NFL 2015 Part 2: There's a Flag Down

[LuckyNumber7]

I'll try to dig up some references for you, but I'm pretty sure that Burke's research shows that using probabilities from the current game does a worse job of predicting future results in that game than using league-average results (which is what the model uses). In other words, using current-game probabilities (or making adjustments for such) is another small sample size mistake.

Also, saying that the Patriots had the ability to 'easily' get the ball back seems really optimistic to me.

I'm by no means totally discounting odds from a model, I just really don't think it does it justice. It seems to me as though those odds were good for "any game on any matchup." Meaning, a compiled statistic of all teams with that remaining time and that current score.

On the idea of using current in-game stats... you don't have to run into small sample errors from doing that. Instead, just use the splits, which could be conditionally set from the in-game stats. Playing Denver is a big deal. A limping Denver offense and a solid defense were big deals.

 

[speedracer]

I'm by no means totally discounting odds from a model, I just really don't think it does it justice. It seems to me as though those odds were good for "any game on any matchup." Meaning, a compiled statistic of all teams with that remaining time and that current score.

On the idea of using current in-game stats... you don't have to run into small sample errors from doing that. Instead, just use the splits, which could be conditionally set from the in-game stats. Playing Denver is a big deal. A limping Denver offense and a solid defense were big deals.

Yes, the model computes probabilities from historical outcomes of all games with given score-down-distance states.

I think Burke has tried to make adjustments based on team strength and has found that it just doesn't make that much of a difference. E.g. if the average defense stops 2-point conversions at a 55% rate, Denver's defense might stop them at a 60% rate or something like that. Over the course of a game those percentages compound to make Denver's defense elite, but over 6 minutes it won't make nearly as much of a difference.

 

[speedracer]

What a horrible coaching decision by Arians to have Palmer throw

Quoted for truth.

 

[LuckyNumber7]

Yes, the model computes probabilities from historical outcomes of all games with given score-down-distance states.



I think Burke has tried to make adjustments based on team strength and has found that it just doesn't make that much of a difference. E.g. if the average defense stops 2-point conversions at a 55% rate, Denver's defense might stop them at a 60% rate or something like that. Over the course of a game those percentages compound to make Denver's defense elite, but over 6 minutes it won't make nearly as much of a difference.

It's not just that, though. Three main unaccounted for factors:

- Denver's unrelenting end zone D
- Tom Brady's ability to move the ball down the field
- Denver's inability to get their offense going anywhere

Those are three big factors, that stray very far from the norm. And it proved to be true, in this case.

 

[LemonMelon]

There's a reason why redzone percentages are listed separately. The Pats were driving with ease only to sputter out inside the 20. Thing is, they could have won the game with three relatively high percentage FGs. Actually punching it in was far more difficult and, as it so happened, unnecessary.

I'm not surprised Belichick went for it though, he's always been the type to go for it with the assumption that historical probability would work in his favor. In this case, it didn't, because they were being dominated. The Broncos were shutting them down again and again in short yardage situations (13% on third down!!) while allowing the occasional long play that kept the Pats alive.

 

[speedracer]

It's not just that, though. Three main unaccounted for factors:

- Denver's unrelenting end zone D
- Tom Brady's ability to move the ball down the field
- Denver's inability to get their offense going anywhere

Those are three big factors, that stray very far from the norm. And it proved to be true, in this case.

All teams play better defense in the end zone. All offenses play better in midfield when given four downs. I assure you that Burke knows these things. I would tend to doubt that the Broncos D or the Patriots O are anomalous in this regard.

Most teams play somewhat predictably (and as a result inefficiently) late in the fourth quarter when nursing a lead. That's built into Burke's model as well, although perhaps you could convince me that Gary Kubiak is an outlier in this regard.

 

[mikal]

One thing that hasn't been accounted for in this discussion is that sometimes the coaches go off of the personnel on the field. Denver was going without their two starting safeties on at least two of those drives. Another factor is maybe there was a match up that they thought they could take advantage of or a play that they really thought would be successful. There's much more to the decision-making then just charts.


