Random thoughts on the Chargers game

[themunn]

You don't go for it on 4th and 2 from your own 40 or wherever we were. Almost ever. You don't get it, game is over. You don't even do that in college.

There's VERY strong evidence that NFL coaches should go for it on 4th down a LOT more than they do currently.

http://www.nytimes.com/2014/09/05/upsho ... 0002&abg=1
http://static.espn.go.com/nfl/columns/g ... 53717.html
http://www.sbnation.com/2014/9/8/611961 ... n-4th-down
http://eaglesrewind.com/2013/08/13/nfl- ... 4th-and-1/

No, what you have is statistical evidence that based on the current number of attempts on 4th down extrapolated to a more extensive number, success rate on 4th down would appear to suggest that teams would increase their "expected win rate" by attempting it more.

Of course, one look at the success rate of 4th down conversions tells you all you need to know - 24 teams had a better conversion rate on 4th down than third down in 2013. The "VERY strong" evidence is based on an absurdly high and unsustainable conversion rate (16 teams converted more than 50% of their fourth downs. ZERO teams converted 50% of their third downs).

What you would find is that teams attempt to go for it on 4th down when they have a better chance of making it and little to lose, artificially inflating the statistics. If more teams went for it, conversion rate would regress to the mean and the points added by attempting to convert would result in teams regularly giving away free points to the opposition (attempting to convert a 4th down from your own 30 yard line means you either have another opportunity to drive 70 yards to score... or you give the opposition 3 points for free.

Statistics are only useful if you understand that as soon as you start changing one measure (e.g. frequency of 4th down attempts), they are garbage.
It's like analysing a QB's completion percentage at different parts of the field and hypothesising that since (for example) Brees completes 70+% of his passes thrown everywhere except at the left sideline, where he completes just over 55% and deciding that he could be more effective if he simply stopped throwing the ball to the left sideline. The reality is what we seen against Green Bay - ignore one part of the field and your defenders can key in on the other parts of the field, and completion percentage drops across the board.

You put too much stock in statistics.

 

[grizbob]

You don't go for it on 4th and 2 from your own 40 or wherever we were. Almost ever. You don't get it, game is over. You don't even do that in college.

There's VERY strong evidence that NFL coaches should go for it on 4th down a LOT more than they do currently.

http://www.nytimes.com/2014/09/05/upsho ... 0002&abg=1
http://static.espn.go.com/nfl/columns/g ... 53717.html
http://www.sbnation.com/2014/9/8/611961 ... n-4th-down
http://eaglesrewind.com/2013/08/13/nfl- ... 4th-and-1/

No, what you have is statistical evidence that based on the current number of attempts on 4th down extrapolated to a more extensive number, success rate on 4th down would appear to suggest that teams would increase their "expected win rate" by attempting it more.

Of course, one look at the success rate of 4th down conversions tells you all you need to know - 24 teams had a better conversion rate on 4th down than third down in 2013. The "VERY strong" evidence is based on an absurdly high and unsustainable conversion rate (16 teams converted more than 50% of their fourth downs. ZERO teams converted 50% of their third downs).

What you would find is that teams attempt to go for it on 4th down when they have a better chance of making it and little to lose, artificially inflating the statistics. If more teams went for it, conversion rate would regress to the mean and the points added by attempting to convert would result in teams regularly giving away free points to the opposition (attempting to convert a 4th down from your own 30 yard line means you either have another opportunity to drive 70 yards to score... or you give the opposition 3 points for free.

Statistics are only useful if you understand that as soon as you start changing one measure (e.g. frequency of 4th down attempts), they are garbage.
It's like analysing a QB's completion percentage at different parts of the field and hypothesising that since (for example) Brees completes 70+% of his passes thrown everywhere except at the left sideline, where he completes just over 55% and deciding that he could be more effective if he simply stopped throwing the ball to the left sideline. The reality is what we seen against Green Bay - ignore one part of the field and your defenders can key in on the other parts of the field, and completion percentage drops across the board.

You put too much stock in statistics.

Well, according to this formula you're both wrong...




2 parts vodka plus 1 part OJ plus 1 part cranberry juice = :179422: :lol:

 

[RolandDeschain]

You put too much stock in statistics.

That's a good one. Also, you've got some nice faulty syllogism going on. I really don't have the patience to debate it with you and your mind won't change regardless, so let's just leave it at that, you can walk away thinking you "won" and we'll call it good.

 

[DavidSeven]

I'm a bit surprised we never saw Sherman on Gates. However, we shut down Graham twice last year with LB's and Kam. When you can isolate Kam or Wright like that consistently, that is a matchup we're not always going to win.

