Stopping the New Romo Myth

[TheDude]

I don't disagree with you, Adam, or PH. Two are speaking about statistics and you are looking at specific variables that 'bent' the game. If you marry the two then you get a more secular view.

The reply probably isnt to you, but I am heading out into 1 1/2 hours of traffic screwing around on this

So when you are at a casino, Im sure many have seen the history of of roulette spins. This is posted to drive the sucker/trend bet. Say black has come up 5 times in a row. The odds of that occurring (std 38 american numbers) are 47.3684%*47.3684%*47.3684%*47.3684%*47.3684% or 2.3848%. So many compute that red is a sure bet. However, the odds of black on the next spin is remains at 47.3684%. The next play is agnostic to preceding history. so fine over time, the law of large numbers should hold and it smooths. That is why there are table limits so whales can't infinitiely bank roll loss runs like this waiting for the law of large numbers to return (i.e. bet $100-lose, bet $200 next - lose, bet $400 lose, etc.)


Same thing with a penny. If I flip heads 10 times in a row, that was interesting and a 0.098% chance. But the next flip is still 50/50. The history doesn't change the near term results.

In football, Once a team gets a first down, everything resets. If the start on the 20, the probability of scoring is low, get a first down to the 30, odds increase. Once you are ending a drive on a final set of downs, the next play (or set of plays) is not beholden to ANY history of the drive.

This is a standard misconception of stats and idiosyncratic predictability. To prove with data I took all plays in the nfl from 1998 to 2014. I looked at the result of drives that's last line of scrimmage was inside the opponents 28 yard line for 3 sets:

• All drives
• Only drives that started inside the the driving teams 20 yard line
• Drives that started inside the opponents 40 yard line

HIghlights

• An offensive score resulted > 81% of the time
• TD % was highest for drives starting inside the teams own 20. (average starting field position is the 16 yard line)
• A blocked FG is exactly 0.3% in all cases

Summary
Leaving a situation that results in a score 80% with no points is similar to leaving a runner on third with no outs. It is agnostic to whether the shortstop made an error or if the defense allowed an opposing QB a higher rating than normal.

Washing situations like these away in football in "the offense played fine" ultimately means the team leaves 80% opportunities for a score on the table every week on average.

The per drive stat is interesting over a history, but CANNOT be assumed to hold for a given game. especially if the focus is on start vs finish

 

[percyhoward]

The fact that the drive result maintained an average is irrelevant.

Certainly not as relevant as the average itself, no.

Your average includes this wasted drive. It is calculated based on historical drives. HIstory has no bearing on the next play

The offense had opportunities to score more than its average, and didn't seize them. The offense also had opportunities to score less than its average, but that didn't happen either. A lot of things could have happened, but what did happen was that the offense matched its average. The defense had chances to hold the Packers below their average, and didn't seize them. The defense had chances to hold the Packers right at their average, and that didn't happen either. What did happen was that the Packers scored higher than their average.

The offense matched its average, while the defense did not hold the Packers to theirs. Any discussion of specific plays that involved the fault of either the offense or defense -- and that are reflected in the score -- is anecdotal.

Dallas didn't need to score at a normal rate - it need higher. History also points to 21 points not being enough.

But we're comparing the Dallas offense to the Dallas defense, and saying the offense could have scored more points doesn't make the defense's performance any better than it was.

 

[jobberone]

The reply probably isnt to you, but I am heading out into 1 1/2 hours of traffic screwing around on this

So when you are at a casino, Im sure many have seen the history of of roulette spins. This is posted to drive the sucker/trend bet. Say black has come up 5 times in a row. The odds of that occurring (std 38 american numbers) are 47.3684%*47.3684%*47.3684%*47.3684%*47.3684% or 2.3848%. So many compute that red is a sure bet. However, the odds of black on the next spin is remains at 47.3684%. The next play is agnostic to preceding history. so fine over time, the law of large numbers should hold and it smooths. That is why there are table limits so whales can't infinitiely bank roll loss runs like this waiting for the law of large numbers to return (i.e. bet $100-lose, bet $200 next - lose, bet $400 lose, etc.)


Same thing with a penny. If I flip heads 10 times in a row, that was interesting and a 0.098% chance. But the next flip is still 50/50. The history doesn't change the near term results.

In football, Once a team gets a first down, everything resets. If the start on the 20, the probability of scoring is low, get a first down to the 30, odds increase. Once you are ending a drive on a final set of downs, the next play (or set of plays) is not beholden to ANY history of the drive.

This is a standard misconception of stats and idiosyncratic predictability. To prove with data I took all plays in the nfl from 1998 to 2014. I looked at the result of drives that's last line of scrimmage was inside the opponents 28 yard line for 3 sets:

• All drives
• Only drives that started inside the the driving teams 20 yard line
• Drives that started inside the opponents 40 yard line

HIghlights

• An offensive score resulted > 81% of the time
• TD % was highest for drives starting inside the teams own 20. (average starting field position is the 16 yard line)
• A blocked FG is exactly 0.3% in all cases

Summary
Leaving a situation that results in a score 80% with no points is similar to leaving a runner on third with no outs. It is agnostic to whether the shortstop made an error or if the defense allowed an opposing QB a higher rating than normal.