Sent from my iPhone using Tapatalk

 

[speedracer]

I'm not surprised Belichick went for it though, he's always been the type to go for it with the assumption that historical probability would work in his favor. In this case, it didn't, because they were being dominated.

Just last week a team outscored its opponent 31-0 in one half and got outscored 24-0 in the second half. It's generally not too helpful to put so much stock in the events of one game.

 

[Headache in a Suitcase]

Death. Taxes. Carson Palmer absolutely shitting the bed in a playoff game.

 

[MrPryck2U]

I would have rather seen Pats/Panthers than Broncos/Panthers, but I'll be rooting for Peyton to win another SB and ride off into the sunset. I would think that Carolina will be favored to win. Whoever wins, I just want a close game. No 44-8 blow outs.

 

[LemonMelon]

I probably wouldn't have watched at all if it had been Pats/Seahawks, so I've been satisfied with any possible outcome for a week or so now.

Hoping Denver wins so the Eli > Peyton truthers will have no ammo left.

 

[MrPryck2U]

I probably wouldn't have watched at all if it had been Pats/Seahawks, so I've been satisfied with any possible outcome for a week or so now.

Hoping Denver wins so the Eli > Peyton truthers will have no ammo left.

I call double BS on this. You would have watched the Super Bowl regardless of the match-up. It's still the Super Bowl. Where you planning on going shopping at the mall for 6 hours? Good Lord even. The "truthers" ? would still have a debate. Eli is 2 for 2. If Peyton wins, he'll be 2 for 5. The only way he'll be "better" than Eli is to win 3 Super Bowls.

 

[LuckyNumber7]

I call rebuttal BS. If a QB can make it to 5 Super Bowls, and only win two, and have better career numbers, he's better than a guy who only made it to two Super Bowls and won both.

 

[MrPryck2U]

I call rebuttal BS. If a QB can make it to 5 Super Bowls, and only win two, and have better career numbers, he's better than a guy who only made it to two Super Bowls and won both.

There's no doubt that Peyton has accomplished more in the NFL and has more wins, BUT Eli is perfect in the Super Bowl and Peyton has lost twice. Now, naturally, I'm just having some good natured fun with LM, but Eli has achieved more success in the Super Bowl than his brother. A victory in 2 weeks will only equal Eli's success. Another famous Denver QB was 2 for 5 in the SB, but he managed to save his best for last.

 

[LemonMelon]

If we were talking about golf, tennis or boxing, championships would be a reasonable measure of greatness. Since we're talking about football, a team sport, one could use the Giants' consistent playoff absences against Eli just as easily as you could use the rings in his favor.

But please, feel free to argue that Trent Dilfer and Brad Johnson are better QBs than Dan Marino and Fran Tarkenton. I'd like to see that.

 

[MrPryck2U]

Nobody should argue that, but the fact remains that Eli has achieved more Super Bowl success than his brother. Marino and Fran are legends, of course. But, a thorn in their side is the fact that neither has won a Super Bowl. Especially Fran who was 0 for 4. Same thing with Jim Kelly.

 

[bono_212]

I call double BS on this. You would have watched the Super Bowl regardless of the match-up. It's still the Super Bowl.

It's stupidly easy to watch something else instead of the Super Bowl. I've done it at least twice since I started watching regularly because of match ups I just had zero interest in. You just turn the TV off, or watch a movie. It's not like the beam it into your brain or something.

I call rebuttal BS. If a QB can make it to 5 Super Bowls, and only win two, and have better career numbers, he's better than a guy who only made it to two Super Bowls and won both.

Thank you, Jesus, this is one of the stupidest arguments in sports history.

 

[LuckyNumber7]

Nobody should argue that, but the fact remains that Eli has achieved more Super Bowl success than his brother. Marino and Fran are legends, of course. But, a thorn in their side is the fact that neither has won a Super Bowl. Especially Fran who was 0 for 4. Same thing with Jim Kelly.

Look at it this way:

Peyton Manning has 5 conference championships. Eli has 2. Peyton Manning, over his career, has been able to get to the super bowl twice as much as Eli has through his career.

It's almost like that doesn't count for anything, by your definition of success. If Peyton wins this super bowl (big if...), Eli doesn't have anything on him whatsoever.