Good thoughts as always.

On the quoted point:

I could be wrong, but I think the issue there is that Gates lines up like a true TE, as opposed to playing out of the slot like Graham and occasionally VD. Thus, if you put Sherman on him, you have a skinny corner setting the edge against a team that was more than willing to run. I think SD may have also consciously decided to isolate Gates away from Sherman on certain occasions.

This is a pic I saw on Twitter illustrating that point (Gates lined up near the RT -- Sherm has to move because no WRs lined up on the defensive left side):

 

[Tical21]

I get it. We should always kick onsides kicks, always go for two, let teams score to save time, etc. It doesn't matter if a 4th and 2 is successful 75% of the time. None of that takes situation or team strengths into the equation. If we go for that and don't convert, the game is over, period. You're putting the entire game on converting a 4th and 2. You don't need to do that when you have a great defense and there is still 8 minutes left. Teams with bad defenses still punt the ball there. We're supposed to have the best defense in the game, aren't we? More importantly than any of that, is the message you are sending to them when you don't let them do their job.

 

[themunn]

You put too much stock in statistics.

That's a good one. Also, you've got some nice faulty syllogism going on. I really don't have the patience to debate it with you and your mind won't change regardless, so let's just leave it at that, you can walk away thinking you "won" and we'll call it good.

When you can actually make a point which doesn't start with "DVOA says this", "advanced statistics say this" or "research into these arbitrary figures says this" then maybe.

Or at the very least, if, when someone actually comes out and says "your statistics are wrong" you could man up and actual try to discuss things without saying "BUT BUT BUT THE STATS SAY NO", or "I'm gonna be the bigger man and concede this argument because I don't have a website with some stats to prove you wrong".

 

[Laloosh]

I'm a bit surprised we never saw Sherman on Gates. However, we shut down Graham twice last year with LB's and Kam. When you can isolate Kam or Wright like that consistently, that is a matchup we're not always going to win.

Good thoughts as always.

On the quoted point:

I could be wrong, but I think the issue there is that Gates lines up like a true TE, as opposed to playing out of the slot like Graham and occasionally VD. Thus, if you put Sherman on him, you have a skinny corner setting the edge against a team that was more than willing to run. I think SD may have also consciously decided to isolate Gates away from Sherman on certain occasions.

This is a pic I saw on Twitter illustrating that point (Gates lined up near the RT -- Sherm has to move because no WRs lined up on the defensive left side):

Looked like his biggest catches came out of that formation. Kam, Bobby and K.J. all got burned. Will be interesting to see how we handle it moving forward. I think Kam and the LB's just have to do a better job.

Tackling was pretty horrible too. Seemed liked Kam was trying to knock people out rather than tackle.

 

[RolandDeschain]

When you can actually make a point which doesn't start with "DVOA says this", "advanced statistics say this" or "research into these arbitrary figures says this" then maybe.

Or at the very least, if, when someone actually comes out and says "your statistics are wrong" you could man up and actual try to discuss things without saying "BUT BUT BUT THE STATS SAY NO", or "I'm gonna be the bigger man and concede this argument because I don't have a website with some stats to prove you wrong".

You are blind. You go by points per game which doesn't take opponent skill or opposing defensive difficulty into account at ALL, and claim it's a better figure. You're the one being a stat monkey. You don't even understand it I'm guessing, either. What a joke. This is why I don't want to discuss it with you, because you can't even look at it logically.

I get it. We should always kick onsides kicks, always go for two, let teams score to save time, etc. It doesn't matter if a 4th and 2 is successful 75% of the time.

Yes, THAT'S what I said. :roll:

 

[ivotuk]

I get it. We should always kick onsides kicks, always go for two, let teams score to save time, etc. It doesn't matter if a 4th and 2 is successful 75% of the time. None of that takes situation or team strengths into the equation. If we go for that and don't convert, the game is over, period. You're putting the entire game on converting a 4th and 2. You don't need to do that when you have a great defense and there is still 8 minutes left. Teams with bad defenses still punt the ball there. We're supposed to have the best defense in the game, aren't we? More importantly than any of that, is the message you are sending to them when you don't let them do their job.

I don't believe that for a second. A defense doesn't care if a coach goes for it on 4th down. It could also mean that the coach believes in his defense that if they don't make it on 4th, that the defense can hold.

A coach going for it on 4th down means he thinks the odds are in his favor. Or that he thinks he needs to change something to spark his offense. Pete has done that in the past, with mixed results.

