Predictions: Dolphins at Saints

[Spider-Dan]

Right, but our season includes the losses, too.  You can just take the games you play well as an indication of how good you are.  This same defense got some points laid on us a few times, too.

But that's the point: if the defense had played as well in the first half of the season as they are now, the Dolphins would only have two losses.  The offense hasn't had that same kind of wild swing.  It's not like they went from 10 points a game in the first half of the season to putting up 30 points a game now.  The offense has been mostly mediocre, but coupled with a top defense that can be enough to get the job done.

Ultimately, we're not going to see anything "of value" until a playoff win... maybe two.

Win @ TEN?  The Titans are really banged up and missing their best player.
Win vs. NE?  Rookie QB hit the wall at the end of the season (even more so if they lose to JAX).
Win against TEN in the Wild Card?  Derrick Henry is rusty.  BAL?  Decimated by injury.
etc.

Nothing MIA can do will change the narrative until/unless they beat KC, BUF, or LAC in the playoffs.

[Dolfanalyst]

But that's the point: if the defense had played as well in the first half of the season as they are now, the Dolphins would only have two losses.  The offense hasn't had that same kind of wild swing.  It's not like they went from 10 points a game in the first half of the season to putting up 30 points a game now.  The offense has been mostly mediocre, but coupled with a top defense that can be enough to get the job done.

Ultimately, we're not going to see anything "of value" until a playoff win... maybe two.

Win @ TEN?  The Titans are really banged up and missing their best player.
Win vs. NE?  Rookie QB hit the wall at the end of the season (even more so if they lose to JAX).
Win against TEN in the Wild Card?  Derrick Henry is rusty.  BAL?  Decimated by injury.
etc.

Nothing MIA can do will change the narrative until/unless they beat KC, BUF, or LAC in the playoffs.

You have to consider the pass offenses the Dolphins have faced during the win streak.  In terms of season EPA per pass dropback league ranks they are as follows:

Houston 31st
Baltimore 17th
Jets 27th
Panthers 32nd (last)
Giants 30th
Jets 27th
Saints 25th (plus Ian Book)

That makes a defense's job awfully easy.  On top of that, the defense last year feasted on poor passing teams and was exploited by good ones, and made their living largely off of takeaways, which are about 50% random and can't be counted on.

This team is in for it when it plays against a good passing team.  Unless they can get lucky with turnovers, they won't defend them well, and they won't keep pace offensively.

[Spider-Dan]

Why would you just consider the pass offense rank, instead of the total offense?  That doesn't make sense, especially at the end of the season when the weather gets worse, high-powered passing offenses traditionally fall off, and power running teams start producing more results.

[Dolfanalyst]

Why would you just consider the pass offense rank, instead of the total offense?  That doesn't make sense, especially at the end of the season when the weather gets worse, high-powered passing offenses traditionally fall off, and power running teams start producing more results.

Run defense isn't a significant predictor of season win percentage in the NFL when included in a regression model with pass offense, run offense, and pass defense.  Teams can have relatively poor run defenses and still have very good season records, or conversely they can have great run defenses and have relatively poor season records.

Pass offense and pass defense, in that order, are by far the strongest predictors of season win percentage in the NFL.  Run offense comes in a distant third of the four major facets of the game, and run defense isn't even significant, statistically speaking.

So when you're determining what meaning to make of the seven-game win streak in terms of this team's quality, look first at how it passed the ball and how it defended the pass during that period.  The Dolphins' pass defense's job has been very easy during the win streak, as noted in my post above.  Likewise its pass offense has been nothing special, at 12th in the league with 0.091 EPA per pass dropback during that period.  Green Bay by contrast has had 0.275 EPA per pass dropback during that period, a whopping difference.

The Dolphins have been doing it with smoke and mirrors against weak opponents.  They're destined to be tripped up against high-quality passing teams they can't defend adequately and with which they can't keep pace offensively.

