NFR Nate Silver predicts Seattle and New England in the Super Bowl

#16
While that might explain why the Redskins didn't score another TD after their first two, it doesn't explain why the Hawks scored 24 unanswered points afterwards.

K
Ummmmm, huh? They would have scored more but they wouldn't have?

If cousins had come out in the second half it would have been a different game. That said nobody want to face them right now, kinda like the Giants have done the last couple years.
 

Kent Lufkin

Remember when you could remember everything?
#17
Ummmmm, huh? They would have scored more but they wouldn't have?
Perhaps I mis-wrote. I meant to say that if RGIII hadn't been injured, the Redskins would conceivably scored more points. But since he left the game injured, they didn't. Before his departure, he was definitely a factor on the offensive side of the ball.

But even if he had remained in the game, since he didn't play defense as well as offense, aside from being a psychological letdown for the entire team when he left, his presence one way or another wasn't a factor in preventing the Hawks from scoring 24 unanswered.

K
 
#18
Gotcha.

Personally I think ATL is overrated and you guys will blow through them. I think we ate gonna have a tough one against a scorned Rodgers. Funny how the 2 best QBs in the NFL are from SF and neither play for the Niners.

Best weekend in football!
 

Lugan

Joe Streamer
#22
Speaking as someone who is professionally handy with statistics, I'm surprised Nate didn't admit that predicting elections and predicting the winner of a sports event are like comparing apples to burritos.

An election unfolds slowly over time, across 100+ million people all doing a single very similar task (choosing candidates and voting for those candidates). The important thing in elections is that the 100+ million voters' actions all happen in parallel - meaning, they are not dependent on each other. Sure, individuals are influenced by each other and the crowd's choices ebb and flow together somewhat, but not that much. This makes predicting elections a more systematic endeavor, and thus Nate was able to produce a prediction about it with a high degree of statistical confidence. Near the end of the election cycle, Nate was up to something like 93% chance of Obama winning, and he got all but one state's results correct.

Games (or in the playoffs, a series of a few games) are entirely different. Games are a linear series of plays with a small number of players on the field on each side. The actions of players are somewhat predictable based on past player and team statistics, but there is still huge variation in the success and failure of those actions: Did the QB complete the 40 yard pass, or did the receiver have butter fingers? Did the linebacker sack the QB, or did he get away to keep the drive alive? Did the punt returner hold onto the ball when tackled, or did he fumble at his own 10 yard line? Etc. Lots of random stuff happens, each of those random things effects the ensuing line of events, and that makes prediction hard. Plus, the statistical sampling is just far smaller than 100+ voters. That's not to say Nate (and the guys on ESPN, etc.) can't make predictions based on some data, but it does mean that getting to anything like the statistical confidence of an election prediction is impossible. In the end, I'm guessing Nate's confidence underlying his Super Bowl prediction is very low.
 

jimmydub

Active Member
#23
An election unfolds slowly over time, across 100+ million people all doing a single very similar task (choosing candidates and voting for those candidates). The important thing in elections is that the 100+ million voters' actions all happen in parallel - meaning, they are not dependent on each other. Sure, individuals are influenced by each other and the crowd's choices ebb and flow together somewhat, but not that much. This makes predicting elections a more systematic endeavor, and thus Nate was able to produce a prediction about it with a high degree of statistical confidence. Near the end of the election cycle, Nate was up to something like 93% chance of Obama winning, and he got all but one state's results correct.

Games (or in the playoffs, a series of a few games) are entirely different. Games are a linear series of plays with a small number of players on the field on each side. The actions of players are somewhat predictable based on past player and team statistics, but there is still huge variation in the success and failure of those actions: Did the QB complete the 40 yard pass, or did the receiver have butter fingers? Did the linebacker sack the QB, or did he get away to keep the drive alive? Did the punt returner hold onto the ball when tackled, or did he fumble at his own 10 yard line? Etc. Lots of random stuff happens, each of those random things effects the ensuing line of events, and that makes prediction hard. Plus, the statistical sampling is just far smaller than 100+ voters. That's not to say Nate (and the guys on ESPN, etc.) can't make predictions based on some data, but it does mean that getting to anything like the statistical confidence of an election prediction is impossible. In the end, I'm guessing Nate's confidence underlying his Super Bowl prediction is very low.
The cool thing about the advanced metrics in football is that they do seem to be showing patterns. The most overrated team in football this season was Indianapolis, despite their winning record. They went out in a hurry. The Hawks are the best team to end a season over the last decade based on weighted DVOA, and were among the best teams in the league this year even before they started annihilating opponents. DVOA has been set up to account for variables that otherwise don't show up in box scores or player cards. When Marshawn Lynch fumbled at the goal line, his DYAR took a serious nosedive, as did the DVOA for the rest of the team. It wouldn't have been as big of an impact if it were nearer the fifty yard line. Those kinds of things are accounted for, and show up in the advanced stats. Any time the offense fumbles the ball, it goes against their DVOA.

While there are certain things that DVOA doesn't account for, it does measure a team's efficiency very well. When it comes to evaluating a team and how they'll perform under pressure, that's when good old fashion football culture takes over. DVOA can't predict how players play each other, but the Seahawks have played a really good sampling of games as far as different schemes are concerned. It's not a perfect system, but it's the best I've found.

If you watch the video of him on ESPN, he actually says he would bet on the Seahawks based on the line that Vegas is putting out for the Super Bowl. That would tell me he does have confidence in the outcome, between the statistics he's viewing and the results he's seeing on the field.
 
#24
Hahahaha!! I would like to see the Hawks win it... My dad told everybody that if they win, he will shave his head and beard (always had hair on his head, I have only seen him with out a beard twice in 29 years).
We are guessing if he does it, it wont come back...
 
#30
The cool thing about the advanced metrics in football is that they do seem to be showing patterns. The most overrated team in football this season was Indianapolis, despite their winning record. They went out in a hurry. The Hawks are the best team to end a season over the last decade based on weighted DVOA, and were among the best teams in the league this year even before they started annihilating opponents. DVOA has been set up to account for variables that otherwise don't show up in box scores or player cards. When Marshawn Lynch fumbled at the goal line, his DYAR took a serious nosedive, as did the DVOA for the rest of the team. It wouldn't have been as big of an impact if it were nearer the fifty yard line. Those kinds of things are accounted for, and show up in the advanced stats. Any time the offense fumbles the ball, it goes against their DVOA.

While there are certain things that DVOA doesn't account for, it does measure a team's efficiency very well. When it comes to evaluating a team and how they'll perform under pressure, that's when good old fashion football culture takes over. DVOA can't predict how players play each other, but the Seahawks have played a really good sampling of games as far as different schemes are concerned. It's not a perfect system, but it's the best I've found.

If you watch the video of him on ESPN, he actually says he would bet on the Seahawks based on the line that Vegas is putting out for the Super Bowl. That would tell me he does have confidence in the outcome, between the statistics he's viewing and the results he's seeing on the field.
Man if this was the process going through my head when I watched a game for entertainment then the fun would be gone!

I guess if I was a "wise guy" in Vegas looking for an edge this would be what I was looking for....