Have the Dolphins Really Improved this Year? | Page 7 | FinHeaven - Miami Dolphins Forums

Have the Dolphins Really Improved this Year?

shouright said:
You mean for the sample consisting of the 2013 Miami Dolphins?
Click to expand...

Yes... this is a Miami Dolphins message board and per your earlier post you were using this to discuss improvement within this year, right? Earlier you used the stat after the 3rd game as evidence that this model worked and my previous post said it (and I believe graphically shows) that the model is not good at predicting performance on either a weekly or an average basis.

I don't know if it is your intent but your way of answering with a question/implication instead of just saying "yes it is" or "no it isn't" to an opinion / assertion that this model doesn't work is really annoying. You have chided other posters for not doing analysis. I did one, questioned your methodology in what I hoped was a civil, organized manner and I really feel like you haven't heard a word of it based on your responses because you really don't address what I bring up.
 
I wish you were Teddy Ruxpin

imagesqtbnANd9GcQjNnzYw_lMH5kMQeNB9baA5t-1.jpg
 
tazthenomad said:
Yes... this is a Miami Dolphins message board and per your earlier post you were using this to discuss improvement within this year, right? Earlier you used the stat after the 3rd game as evidence that this model worked and my previous post said it (and I believe graphically shows) that the model is not good at predicting performance on either a weekly or an average basis.
Click to expand...
Not as evidence that the model "worked," but as evidence that the team wasn't playing in a way associated with high-level play based on a model derived from many years of full NFL team data, that accounts for over 80% of the variance in win percentage. In other words, the team wasn't playing at a high level based on that data.

We both know that as the sample size increases, the model is going to do a better job of predicting win percentage, and as the sample size decreases, it'll likely do a poorer job. However, the Dolphins' first three games were less consistent with what the model would predict for a highly above average team, and more consistent with what it would predict for a slightly above average team. Hell, the intercept in the model is 8.03, and the Dolphins' average number of predicted regular season wins based on its performance during its early-season 3-0 run was but a mere 9.4. They were undefeated at that point largely via "smoke and mirrors" in other words.

Of course we also know there's nearly 20% of the variance in win percentage unaccounted for by the model, and so that will always be a factor to consider in determining its accuracy as applied to any one team in any one year.

I don't know if it is your intent but your way of answering with a question/implication instead of just saying "yes it is" or "no it isn't" to an opinion / assertion that this model doesn't work is really annoying. You have chided other posters for not doing analysis. I did one, questioned your methodology in what I hoped was a civil, organized manner and I really feel like you haven't heard a word of it based on your responses because you really don't address what I bring up.
Click to expand...
My apologies then. Believe me, I'm paying close attention. :up:

Do you feel I've addressed what you've brought up in this post? If not, I'd be happy to clarify further. Just let me know.
 
The value of this thread was shouright isolating low penalty numbers and high interception numbers as key variables saving our season. That has gone largely unreported, or merely accepted as the norm for our team. For all the happy adjustments around here, none of them are downward, i.e. any type of acceptance that we can't count on penalties and interceptions to save our hide next season or beyond.

The overall level of the team needs to improve quite a bit to make sure we aren't dependent on variables like those, that tend to fluctuate wildly.

It starts with pass offense, which can overcome low grades in less vital areas. We aren't anywhere close to where we need to be. I'd also argue that the pass defense as a whole isn't great. I look at YPA minus interceptions.

I would agree with the posters who aren't comfortable using blocks of games to project win totals, or how a team's level has shifted from one point of a season to another. Sixteen isn't a magnificent sample size to begin with, let alone when it's sliced. Natural waves.

Shouright always spotlights the most relevant categories so I have no idea why so many posters are petrified of his threads.
 
FinfanInBuffalo said:
I'll answer this for Shouright:

Category Correlation
ANY/A1 0.55
Passer Rating 0.51
NY/A2 0.50

According to his criteria, YPA and wins are largely independent.
Click to expand...
Now throw in what the defense does with regard to the opposing team's YPA, while considering sacks and interceptions (other variables associated with the passing game and quarterback play), and you get this:

http://nflminds.blogspot.com/2011/06/adjusted-net-yards-per-passing-attempt_20.html

...a 0.84 correlation, along with a whole host of Super Bowl appearances and victories.