Great take on the gameTical, thanks for posting.

 

[Tical21]

When you can actually make a point which doesn't start with "DVOA says this", "advanced statistics say this" or "research into these arbitrary figures says this" then maybe.

Or at the very least, if, when someone actually comes out and says "your statistics are wrong" you could man up and actual try to discuss things without saying "BUT BUT BUT THE STATS SAY NO", or "I'm gonna be the bigger man and concede this argument because I don't have a website with some stats to prove you wrong".

You are blind. You go by points per game which doesn't take opponent skill or opposing defensive difficulty into account at ALL, and claim it's a better figure. You're the one being a stat monkey. You don't even understand it I'm guessing, either. What a joke. This is why I don't want to discuss it with you, because you can't even look at it logically.

I get it. We should always kick onsides kicks, always go for two, let teams score to save time, etc. It doesn't matter if a 4th and 2 is successful 75% of the time.

Yes, THAT'S what I said. :roll:

Maybe not, but you'll agree that it is all in the same vein, right? If you're going to go for 4th downs, why stop there? These types of things might hold water if you're quite inferior to your opponent, but if you're the better team, you have to believe you'll eventually win if you can keep surviving.

 

[Tical21]

I'm a bit surprised we never saw Sherman on Gates. However, we shut down Graham twice last year with LB's and Kam. When you can isolate Kam or Wright like that consistently, that is a matchup we're not always going to win.

Good thoughts as always.

On the quoted point:

I could be wrong, but I think the issue there is that Gates lines up like a true TE, as opposed to playing out of the slot like Graham and occasionally VD. Thus, if you put Sherman on him, you have a skinny corner setting the edge against a team that was more than willing to run. I think SD may have also consciously decided to isolate Gates away from Sherman on certain occasions.

This is a pic I saw on Twitter illustrating that point (Gates lined up near the RT -- Sherm has to move because no WRs lined up on the defensive left side):

I agree, they rolled out this formation a few times during the game to get Gates isolated. You see Sherm sometimes setting the edge in our cover 3 against different formations when there aren't any receivers on his side, but I can't recall seeing us play against trips like this before. You're right that when we see trips, it is probably an automatic adjustment for Sherman to go to the trips side, even when we're in zone. We might even have a sight adjustment to man coverage. You would have to look at the tape.

Man is what this formation is trying to expose. I agree it would leave Sherman exposed and vulnerable to the run if he stayed on the short side, but when stuff isn't working, I would think about taking some of those risks. Especially on plays like this where it is 2nd and 12. Even if they run, you still have a decent shot at bringing up 3rd and long. I'm a little surprised we never tried it. If we played them again next week and had a chance to prepare for it, I think you'd see Sherman on him in certain situations.

 

[BirdsCommaAngry]

It seems like if a team goes for it more frequently on 4th down, their conversion percentage would be somewhat higher, or possibly much higher, than their typical 3rd down conversion rate because the additional play would tend to create a shorter distance to be needed and the option to punt would prevent attempting to convert less ideal distances. Thus, for an effective enough offense there are times where they really should be going for it when most other coaches would currently not, and for us that 4th and 2 was probably one of them.

What needs to be understood is rule of thumbs like "You don't go for it on 4th down in your own territory until you have to toward the end of a game" are more cultural than factual (i.e. they depend more on popular opinion and all the biases that opinions are subjected to than truth). They do serve a very useful purpose, like protecting less-established coaches from losing their jobs over the negative backlash of failed 4th down attempts that fans aren't accustomed to seeing yet. However, those kind of rules are poor at answering hypothetical questions (Like "Would we win more if we attempted more 4th down conversions in certain situations?") because such rules make no attempt to embrace the unknown or challenge people to succeed where others have failed.

To be as successful as possible, it takes more than following rules of thumb. We should know this by now considering one of our hobbies is currently to watch a franchise run by individuals who have made it their prerogative to shatter some previous rules of thumb ("Short QBs are bad", "Player's coaches can't succeed in the pro's", "When there's a player you want in the draft, you trade up to make sure you get him", "The modern era of football is all about having an elite pocket passing offense", etc.). I hope those whom are questioning the statistical basis behind ideas like going for it more on 4th down are doing so because you're attempting to solicit a more refined argument and not because of some illusion like handling something the way people usually handle it being the best course of action.

 

[Tical21]

If you go for 4th and 2 from your own territory ten times, do you make what, 6 or 7? That is three or four times where you just handed the other team points. Here ya go. Imagine doing that twice in a game.

