Perspective Thread: How many Super Bowls should the Patriots have won by now?

[zydecochris]

Since 2001, New England has made the playoffs twelve times, appearing in 9 conference championship games and making it to the Super Bowl on six occasions. At this point, the Patriots have won the Super bowl four times over a fourteen-year span. By these metrics, if the New England Patriots were a poker player, they would have made nine out of 15 final tables, cashed six times, and gone on to play heads up for the title 6 times, so far winning four!

Yet Chase Stuart of footballperspective.com asks the question: how many Super Bowls should the Patriots have won?

Stuart calculates the odds of the Patriots winning the Super Bowl 2001 - 2014 according to expected winning percentage and point-spread betting. Here are his findings:



So in 2001, 2003, 2004, and 2014, the Patriots should have been expected to win 0.754 Super Bowls. Instead, the team won 4.000! How insane is that?!?

But what about the other eight years? According to Stuart, the ’05 Patriots had a 10.2% chance of winning all the games in the postseason, 06' Patriots had a 7.1% chance of winning all the games in the postseason, 07' Patriots had a 57.5% chance of winning all the games in the postseason, the 09' Patriots had an approximate 3% of winning all the games in the postseason, 10' Patriots had a 33.7% chance of winning the postseason, 11' Patriots had a 33.9% chance of winning it all, the 12' Patriots had a 26.6% of winning all the games in the postseason, and the 13' Patriots had a 12% chance of claiming the Super Bowl)

If put on a chart, the numbers would look like this:




As you can see, the Patriots massively overachieved from 2001 to 2004 (other than missing the playoffs in 2002), and then the team underachieved in the postseason for about a decade (other than missing the playoffs in 2008), winning no Super Bowls despite being expected to win about 1.8 of them.

But as Stuart points out, if you sum the Super Bowl probabilities for New England in each postseason, the Patriots should have been expected to win about 2.6 Super Bowls from ’01 to ’14, instead of the 4.0 they actually won.

Great thread.

 

[zydecochris]

1) This thread made me think about how many elements, many of them random, have to fall in place for a team to win a Super Bowl.

2) I won't second guess the math that leads to 2.6, but it's virtually worthless. All it says is that if you multiply out these 12 sets of probabilities, the mean number of successful outcomes for the Patriots turns out to be 2.6. But that falls into the logical trap that some call "the flaw of averages;" In this case, 2.6 is the average of a specific distribution of potential positive outcomes between zero and 12. It has nothing to do with real life.

3) It's impossible to say that the Pats "should" have won an SB in which they did not play. We have no way of knowing what would have happened in an 06 SB v. the Bears. Brady might have been knocked out of the game on his first play from scrimmage. A number of usually reliable players could have all played the worst games of their professional careers, making error after error that gave the victory to an inferior Chicago team.

4) So, how many "should" they have won? I want to be facetious and say "4," because that's the only verifiable answer. But, if you want me to give a number, my thought process would be that they made it to 6, so they "should" have won somewhere between "0" and "6." Interestingly enough, the average of 0 and 6 is "3," which is a rounding up of 2.6. But, I'll just chalk that up to randomness and say, "somewhere between 0 and 6."

Again, great thread, really interesting. I think the first thread I ever started was on how football should be thought of as probabilistic (involving an element of chance, such as backgammon or risk) instead of deterministic (like chess), so let me go with similar simplified probabilistic arguments here. However, I'm not going to consider point spreads and so forth, I think that is too complicated to factor in accurately.

From the simplest point of view, the Patriots have gone to 9 conference championships since 2001. I would argue that by the time you get to the conference championships, the teams are approximately equal on average (probably sometimes the Patriots are better and sometimes worse). Thus, one would expect that with a big enough sample size the Patriots would have won half of their conference championships, or 4.5. The actual number is 6, which is 33% greater than what would be expected with equal odds.

The Patriots have appeared in 6 superbowls. Again assuming that at that level the teams are approximately equal, with a big enough sample size the Patriots should have won 3 of them. The actual number is 4, which is (coincidentally) again 33% greater than what one would expect from equal odds. In summary:

Patriot Conference appearances - 9 Expected wins equal odds - 4.5 Actual Wins 6
Patriot SuperBowl appearances - 6 Expected wins equal odds - 3 Actual Wins 4

Why have the Patriots been 33% more successful than would have been expected? Part of it is doubtlessly simply the small sample size, there is certainly some element of luck. I guess one could argue that assuming equal odds isn't right because the Patriots have been more talented than their opponents, but I question that assertion.

Beyond luck, I would argue that another important factor is the superior performance of TB and BB in the clutch. TB's string of strong clutch performances on the biggest stages is well documented (for example, compared to Peyton Manning). And, BB is absolutely a rock in these big games. I will always consider BB's defining moment to be in the last minute of the last Super Bowl when he noticed confusion (and perhaps a little panic) on the Seattle sideline and calmly decided to not call a time out, leading to the pass attempt that was intercepted (on a defensive play that had been practiced many times). The calm clutch play of TB and BB is the Patriot's secret weapon.

 

[PatsFanSince74]

Again, great thread, really interesting. I think the first thread I ever started was on how football should be thought of as probabilistic (involving an element of chance, such as backgammon or risk) instead of deterministic (like chess), so let me go with similar simplified probabilistic arguments here. However, I'm not going to consider point spreads and so forth, I think that is too complicated to factor in accurately.

