Stopping the New Romo Myth

[ufcrules1]

Always a bunch of shoulda coulda's in there.

And you're right too.. always some should woulda coulda's in there. That is why stats alone can't pain the whole picture even if we badly want them to.

 

[percyhoward]

Does that mean it was due entirely because our defense sucked bad or maybe could it be that Rodgers turned it up against us because it was a playoff game at home and he is Aaron Rodgers after all?

When a light is turned on and you can see better, is that due entirely to the fact that there's more light, or maybe could it be that there's less darkness?

 

[ufcrules1]

When a light is turned on and you can see better, is that due entirely to the fact that there's more light, or maybe could it be that there's less darkness?

When a team has a below average defensive game, is that due 100% to them sucking or does it have anything to do with the opposing QB being elite and playing at home with all the cards on the table? There is no way to quantify this, you have an opinion and I have an opinion. At the end of the day, our team only put up 21 points which is rarely good enough to win a playoff game. You can look at time of possession, where our possessions started, and any other tidbit statistics you want but 21 points just isn't going to get it. They outscored us and we lost as a team, not simply because our defense sucked.

 

[percyhoward]

I don't see the relevance of # of possessions.

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. Otherwise, you're expecting them to score at higher than their normal rate in this game. There's no reason to believe that two teams that scored 47 points on 16 possessions would have been shut out if given 5 more possessions.

An offense can't score without the ball, and you can't judge a defense's performance based on possessions that it did not have to face.

 

[percyhoward]

There is no way to quantify this, you have an opinion and I have an opinion.

No, I just gave you a fact. You attached a couple of meanings to it, and deemed one of them to be my opinion.

 

[ufcrules1]

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

No, not necessarily. Just because the average is 21 possessions and they only had 16 doesn't mean more points would be scored if there were 5 more total possessions. We can look at past stats and say theoretically based on the stats alone, they should have scored more but it is far from an absolute. It's true that you can't just look at past performance to figure out what the future holds. Especially in football.[/quote]

 

[percyhoward]

No, not necessarily. Just because the average is 21 possessions and they only had 16 doesn't mean more points would be scored if there were 5 more total possessions.

For the number of possessions not to matter, it would have to be certain that no more points would be scored -- not just "possible."

 

[ufcrules1]

For the number of possessions not to matter, it would have to be certain that no more points would be scored -- not just "possible."

I didn't say they didn't matter, I just made the point that there is no guarantee whatsoever that if each team had 2.5 more possessions more points would be scored. So it's possible more points could have been scored and also possible no points could have been scored.

 

[jobberone]

And you're right too.. always some should woulda coulda's in there. That is why stats alone can't pain the whole picture even if we badly want them to.

Don't get me wrong although you're right there is more to the game than stats. And you're correct that some stats are not predictive.

However, many stats have a very high rate of correlation and are thus highly predictive. I won't bore with the different tests but anyone can do a thought experiment.

If you jump from a plane and have a chute failure you are likely falling to your death. Not only is this intuitive but you can test this by examining all the incidences of chute failures ending in death. Without looking it approaches 1 which is highly correlative and in this instance has a positive correlation.

TO ratio is highly predictive. PPG is highly predictive. When you say stats aren't predictive then that has a very high negative correlation.

 

[ufcrules1]

Don't get me wrong although you're right there is more to the game than stats. And you're correct that some stats are not predictive.

You know the popular saying about stats. "They are like Bikinis, what they reveal is most certainly interesting but what they hide is essential" I would agree with that saying. I think they are very interesting but I see plenty of flaws when people bring them up here. I still always enjoy when posters post them because I can appreciate the time it takes to look up a gazillion random stats. I'm glad someone is willing to do it.

 

[CowboyRoy]

Our defense was 19th in weighted DVOA and 16th in points allowed per possession. (We were also No. 1 in turnovers created per possession.)

Timely turnovers was the one thing we did great.