Washing situations like these away in football in "the offense played fine" ultimately means the team leaves 80% opportunities for a score on the table every week on average.

The per drive stat is interesting over a history, but CANNOT be assumed to hold for a given game. especially if the focus is on start vs finish

Nice work. I understand the statistics above well as a reluctant student. I don't think anyone of reason would argue with you about the last set of stats so I'm not sure where you want to go with this. I wouldn't be surprised I just missed it as I can't read it again. If you get the same number of possession and your Sc% and RZ% is less than your opponent then you may very well lose.

Anyone my wife is going to get me as we have company so you enjoy your drive and I'll play nice at home.

 

[ufcrules1]

But we're comparing the Dallas offense to the Dallas defense, and saying the offense could have scored more points doesn't make the defense's performance any better than it was.

Right, they could have BOTH played better. Not just the defense. You act like the offense couldn't have done any better and that is just not true. Rodgers
went above and beyond which is what a champion does. Not because our average defense sucks but because he did something special out there with their backs up against the wall and the season on the line. That is why he is the best QB in the NFL.

Nobody is saying the defense didn't struggle, but to point at the defense and blame them solely for the loss just isn't fair in my opinion.

 

[percyhoward]

Leaving a situation that results in a score 80% with no points is similar to leaving a runner on third with no outs. It is agnostic to whether the shortstop made an error or if the defense allowed an opposing QB a higher rating than normal.

Washing situations like these away in football in "the offense played fine" ultimately means the team leaves 80% opportunities for a score on the table every week on average.

The per drive stat is interesting over a history, but CANNOT be assumed to hold for a given game. especially if the focus is on start vs finish

You've got the tail wagging the dog. The wasted opportunity there was just a part of this overall performance that you're ostensibly judging -- and doesn't characterize it at all. How does an event which isn't at all representative of the larger sample become so important in your mind?

 

[ufcrules1]

The point is that our offense played much better than our defense did in that game. So the defense should get more of the blame.

I don't think our opinions are far apart. I think it is simple really, the offense could have played better and the defense could have played better. Wouldn't that be considered a team loss? If someone wants to take the stance of the offense having zero blame, it would be a shortsighted position in my mind from watching the game.

 

[ufcrules1]

And again, that's pretending the overturn of Dez's catch was correct.

That was a catch by Dez IMO however, I think it was a low % play in the first place. Perhaps the smarter play on 4th down and the season on the line would have been to throw a nice safe pass to Beasley who was wide open and would have easily picked up the first down. Then we could have continued the drive all while eating up clock thus putting GB in an extremely bad position. Romo could have then been the hero that Rodgers turned out to be.

 

[percyhoward]

To point at the defense and blame them solely for the loss just isn't fair in my opinion.

I think you're overstating the other side's argument there.

 

[AdamJT13]

I don't think our opinions are far apart. I think it is simple really, the offense could have played better and the defense could have played better. Wouldn't that be considered a team loss? If someone wants to take the stance of the offense having zero blame, it would be a shortsighted position in my mind from watching the game.

Every loss is a team loss. But some losses are more the fault of the offense, some are more the fault of the defense, and the special teams might even be the biggest culprit sometimes.

 

[ufcrules1]

I think you're overstating the other side's argument there.

Well, that is how I see it coming across. Feel free to correct me since you are on the other side. I could be reading into it wrong but from what I'm gathering
you think the offense played perfectly fine but the defense had a horrible game, thus giving all the blame to our defense.

If your stance is the offense played "ok" but didn't put up enough points (only 21 which is almost never enough which has been pointed out) and could have played better, then we would be on the same page. I can freely admit the defense could have played better but I simply can't give them all the blame. Maybe 60/40 on the blame scale. What is your blame scale numbers?

 

[ufcrules1]

Every loss is a team loss. But some losses are more the fault of the offense, some are more the fault of the defense, and the special teams might even be the biggest culprit sometimes.

Agreed, and this game was more the fault of the defense but how much of the blame does the offense get? That is what the debate has been for the last few pages. You and PH have provided stats basically saying ..."Hey, the offense shouldn't have any blame based on these stats".

 

[percyhoward]

I can freely admit the defense could have played better but I simply can't give them all the blame. Maybe 60/40 on the blame scale. What is your blame scale numbers?

I got offense 10, defense 40, Blandino 50.

 

[AdamJT13]

Perhaps the smarter play on 4th down and the season on the line would have been to throw a nice safe pass to Beasley who was wide open and would have easily picked up the first down.