 

[LemonMelon]

I would honestly rather watch a couple of movies than the Patriots, Seahawks and Coldplay.

Luckily, the game on tap is pretty interesting. It's a matchup we haven't seen yet and there are some interesting narratives here.

 

[phanan]

I think the Broncos will get destroyed. At least the Patriots or Steelers would have made it more interesting.

 

[LemonMelon]

Nah. The league's #1 defense isn't going to get destroyed by the league's #1 offense. Didn't happen to the Seahawks in Super Bowl 48, didn't happen to the Giants in Super Bowls 42 and 46 and it's not going to happen now.

The key matchup of the game will be the Panthers' OL vs. the Broncos' DL. If Cam gets 10-15 seconds to throw on every pass like last night, it will be a one-sided game. If Denver can wreck Cam the way they did Brady, the Broncos will win.

 

[MrPryck2U]

Look at it this way:

Peyton Manning has 5 conference championships. Eli has 2. Peyton Manning, over his career, has been able to get to the super bowl twice as much as Eli has through his career.

It's almost like that doesn't count for anything, by your definition of success. If Peyton wins this super bowl (big if...), Eli doesn't have anything on him whatsoever.

I'm talking SUPER BOWL SUCCESS! I've emphasized that numerous times. I have already said that Peyton has accomplished more in the NFL and won more games. BUT, Eli was won more Super Bowls, therefore he has achieved more success in the SUPER BOWL than his brother.

 

[deep]

If Denver can wreck Cam the way they did Brady, the Broncos will win.

but Cam is (way) better than Brady, so Panthers win, not a close game.

 

[GirlsAloudFan]

Let me get some of what you're smoking, deep.

 

[MrPryck2U]

but Cam is (way) better than Brady, so Panthers win, not a close game.

Damn, now I'm really pulling for Denver. If the Panthers don't win, they'll be the '98 Vikings all over again. Cam is very good, but he will have a long off season if the Panthers don't win. Though not as bad an off season as Brady after the 2007 season.

 

[LemonMelon]

but Cam is (way) better than Brady, so Panthers win, not a close game.

Nobody and I mean nobody plays consistently good football with a defender in their face. Maybe they can pull off a highlight reel throw once in a while, but you can't build an offense around that kind of pressure. The Panthers' OL completely dominated last night and if they do the same again, they'll win in a laugher, but they are about to face the league's best pass rush.

I see the Panthers winning because they're probably the league's most balanced team, but you're ignoring history if you think a battle tested defense is about to get blown out on the big stage. That hardly ever happens.

Another important stat: the Panthers are currently sitting at +20 in turnover margin. That's a formidable stat and it helps to explain their dominance, but it's not something to rely on. The Chiefs were +15 and riding high, but they lost in New England with a -1 game. If the Broncos can be disciplined with the ball, that will affect the Panthers' efficiency in every aspect.

 

[Hewson]

Look at it this way:

Peyton Manning has 5 conference championships. Eli has 2. .

Peyton only has 4 conference championships.

 

[phanan]

Nah. The league's #1 defense isn't going to get destroyed by the league's #1 offense. Didn't happen to the Seahawks in Super Bowl 48, didn't happen to the Giants in Super Bowls 42 and 46 and it's not going to happen now.

The key matchup of the game will be the Panthers' OL vs. the Broncos' DL. If Cam gets 10-15 seconds to throw on every pass like last night, it will be a one-sided game. If Denver can wreck Cam the way they did Brady, the Broncos will win.

The Broncos were overmatched two years ago and had a better offensive team. I don't see it being much better this year, even with a better defense.

 

[LemonMelon]

The Broncos were overmatched two years ago and had a better offensive team. I don't see it being much better this year, even with a better defense.

They've gone from 22nd in scoring D in 2013 to 4th in 2015 and just finished smothering two of the best offenses in football. Barring 5-6 takeaways, I don't know how Carolina intends to score enough points for a blowout.

I'm calling 21-17 Carolina, with a failed onside kick sealing the deal.

 

[DaveC]

Those were some spicy #hottakes, LM.