 

[CowboyRoy]

The 19th ranking was not a mirage. I'm actually a bit shocked how much fans are bagging our our defense. They played their hearts out last year and were not all that bad. Sure, we certainly needed more of a pass rush but I thought overall our defense was average. I have no faith in Randle from what I have seen of him and don't see him as an every down back at all. I guess we will just have to wait and see how the season goes. I genuinely hope I'm wrong about Randle and he tears it up. The team doesn't seem like they are very comfortable at RB right now IMO. They appear to be a little uncomfortable and I don't blame them. DMF has only had one full season since 2008 so you sure can't count on him. Williams has played 5 games since 2011, so you can't count on him, and Randle is a second year player that has had some off the field issues and has never been given the task of carrying the ball very much. I think this year we will get the answer of if it was our line, DM, or both that led to our success last year.

What do you mean by what you have seen of Randle? His college career bares him out as a carry the load RB that didnt have durability issues. He is a 3 down back that can block and is a great pass catcher. He can do it all. That is exactly why they drafted him. He was insurance in case Murray went down, as he always did. They wanted another Murray. And every time he has gotten the ball, he has showed up and done well in the little they have given him.

So other then being an idiot and stealing underwear and saying some dumb stuff, what is it that has given you no faith from what you have seen?

 

[CowboyRoy]

Our defense was much improved last year, and deserves some of the credit for helping us get that far. But that's irrelevant to my point that you're supposedly responding to -- that even an average defensive performance could have made a difference. If we hold Rodgers to his average passer rating we probably win the game, so I'd call that a significant difference.

Also, when you talk about the 26-21 score, you should account for the fact that it was a slowed-down game, with only 8 possessions for each team. Dallas averaged 10.6 possessions per game during the regular season, and the Packers averaged 10.2. Apply that to this game, and it's GB 33 Dal 28. That would put us 1 point below our per-game average, and the Packers 3 points above theirs. And that' point total for us assumes that the blocked FG would have been missed anyway, and we would have come away with zero points after a 1st-and-goal at the 1 if Dez's catch had correctly stood.

The defense was definitely improved no question. But it looked a lot better than it did because of the time of possession and always being fresh from sitting on the sidelines. To discount that would be silly. And AGAIN...........management agreed they were anything special.

 

[TheDude]

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. Otherwise, you're expecting them to score at higher than their normal rate in this game. There's no reason to believe that two teams that scored 47 points on 16 possessions would have been shut out if given 5 more possessions.

An offense can't score without the ball, and you can't judge a defense's performance based on possessions that it did not have to face.

I understand the pro rata calulation, but this is a typical misuse of stats when trying to assign a large sample (hopefully based on the large of law numbers) to an idiosyncratic event. This "scoring per drive" can measure effectiveness throughout a season, but in a single game each team should get the same possessions +/- 1 for the end of half/game. This can change if you score on onside kicks drives. Therefore, using this stat as "predictive" or "statistically significant" to a game, the team with the higher average scoring per drive will ALWAYS win. The only way for dallas to win is have a higher scoring/drive than GB in that game. The game is not obligated to follow historical patterns, each game is unique. Unique situations, plays, injuries, calls etc.

Its similar to banks holding reserves to cover loan defaults. With 1 million customers, banks can anticipate the amount of losses by customer FICO or credit score, etc. perhaps the model says historically 1% of loans will default. The bank does not include 1% losses on every customer and then make a decision on one customer.

The game was purposely slowed down by both teams. I would assume this is done more strategically than serendipitously having each team have 12 play, 10 minutes drives.

 

[CowboyRoy]

Past performance is not indicative of future results. You can say the defensive performance was 17 points below the average defensive performance against Rodgers in 2014 but does that mean it was due entirely because our defense sucked bad or maybe could it be that Rodgers turned it up against us because it was a playoff game at home and he is Aaron Rodgers after all? You have your opinion on that and I have mine. Stats don't prove either of us right or wrong.



With regards to us getting a turnover on one of Rodgers 35 attempts, I would say judging by the stats (I'm being funny here) that the odds of getting a turnover on one of his 35 pass attempts would be ridiculously low consider he only had 5 interceptions all year on 520 pass attempts.

The most important thing to remember here is that Rogers was limping around all game long. They werent even sure 100% if he was going to play in the game.