I don't know that I've ever made a Madden reference when discussing a game, but here goes. People in Madden all the time go for 4th downs, even on their own side, right? You see it a lot with good players, too. You know what, I love it! If you're going to do that, I know I've already got you beat. Because you're not going to get every one. And you're going to hand me the ball and I'm going to score. Then you're going to have to march up the field against me again. You may extend a couple drives. You may even score TD's on drives where you wouldn't otherwise. But those one or two times where I stop you are going to determine the outcome of the game.

 

[Scottemojo]

I don't understand trying to boil the loss down to a 4th and 2 or any single play. Our coaches didn't have a plan for the heat besides they gotta play in it too, and San Diego did have a great plan. Our defense looked slow and like they didn't recognize formations.

It felt like a game we would not win for most of the duration, and frankly I was surprised it stayed close enough to steal. Which is what it would have been, stealing. Because our team was outplayed and out thought. Turns out, we are almost good enough to win that way. But one single 4th and 2 is why we lost? What guarantee do you have that an offense that was struggling to execute in the 4th quarter would have done a damn thing with the other 58 yards they had to go?

 

[themunn]

When you can actually make a point which doesn't start with "DVOA says this", "advanced statistics say this" or "research into these arbitrary figures says this" then maybe.

Or at the very least, if, when someone actually comes out and says "your statistics are wrong" you could man up and actual try to discuss things without saying "BUT BUT BUT THE STATS SAY NO", or "I'm gonna be the bigger man and concede this argument because I don't have a website with some stats to prove you wrong".

You are blind. You go by points per game which doesn't take opponent skill or opposing defensive difficulty into account at ALL, and claim it's a better figure. You're the one being a stat monkey. You don't even understand it I'm guessing, either. What a joke. This is why I don't want to discuss it with you, because you can't even look at it logically.

No, you don't want to discuss it with me because you can't hold a discussion, you avoid ever actually addressing the points that people make by hiding behind an alternative argument. I mean, first you started with

but it's not like a defense can go from horrific to elite in one off-season.

I gave you the example of the Saints (whilst also decrying DVOA as a crock), but you didn't once come back to say "OK, maybe the Saints are a good example". You mentioned their DVOA (again), but said yardage was an awful way to measure defense. Maybe so, but I also mentioned their points per game. They were 32nd in both points and yards allowed in 2011. 4th in both points and yards allowed in 2012. DVOA or no DVOA, that is an elite defense.

I mean, the Chargers were in the position last year that the Saints were in 2011 (32nd in defensive DVOA). 10th is maybe not an elite defense, but then again, elite defenses RARELY allow teams to score 21 points on them at home (it happened to us once in our last 15 home games), but it happens to good defenses often.
Do you think it's possible that a team who drafted a linebacker and cornerback with their first two draft picks, have a fit and healthy Ingram and Freeney (who combined to miss 24 games in 2013) could have improved slightly (not even allowing for their guys that missed the game on Sunday).




And this is my last word on DVOA, if you choose to ignore the above which actually relates to the original discussion (i.e. that the Seahawks offense played well in its opportunities) then feel free to pick at the below (because I know that what actually happens on the field doesn't matter for you, only how it relates to the stats). Then again, maybe your opinion will change since the guys at FO have only moved Seattle's offensive DVOA from 43.3 in week 1 to 38.4. I'd have thought such an atrocious offensive performance against the worst defense of 2013 would have resulted in a massive drop.

In 12 games against common opponents last year the Chargers and Broncos allowed 256 and 290 points respectively. That takes into account opponent difficulty right? In fact, given a few home/away differences it should be the exact same.
So why are the Broncos the 15th ranked defense by DVOA and San Diego the 32nd?

DVOA/DYAR is the same system that says that Anquan Boldin was a better receiver in 2013 than Josh Gordon, Kaepernick was better than Wilson, and Rashad Jennings and Steven Ridley were both (apparently) significantly better than Peterson and Gore last year. It's the same system that said that Chicago's defense in 2012 was better than ours was not only in 2012... but better compared to the league average in 2012 than our own was in 2013! Can you imagine! A defense that couldn't stop a rookie QB driving 90+ yards and 80 yards up the field in the 4th quarter and overtime! The 6th best defensive DVOA of all time!

It doesn't pass the eyeball test. Neither does points per game or yardage for that matter, but at least they are simple, quantifiable and non-arbitrary.
Anybody who had watched both Denver and Seattle play last year could have predicted what would happen in the Superbowl, yet Denver's offensive DVOA was 7.5% greater than average than Seattle's defense. Their offense was "33% better than the average offense" despite playing just 3 teams with defenses above average (by their own statistical measure). It's a nonsense statistic made up to satisfy people who are easily convinced by fancy numbers, based on a system developed by 2 sports writers and a "sports statistician" whose majority contribution to sport is all to baseball.