So, I'd say how Ryan Tannehill performs with regard to YPA is of some importance. He contributes to a large part of that equation. :)
 
shouright said:
Now throw in what the defense does with regard to the opposing team's YPA, while considering sacks and interceptions (other variables associated with the passing game and quarterback play), and you get this:

http://nflminds.blogspot.com/2011/06/adjusted-net-yards-per-passing-attempt_20.html

...a 0.84 correlation, along with a whole host of Super Bowl appearances and victories.

So, I'd say how Ryan Tannehill performs with regard to YPA is of some importance. :)
Click to expand...

Hey, I found a video of Shouright back peddling

[video=youtube;n_3v-_p3ESo]https://www.youtube.com/watch?v=n_3v-_p3ESo[/video]
 
Awsi Dooger said:
The value of this thread was shouright isolating low penalty numbers and high interception numbers as key variables saving our season. That has gone largely unreported, or merely accepted as the norm for our team. For all the happy adjustments around here, none of them are downward, i.e. any type of acceptance that we can't count on penalties and interceptions to save our hide next season or beyond.

The overall level of the team needs to improve quite a bit to make sure we aren't dependent on variables like those, that tend to fluctuate wildly.

It starts with pass offense, which can overcome low grades in less vital areas. We aren't anywhere close to where we need to be. I'd also argue that the pass defense as a whole isn't great. I look at YPA minus interceptions.

I would agree with the posters who aren't comfortable using blocks of games to project win totals, or how a team's level has shifted from one point of a season to another. Sixteen isn't a magnificent sample size to begin with, let alone when it's sliced. Natural waves.

Shouright always spotlights the most relevant categories so I have no idea why so many posters are petrified of his threads.
Click to expand...
And that's my take as well (the bolded part above). The Dolphins have the foundation needed for winning in today's NFL precisely backwards at this point. They're staying afloat via discipline (i.e., few penalty yards) and pass defense, while hoping Ryan Tannehill performs well and improves. That needs to be reversed. Ryan Tannehill needs to become a quarterback with sustained success that can be relied upon week in week out, while the pass defense remains strong, and then those other variables (penalties, etc.) can likely go either way. We won't be bound by them to simply be average.
 
FinfanInBuffalo said:
Doesn't explain your 4000 posts harping on YPA and wins when you believe they are largely independent.
Click to expand...
I don't believe they're largely independent. The task is having a high YPA, while playing defense such that the other team has a low YPA. In that sense, YPA and wins are largely dependent, and it's important that you have a quarterback who can contribute what's needed to that equation.
 
shouright said:
I don't believe they're largely independent. The task is having a high YPA, while playing defense such that the other team has a low YPA. In that sense, YPA and wins are largely dependent, and it's important that you have a quarterback who can contribute what's needed to that equation.
Click to expand...

So you subjectively interpret correlations to fit your expectations? You did post this didn't you?

shouright said:
Moderately strong correlation there. 0.51. The variables are largely independent.
Click to expand...

Are you "interpreting" now or were you "interpreting" then?
 
FinfanInBuffalo said:
So you subjectively interpret correlations to fit your expectations? You did post this didn't you?

Are you "interpreting" now or were you "interpreting" then?
Click to expand...
I'd interpret it the same way here, as well. Having a quarterback who can contribute what's needed in the way of YPA is only half the battle in today's NFL. You also need a pass defense that can contend with the league's best quarterbacks, and thereby limit the YPA they can accumulate.

In other words, winning at a high level is "largely independent" of what only your quarterback can do. You also need a good pass defense. Fortunately the Dolphins have that as we speak.

However, no matter how good your pass defense, you aren't getting there without a quarterback who can contribute what's needed, hence my emphasis on Tannehill this season.
 
You must log in or register to reply here.
Back
Top Bottom