 

[jobberone]

You know the popular saying about stats. "They are like Bikinis, what they reveal is most certainly interesting but what they hide is essential" I would agree with that saying. I think they are very interesting but I see plenty of flaws when people bring them up here. I still always enjoy when posters post them because I can appreciate the time it takes to look up a gazillion random stats. I'm glad someone is willing to do it.

I can see why when one is so entrenched it is difficult to pivot even with the facts.

 

[TheDude]

You misinterpreted my post. I was saying that the defensive performance against Rodgers was 17 points below the average defensive performance against him in 2014. It was a "defense played poorly" argument -- not a "regression to average" argument, an example of which would be my trying to convince people that the defense would have played better if they faced Rodgers 16 more times.

Also, it doesn't make a whole lot of sense to insist that a turnover that occurred carries more potential weight than one that did not. So when you dismiss my points with "the lost turnover at the worst time is always the first place to look," you might also want to consider the value of the turnover that the Packers didn't lose. And that doesn't necessarily just mean Cobb's fumble on the kick return that we didn't get, but also any of Rodgers' 35 attempts in the game that could have been intercepted...but were not.

It doesn't make a whole lot of sense to run 800 hypotheticals for what the Dallas defense didnt do to lower Rogers to his average (i.e. they played poorly or whatever) when leaving Romo (rating at 143) static.

A tangible turnover can always be viewed more to impact than potentially 3-100% (1 or 35 INTS for rogers) of the passes that might have been intercepted. Turnovers are more on the tails of play outcomes (i.e. the frequencies of turnovers form total plays) and not all turnovers are equal. An example from a previous debate this week - The 2010 opener vs Washington. The Choice fumble as time expired in the half was infinitely more costly than an INT on a hail mary.

Dallas had opportunities to win this game with the defense playing exactly as it did. They could have done more, but 21 points wasnt enough

 

[TheDude]

I think the stats are very important esp things like TOs and FP. But you're right in that teams avoid big negative plays, make some positive big plays, overcome mistakes and capitalize when they can.

You're right about the end of the half. Might have made the difference in the game. Always a bunch of shoulda coulda's in there.

yeah, I get season stats as a benchmark, but single game standalone stats just dont provide context

Perfect example is the NFC Championship game last year. Wilson throws 4 picks and has a 44 Rating. They lose the turnover battle -2

They lost primarily due to GB settling for an 18 yard and 19 yard FG in the 1st qtr off turnovers - starting on Seattle's 19 and 23 yard line respectively.

Points are really all that matter.

 

[jobberone]

yeah, I get season stats as a benchmark, but single game standalone stats just dont provide context

Perfect example is the NFC Championship game last year. Wilson throws 4 picks and has a 44 Rating. They lose the turnover battle -2

They lost primarily due to GB settling for an 18 yard and 19 yard FG in the 1st qtr off turnovers - starting on Seattle's 19 and 23 yard line respectively.

Points are really all that matter.

See my follow up post above. I'm agreeing that stats aren't everything but they can be highly correlative and predictive. One must look at the data then use that powerful computer between your ears to interpret the data. But it would be a mistake to dismiss it as not being predictive over time. There are going to be outliers no matter for the most part with stats and football. Exceptions shouldn't be used to dismiss the statistics though.

 

[percyhoward]

I understand the pro rata calulation, but this is a typical misuse of stats when trying to assign a large sample (hopefully based on the large of law numbers) to an idiosyncratic event. This "scoring per drive" can measure effectiveness throughout a season, but in a single game each team should get the same possessions +/- 1 for the end of half/game. This can change if you score on onside kicks drives. Therefore, using this stat as "predictive" or "statistically significant" to a game, the team with the higher average scoring per drive will ALWAYS win. The only way for dallas to win is have a higher scoring/drive than GB in that game. The game is not obligated to follow historical patterns, each game is unique.

You can't simply assume that both teams would get shut out on 5 more possessions because "each game is unique."

In this game, the teams had fewer than normal possessions, and scored fewer points than normal. It's not really about understanding how to do a calculation, but about understanding that there may be a connection between those two events.