Edited for language-Mizzou

 

[kearly]

Danny Woodhead is a freakin' gnat. Dude just wouldn't stop getting 6 yard gains when they needed them.

Agreed on all points but especially this. Danny Woodhead is the most under-appreciated RB in the NFL, IMO.

 

[kearly]

You don't go for it on 4th and 2 from your own 40 or wherever we were. Almost ever. You don't get it, game is over. You don't even do that in college.

There's VERY strong evidence that NFL coaches should go for it on 4th down a LOT more than they do currently.

http://www.nytimes.com/2014/09/05/upsho ... 0002&abg=1
http://static.espn.go.com/nfl/columns/g ... 53717.html
http://www.sbnation.com/2014/9/8/611961 ... n-4th-down
http://eaglesrewind.com/2013/08/13/nfl- ... 4th-and-1/

I'm a fan of going for it on 4th down too, but I don't think those articles consider that it was a one score game with the ball being on Seattle's own forty with 9 minutes to go. There are sites that track win probability, I would assume that punting would have the higher win probability outcome due to the factors at hand.

If it had been with 4 minutes to go, or if it had been on the other 40, it would be a closer call.

Pete has an advanced analytics guy on his staff so I'm guessing he knows the percentages on all these scenarios the same way an elite Poker player would.

 

[RolandDeschain]

Maybe not, but you'll agree that it is all in the same vein, right? If you're going to go for 4th downs, why stop there? These types of things might hold water if you're quite inferior to your opponent, but if you're the better team, you have to believe you'll eventually win if you can keep surviving.

All in the same vein to a limited extent, I would say. I also agree with your overall point about eventually winning if you can keep surviving. My question now is, how is that relevant with eight minutes to go in the 4th quarter, down by six, when you've watched your defense get smoked all day long? It's an unusual situation we found ourselves in, defense-wise, and I just can't believe we decided to go conservative there. *shrug*

Doesn't matter if your defense allowed 0 points per game for the last 25 straight games, if they're getting smoked that day, you have to look at THAT way more than what they've done in the past.

 

[Russ Willstrong]

Maybe not, but you'll agree that it is all in the same vein, right? If you're going to go for 4th downs, why stop there? These types of things might hold water if you're quite inferior to your opponent, but if you're the better team, you have to believe you'll eventually win if you can keep surviving.

All in the same vein to a limited extent, I would say. I also agree with your overall point about eventually winning if you can keep surviving. My question now is, how is that relevant with eight minutes to go in the 4th quarter, down by six, when you've watched your defense get smoked all day long? It's an unusual situation we found ourselves in, defense-wise, and I just can't believe we decided to go conservative there. *shrug*

Doesn't matter if your defense allowed 0 points per game for the last 25 straight games, if they're getting smoked that day, you have to look at THAT way more than what they've done in the past.

You bring up good points Roland and it's too bad so many conservative plays happen when coaches are fearful of making unpopular calls or shy from a bold play because of statistics. The issue with quoting previous statistics for 4th down gambles nowadays is that new rules favor the offense so much more and offenses with so many threats don't even have to be run a play to draw a defenders off sides on 4th and 2. You could see right away RW starts walking toward the sideline signaling to PC that they should go for it.
Our situation was that we had 4th and 1.5 yds on our 37 yd line I think. Marshawn was averaging over 5 yds/carry. We had the threat of Percy for fly sweep. On such crucial downs we know that defenders are being watched closely for defensive pass interference and illegal contact besides the usual jumping offsides. We could have run Percy in motion just as we do on a fly sweep whether or not we hike the ball. Using hard counts draw the defense off sides or jump before hiking it. If no defender jumps off sides you still have a Marshawn and RW zone option once Harvin resets.
We had so many options and PC is usually gutsy enough to make such a call when RW points to go for it. What was eating at PC also was that the defense was tired and not making many plays. IMO they didn't go for it probably because they weren't prepared to call a 4th down play and rather than burning a time out to consider it we punted.

 

[RolandDeschain]

Yeah, we're just left to guess at why, and hindsight is always perfect, I totally get that - but I absolutely detested the decision from the first. If some don't want to believe me on it, so be it. Things were getting desperate at that point because our defense was doing its best wet paper bag impression all day, and counting on that to change just seems like the wrong call no matter what.