From the simplest point of view, the Patriots have gone to 9 conference championships since 2001. I would argue that by the time you get to the conference championships, the teams are approximately equal on average (probably sometimes the Patriots are better and sometimes worse). Thus, one would expect that with a big enough sample size the Patriots would have won half of their conference championships, or 4.5. The actual number is 6, which is 33% greater than what would be expected with equal odds.

The Patriots have appeared in 6 superbowls. Again assuming that at that level the teams are approximately equal, with a big enough sample size the Patriots should have won 3 of them. The actual number is 4, which is (coincidentally) again 33% greater than what one would expect from equal odds. In summary:

Patriot Conference appearances - 9 Expected wins equal odds - 4.5 Actual Wins 6
Patriot SuperBowl appearances - 6 Expected wins equal odds - 3 Actual Wins 4

Why have the Patriots been 33% more successful than would have been expected? Part of it is doubtlessly simply the small sample size, there is certainly some element of luck. I guess one could argue that assuming equal odds isn't right because the Patriots have been more talented than their opponents, but I question that assertion.

Beyond luck, I would argue that another important factor is the superior performance of TB and BB in the clutch. TB's string of strong clutch performances on the biggest stages is well documented (for example, compared to Peyton Manning). And, BB is absolutely a rock in these big games. I will always consider BB's defining moment to be in the last minute of the last Super Bowl when he noticed confusion (and perhaps a little panic) on the Seattle sideline and calmly decided to not call a time out, leading to the pass attempt that was intercepted (on a defensive play that had been practiced many times). The calm clutch play of TB and BB is the Patriot's secret weapon.

Definitely an Occam's Razor approach there.

Arguing that teams that get to Conference and League Championship Games are going in general to be "approximately equal" is probably valid "over the long run," but clearly not in every case.

No one has ever suggested that, on the field, the 2000/01 Patriots were equal to the Rams or the 2007/08 Giants were equal to the Patriots in those respective SB's. In 2001, the difference was indeed BB and his game plan and the coolness of Brady and Vinatieri under pressure in the final 90 seconds of the game. In 2008, the difference was a randomly freakish play by a guy who never played another down in the NFL.

Somehow, over time, the scales equalize, but I think the explanation is messier than Occam would allow.

Good comment. Thanks.

 

[zydecochris]

Definitely an Occam's Razor approach there.

Arguing that teams that get to Conference and League Championship Games are going in general to be "approximately equal" is probably valid "over the long run," but clearly not in every case.

No one has ever suggested that, on the field, the 2000/01 Patriots were equal to the Rams or the 2007/08 Giants were equal to the Patriots in those respective SB's. In 2001, the difference was indeed BB and his game plan and the coolness of Brady and Vinatieri under pressure in the final 90 seconds of the game. In 2008, the difference was a randomly freakish play by a guy who never played another down in the NFL.

Somehow, over time, the scales equalize, but I think the explanation is messier than Occam would allow.

Good comment. Thanks.

Great post, good point.

Something just occurred to me. The original excellent post in this excellent thread was by Chiuba linking to an article by FootballPerspective.com that had what appeared to me to be a very complicated analysis (that I admit I didn't read too carefully) that involved point spreads of each game, etc. Their prediction based on this complicated analysis was that the Patriots should have won 2.6 Superbowls.

I was focussing on how the Patriots beat the odds both at the Conference Level and at the Superbowl level. However, to make things even MORE simple, over a large sample size, if the teams were evenly matched, each team appearing in the Conference Final should have a 25% chance of winning the Superbowl. The Patriots have appeared in 9 Conference Finals. Therefore, 25% of 9 is 2.5. And, the complicated analysis predicted 2.6, pretty close.

Fortunately I guess BB and TB didn't read that analysis before they won their four superbowls.

 

[captain stone]

Since 2001, New England has made the playoffs twelve times, appearing in 9 conference championship games and making it to the Super Bowl on six occasions. At this point, the Patriots have won the Super bowl four times over a fourteen-year span. By these metrics, if the New England Patriots were a poker player, they would have made nine out of 15 final tables, cashed six times, and gone on to play heads up for the title 6 times, so far winning four!

Yet Chase Stuart of footballperspective.com asks the question: how many Super Bowls should the Patriots have won?

Stuart calculates the odds of the Patriots winning the Super Bowl 2001 - 2014 according to expected winning percentage and point-spread betting. Here are his findings:



So in 2001, 2003, 2004, and 2014, the Patriots should have been expected to win 0.754 Super Bowls. Instead, the team won 4.000! How insane is that?!?

But what about the other eight years? According to Stuart, the ’05 Patriots had a 10.2% chance of winning all the games in the postseason, 06' Patriots had a 7.1% chance of winning all the games in the postseason, 07' Patriots had a 57.5% chance of winning all the games in the postseason, the 09' Patriots had an approximate 3% of winning all the games in the postseason, 10' Patriots had a 33.7% chance of winning the postseason, 11' Patriots had a 33.9% chance of winning it all, the 12' Patriots had a 26.6% of winning all the games in the postseason, and the 13' Patriots had a 12% chance of claiming the Super Bowl)

If put on a chart, the numbers would look like this:




As you can see, the Patriots massively overachieved from 2001 to 2004 (other than missing the playoffs in 2002), and then the team underachieved in the postseason for about a decade (other than missing the playoffs in 2008), winning no Super Bowls despite being expected to win about 1.8 of them.

But as Stuart points out, if you sum the Super Bowl probabilities for New England in each postseason, the Patriots should have been expected to win about 2.6 Super Bowls from ’01 to ’14, instead of the 4.0 they actually won.

You can have all those numbers; they are utterly meaningless once the actual games begin.

The true number is 7: the four already won, plus 2006, 2007 & 2011.