Hindsight is 20/20 (or 50/50, if you're Cam Newton). That was a pre-snap read because of the extra blitzers. There was no way to anticipate that Beasley would be open (he had a corner and safety right over him), nor to expect to have the time to scan the field for open receivers. He knew for sure, however, that Dez would be one-on-one -- and that if he put the ball up, Dez more than likely would catch it.

 

[TheDude]

You've got the tail wagging the dog. The wasted opportunity there was just a part of this overall performance that you're ostensibly judging -- and doesn't characterize it at all. How does an event which isn't at all representative of the larger sample become so important in your mind?

we are talking past each other. I've laid out the statistical argument above.

Your position is
- 8 drives each were how the game unfolded - drive numbers seem to be unable to be affected by either teams strategy
- the offense scored 21 points on 3 of 8 drives which is inline with some seasonal average against all opponents and all situations.
- How the failed drives end and the time or situation of the game of the failed drive is irrelevant because 3/8 is the average (or 21pts/8) and because the average was met, they blame lies mostly somewhere else.
- The end of drives irrelevant even if it is a turnover on your side of the field because the offense has already done enough.

My position is pretty simple
- 21 points doesnt win playoff games - provided only 16 of 99 games since 2006
- No drive result is in any way shape or form dependent on on past drive history.
- When there is a situation in the game, and every possession is precious in an elimination game, not capitalizing on a an opportunity that has resulted in a score 80% of the time is a wasted opportunity.
- Not scoring is one issue, but the call and result enable GB an extra possession before the half to score 3 points....a 6 point swing.

I'm fine saying the defense didnt do a stellar job, but the game plan was about keeping the DALLAS offense on the field and GB off the field because of Rogers. The strength of the team was the offense, therefore, blowing scenarios is less expected from that side than the defense.

The wagging the dog is accepting that a specific drive outcome is justified if it moves toward a historical average. A situation could have a 99% historical scoring rate, but in your argument the the offense is absolved of failing on that one drive because the 1% isnt the issue its the 2.7 pts per drive that must be maintained according to your argument.

It relates to the total points a team scores in a game.

The 26-21 score seemed low for both teams because there were only a total of 16 possessions in the game. There would normally be 21 for these two teams. So if you're going to compare their score in this game to their average score, you need to compare their number of possessions in this game to their average number of possessions.

I dont want to compare any pts/possessions I want the team to score points when the opportunity is there because points are precious and win games.

Otherwise, you're expecting them to score at higher than their normal rate in this game.

.

Before the game, but most assuredly I do if, as the game unfolds on the field, the 5 failed drives end deep in GB territory and each of the 5 fails had has yielded some 80% success in the entire nfl over 17 years, then yeah failing to score in over 50% of the possessions in that scenario is reason to challenge not absolve.

 

[AdamJT13]

I got offense 10, defense 40, Blandino 50.

Special teams, too, for not getting the field goal before the half. If we make that, the Packers likely don't score before halftime, and we're up 17-7 instead of 14-10.

 

[TheDude]

Special teams, too, for not getting the field goal before the half. If we make that, the Packers likely don't score before halftime, and we're up 17-7 instead of 14-10.

Bingo. of course the botched snap and false start played a role in making it a 50 yd FG but exactly right

Team loss

 

[big dog cowboy]

Special teams, too, for not getting the field goal before the half. If we make that, the Packers likely don't score before halftime, and we're up 17-7 instead of 14-10.

JMO but that is where the game was truly lost.

 

[gimmesix]

Special teams, too, for not getting the field goal before the half. If we make that, the Packers likely don't score before halftime, and we're up 17-7 instead of 14-10.

I always hate assessing blame for losses because there are always so many ifs in games. I know some throw the Green Bay loss at Murray's feet, some throw it at the defense, some at the officials/Blandino (although I think they'd like to throw something more solid at him), but you look back at things like the end of the half and how that might alone have changed the outcome of the game and it's really hard to point to just one thing and say that's it, that's why we lost.

 

[AdamJT13]

- 21 points doesnt win playoff games - provided only 16 of 99 games since 2006

That's not what you showed. Those numbers are for scoring "under 23 points" (or 21, if that was a typo -- I'm not sure, because nothing matches your numbers). But we didn't score 20 points or less -- we scored 21. Teams that have scored exactly 21 points in a playoff game are 8-5 since 2006 (including our loss).

Besides that, most games don't have each team with only eight possessions on which it tried to score, so you'd have to compare it only to similar games. Or look at how often 2.625 points per possession was good enough to win. In both cases, I'm guessing the points we scored should have been enough to win the majority of the time.

 

[TheDude]

I always hate assessing blame for losses because there are always so many ifs in games. I know some throw the Green Bay loss at Murray's feet, some throw it at the defense, some at Blandino (although I think they'd like to throw something more solid at him), but you look back at things like the end of the half and how that might alone have changed the outcome of the game and it's really hard to point to just one thing and say that's it, that's why we lost.

what he said
: