in 2011, Chicago in 2010, Pitt in 2008, GB in 2007 and Denver in 2005 all had byes and were playing teams coming off road wins (in most of those cases, two road wins). Their combined record: 2-4

I think it's likely a result of small sample size combined with multiple endpoints. I mean, there's no particular reason to choose games since 2005 to analyze, and there will always be streaks in the data if you look for them. #1 seeds will normally be pretty strong favorites to win at home. If you combine the two eras you're looking at 26-8 which is probably a little above the win expectation level (which I'd guess is about 2/3) but within the range you'd expect.

Maybe the top seeds have now fine tuned their strategy going into the playoffs. There used to be a lot more conflict over resting guys and going into the playoffs in a non-win mode. 

CalBear said:

I think it's likely a result of small sample size combined with multiple endpoints. I mean, there's no particular reason to choose games since 2005 to analyze, and there will always be streaks in the data if you look for them. #1 seeds will normally be pretty strong favorites to win at home. If you combine the two eras you're looking at 26-8 which is probably a little above the win expectation level (which I'd guess is about 2/3) but within the range you'd expect.

Click to expand...

Was hoping you'd chime in, but thought it might be presumptuous to tag you. 

You're absolutely right about arbitrary endpoints. It's like when the announcer cites some trends in a team's past X games, because if you go X+1 the numbers get a lot more muddy. I suppose if you really wanted to do this right, you'd look at how each seed has performed across all playoff games over a fairly long timeframe and see if there were any trends based on round, home-road, etc. Like I said, I definitely wouldn't bet my house on Chiefs-Saints based solely on this stat.

Wasn't there some stat about the Super Bowl coin toss from a few years ago? Like it came up heads a bunch of times in a row, or the NFC team won it, or something? Can't remember what it was, but it was one of those things that seems statistically unlikely but is also clearly random.

Years ago I used to do an office pool where we picked games against the spread. One year (I think 2006) a few of us figured out that underdogs were absolutely killing it. One guy figured it out before me and started picking dogs in every single game. He ran away with the season-long title. I came around on it too late to catch him, but my dog-heavy strategy did win a number of individual weeks, which came with a pretty hefty payout (I think I cleared like $1,200 that year from a $50 buy-in). Never figured out what, if anything, caused the blip, and as far as I know it never repeated again.

ETA: Here's a column Bill Simmons wrote in Dec. 2006 talking about the performance of underdogs that year. At the time he was writing, they were 136-97-6 (.584)

home field means alot

Deamon said:

Not sure how to answer this, maybe just that 1 seeds are obviously better?  If there's one dominant team in a conference like there sometimes is, they're the top seed so SHOULD make it all the way to the end.  It's the same in basketball... 1 seeds have astronomically better success than 2 seeds in both nba and college basketball.

There's also the fact of who they are playing.  Let's first assume your 'non-No.1 home team" is the 2 seed in the divisional round.  These teams automatically play the strongest seed coming out of weekend one, so there's often a very strong 3 seed as well, which would of course be a much harder opponent than the 4 seed (or even a 5 or 6 who snuck through round 1 and has to play another road game in a row). 

The 2 seed is forced to play a team that just played a home game, while the 1 seed LIKELY is playing a team on back to back road games (unless there's no upsets in wild card round which is rare.)  Back to back road games the heading to a team that is well rested from a bye, and was the top team in the conference, is very difficult (See Indy, See Philly). 

As to why this is more prevalent now than pre 2005, I'm not sure... but I do see the huge advantage to being a 1 seed over a 2 seed besides just the home field in the conf. championship game.  The league seems to have more parity in the regular season lately, but less in playoffs. 

Click to expand...

Tom Brady and Bellichick being great and Buffalo and Jets in their division every year.  The Pats are for 5-1 in home conference championship games, not sure how many of those were as a #1 seed.

zftcg said:

Since 2005, non-No. 1 home teams have gone 5-5. But No. 1 seeds have gone 15-1 (0.938)!

Click to expand...

Out of curiosity do we know the Pats’ number out of this?

How many home teams in these playoffs lost?

SaintsInDome2006 said:

Out of curiosity do we know the Pats’ number out of this?

Click to expand...

17 W

16 W

15  L away

14 W

13 L away

12 L 

11 W

07 W

06 L away

Pats are 5-0 as the #1 seed.  The bolded away are Pats losses to the #1 seeded team.

There have been 24 seasons in the salary cap / free agency era. Yes, in recent seasons the #1 seeds have advanced to the SB quite a bit. But looking at that time span as a whole, 23 times a #1 seed did not get to the Super Bowl (out of 48 teams). Essentially, top seeds reached the Super Bowl 52% of the time.

ZenoRazon said:

How many home teams in these playoffs lost?

Click to expand...

Home teams have gone 5-3 this season so far.

BassNBrew said:

17 W

16 W

15  L away

14 W

13 L away

12 L 

11 W

07 W

06 L away

Pats are 5-0 as the #1 seed.  The bolded away are Pats losses to the #1 seeded team.

Click to expand...

They also won as a 1 seed in 03 (Colts) so 6-0.

ETA: Just realized you were talking specifically about the period I outlined in the OP. Carry on.

Something else I realized that argues in favor of the "statistical noise" theory: I mentioned in the OP that No. 1 seeds are 15-1 over the past 13 years, and were 9-7 over the previous 10. That means that as many 1 seeds advanced to the championship over the last 13 years as did over the previous decade. Or, to put it another way, 1 seeds were 16-10 (.615) in the divisional round from 2005-2017, but 16-4 (.800) from 1995-2004. Then again, if we include this year's divisionals, they're 11-1 over the past six years. 

I still think it's interesting, and I'm not ruling out the possibility that there's some factor contributing to this, but the most likely explanation is that it's nothing and we'll see a reversion to the mean at some point.

Nothing to add on the actual numbers, but if you are looking at the arbitrary cutoff and a reason for the change in outcomes:

Roll back from 2005 to 2002 and you have a non-arbitrary cutoff.  That year the league expanded and went from three divisions to four.  Maybe the restructured playoff allowed teams that dominated the new smaller divisions an easier road to the playoffs, winning them not only the high seed and bye, but a less taxing path to get there?

Just a thought.  Seems like something changed, looking for theories.

Arodin said:

Nothing to add on the actual numbers, but if you are looking at the arbitrary cutoff and a reason for the change in outcomes:

Roll back from 2005 to 2002 and you have a non-arbitrary cutoff.  That year the league expanded and went from three divisions to four.  Maybe the restructured playoff allowed teams that dominated the new smaller divisions an easier road to the playoffs, winning them not only the high seed and bye, but a less taxing path to get there?

Just a thought.  Seems like something changed, looking for theories.

Click to expand...

OK, so testing this hypothesis:

2002-2017

• No. 1 seeds in championship: 18-4 (.818)
  
• Non No. 1 home teams in championship: 5-5 (.500)
  
• Home teams in championship: 23-9 (.719)
  
• No. 1 seeds in divisional round: 22-10 (.688)
  

1986-2001

• No. 1 seeds in championship: 17-9 (.654)
  
• Non No. 1 home teams in championship: 3-3 (.500)
  
• Home teams in championship: 20-12 (.625)
  
• No. 1 seeds in divisional round: 26-6 (.813)
  

The fact that divisional performance has gone in the other direction is the thing that makes me think this is all noise. No. 1 seeds won more than 80% of their divisional games in the first 16-year period, then dropped to below 70% in the next one. In championship games, those numbers were reversed. Certainly would suggest that it could just as easily flip back from 2018-2033, or settle somewhere in between.

But I suspect it's impossible to prove one way or another. Question is, if it goes on long enough do we reach a point where we have to dismiss the noise argument?

Of course, the other possibility is that, now that I've identified a trend, it will definitely end immediately (see: Redskins Rule), so bet it all on Pats-Rams!

The hypothesis is interesting, at least. My intuition would be that having four division winners and two wild cards will result in an overall weaker playoff field than having three division winners and three wild cards. (More competition for the wild card spots). So maybe there is a bit of an inflection in that #1 seeds (who presumably are as good as they were before 2002) have a slightly increased advantage. But yes, that should show up in the divisional round, even more so than in the championship round, if it's a real thing.

I'll show myself out 

Yikes.

So this entire trend was just because of the Pats winning in the playoffs?

Somehow the letter Z seems to be jinxed.

Deamon said:

Yikes.

So this entire trend was just because of the Pats winning in the playoffs?

Click to expand...

Not really. Pats' win yesterday was their first road win in an AFCCG since 2004. During that time they've lost twice to a 1-seed (and once at home as the 2). And obviously, if 15-1 over more than a decade doesn't mean anything, then 0-2 in one year doesn't either.