As for bad weather games, all bets are off on those, since predictors of winning become less reliable the smaller the sample gets (i.e., single bad weather games), and weather can change how teams play.  But if you have a team whose limited salary cap has been devoted to running the ball well and stopping the run in bad weather -- at the expense of passing the ball well and defending the pass well -- you're destined to win a game here or there in the month of December and likely to do poorly the rest of the season.  It's not a way to build a team.

[Spider-Dan]

Given that you were recently arguing against an entire season as a measurement for "team quality," I'm not sure why you're now citing midseason statistical rankings as a meaningful indicator.

[Dolfanalyst]

Given that you were recently arguing against an entire season as a measurement for "team quality," I'm not sure why you're now citing midseason statistical rankings as a meaningful indicator.

They're certainly a less reliable indicator after a mere 15 games than they would be for a major league baseball team after 150, for example.

That's why the Dolphins can play like they have during the win streak with regard to the strongest predictors of winning in the game and still be 7-0.  If that pattern of performance continued over 150 games they'd likely end up somewhere around 70-80.

[Spider-Dan]

Point being: you're happy to cite passing offense rank after 15 games when it supports the conclusion you like, but when the statistics DON'T support the conclusion you like, suddenly there are significant discrepancies between actual team quality and win percentage over a mere 17 games.

If we can discount literal wins and losses when it comes to evaluating teams, then we damned sure can discount passing offense rank.

[Dolfanalyst]

Point being: you're happy to cite passing offense rank after 15 games when it supports the conclusion you like, but when the statistics DON'T support the conclusion you like, suddenly there are significant discrepancies between actual team quality and win percentage over a mere 17 games.

The fact that the Dolphins are 7-0 during the win streak while performing like a team that would typically be 3-4 (or 70-80 over 150 games) certainly supports the point that win percentage over relatively small samples of games isn't as reliable a measure of team quality as win percentage over larger samples.

[Dolfanalyst]

If we can discount literal wins and losses when it comes to evaluating teams, then we damned sure can discount passing offense rank.

The smaller the sample size, the more every measure of team quality should be discounted, including win percentage.

[pondwater]

I haven't compared schedules, but haven't we played mostly the same teams as all the other AFC East teams? But yet we are within one game of Buffalo and NE. At the end of the day the only thing that matters is wins and losses. All this speculation is just nonsense, they play the games for a reason.

[Dolfanalyst]

I haven't compared schedules, but haven't we played mostly the same teams as all the other AFC East teams? But yet we are within one game of Buffalo and NE. At the end of the day the only thing that matters is wins and losses. All this speculation is just nonsense, they play the games for a reason.

Certainly teams' records are what matters in the end because they primarily determine playoff seeds, but what we're talking about here is how to predict how the Dolphins will play in the future on the basis of how they've performed in the past.  There are measures that are better than win percentage at predicting how they will play in the future.

Here's a great article about how wins and losses -- and even point differentials -- aren't the best ways of predicting the future performance of teams:

https://www.nfeloapp.com/analysis/whats-the-best-nfl-game-grade

Think about it very simply -- if team A beats the worst team in the league 3-0 it's credited with a single win.  If team B beats the best team in the league 50-0, it's also credited with a single win.  Certainly there is more information to be gleaned about those two teams from those games than simply that they both experienced "wins."

[Tenshot13]

Point being: you're happy to cite passing offense rank after 15 games when it supports the conclusion you like, but when the statistics DON'T support the conclusion you like, suddenly there are significant discrepancies between actual team quality and win percentage over a mere 17 games.

If we can discount literal wins and losses when it comes to evaluating teams, then we damned sure can discount passing offense rank.

I believe they call that confirmation bias.

[Dolfanalyst]

I believe they call that confirmation bias.

No, just a misunderstanding of what I'm saying.  I clarified that adequately in my responses in my opinion, and if you're still incapable of understanding I can't help you.  My ability to have another person understand something is limited by what's going on within them, over which I have no control.

[Tenshot13]

Ah, smartest guy in the room argument.  

[Dolfanalyst]

Ah, smartest guy in the room argument.  

You mean that's you, or are you possibly incapable of understanding something correctly?