NFL 2013 Season Thread

[Ben Kenobi]

What is your argument? State it out in full. Explain how a low yards per point metric singularly above all other metrics means the Seahawks suck

Over the course of an entire season, average yards per point works out to about 15, over all teams and all games. This is a relatively large sample.

The seahawks having 5 yards per point in the superbowl game means that they were extremely lucky. Not only were they lucky, but Denver was also unlucky in terms of yardage and points scored. Ergo, given neutral luck, and their individual performances in that game, Denver should have won 18-15. Which is about where the bookies (and the general public had it).

Sometimes a team has a freak day where everything goes right, every bounce goes their way. That was the superbowl for the Seahawks.

As for penalties, it's exactly equivalent to boosting Manning by the penalty yardage received. The math doesn't change the difference.

[Al B. Sure!]

Over the course of an entire season, average yards per point works out to about 15, over all teams and all games. This is a relatively large sample.

The seahawks having 5 yards per point in the superbowl game means that they were extremely lucky. Not only were they lucky, but Denver was also unlucky in terms of yardage and points scored. Ergo, given neutral luck, and their individual performances in that game, Denver should have won 18-15. Which is about where the bookies (and the general public had it).

Sometimes a team has a freak day where everything goes right, every bounce goes their way. That was the superbowl for the Seahawks.

First of all, the Seahawks had 341 net yards. You don't believe me, take it up with the NFL.

Therefore, the Seahawks actual yards per point is 341/43 = 7.93

Now my question for you: Why were they lucky? Couldn't they have just played really well?

I went and checked the entire Seahawks 2013 season and post-season minus the Super Bowl. They averaged 14.07 yards per point (one of the very best in the league) with a standard deviation of 5.32.


And I want to see which bookies and what general public had the Super Bowl as an 18-15 victory for the Broncos?

I'm also curious to see the math you used to derive 18-15 from.

[Ben Kenobi]

First of all, the Seahawks had 341 net yards. You don't believe me, take it up with the NFL.

Uh, no they didn't. Taking penalties away from their yardage, they only gained 242 yards.

Therefore, the Seahawks actual yards per point is 341/43 = 7.93

It makes more sense to have penalties included because that gives you a better picture.

Now my question for you: Why were they lucky? Couldn't they have just played really well?

Again, with neutral luck, their performance would have lost them the game. So, yes, they played well but it wasn't enough for them to win.

I went and checked the entire Seahawks 2013 season and post-season minus the Super Bowl. They averaged 14.07 yards per point (one of the very best in the league) with a standard deviation of 5.32.

So that means their performance in this game was about 2 standard deviations from their norm. Coupled with the Broncos scoring at about half, that gives you a game about 3 standard deviations from the expected outcome. This puts us in the 3 tenths of 1 percent range.

And I want to see which bookies and what general public had the Super Bowl as an 18-15 victory for the Broncos?

Broncos were 2.5 faves by Kickoff. Over under was about 48 - so the Seahawks and Broncos D actually played substantially better than expected. 25-22 would have been the expected score.

I'm also curious to see the math you used to derive 18-15 from.

Simple. 15 yards/point divided by their net yardage as I'd already calculated.

[Al B. Sure!]

Let me humor you and do your math with real numbers since who knows what numbers you will pull out of you ass.

The yards per point I calculated a post earlier includes the post-season. I'll just use the data from that sportingcharts.com website for the 2013 season.

The Broncos averaged 12.1 yards per point and allowed 14.3 yards per point against.
The Seahawks 13.0 and 19.0 against.

I'm not sure how you would balance offense and defense but let's split it down the middle. With the Seahawks offense at 13.0 and the Broncos defense allowing 14.3, the middle would be Seahawks at 13.65 yards per point.

Likewise, do the same for the Broncos and you get 15.55 yards per point.

Now take the NFL net yards totals and divide.

Seahawks: 341/13.65 = 24.98 expected points
Broncos: 306/15.55 = 19.67 expected points.

So in your neutral case and if the teams played at their average level all season, the outcome would have been 25-20 Seahawks.

[Ben Kenobi]

teams played at their average level all season

You want the league average, and not the team averages. If we were running bookie numbers, yes, we would want the team numbers and run the ratio of the two.

We already have the individual team performances based on their yardage minus penalties. What we want to get from these is, "how many points could each team expect to score given their on field performance?"

This way we avoid systematic bias in the results. You're double counting, which is why your figures won't work.

[-Jrabbit]

Wait, so now you two are arguing over how to predict the outcome of a game that has already been played?

Good lord...

[Al B. Sure!]

You want the league average, and not the team averages.

What? Why would you want the league average? You think the Broncos and Seahawks are league average teams?

You think their wins were lucky flukes?

We already have the individual team performances based on their yardage minus penalties. What we want to get from these is, "how many points could each team expect to score given their on field performance?"

This way we avoid systematic bias in the results. You're double counting, which is why your figures won't work.

How am I 'double counting'?

[Al B. Sure!]

Wait, so now you two are arguing over how to predict the outcome of a game that has already been played?

Good lord...

I'm being Socratic.

He made an assertion based on a specific metric. I'm using that same metric (and his logic) to disprove him.

[Ben Kenobi]

Wait, so now you two are arguing over how to predict the outcome of a game that has already been played?

Given that this game by back of the envelope calculations was about 3 standard deviations from the expected result? Absolutely.

[Al B. Sure!]

Given that this game by back of the envelope calculations was about 3 standard deviations from the expected result? Absolutely.

You're using league averages instead of the Seahawks and Broncos averages for god knows what reason
You're ignoring the points against figure since the Seahawks had a much better relative defense than the Broncos'.
Yards per point is one metric of many. There's no reason to believe it is a better metric than any other. If there is a reason, you haven't provided any evidence for it.
The ****ING NFL describes net yards WITHOUT penalties.

[Ben Kenobi]

What? Why would you want the league average?

The ratio of the relative in-game team abilities, ideally, should be the ratio of their respective in-game yardages.

To get the expected point total, to run the over/under - doing it your way would probably get us closer to the actual total, but not to the margin between the two teams. Based on the assumption that you would expect the team with greater yardage, on average, to do better than teams with less yardage.

You think the Broncos and Seahawks are league average teams?

I think yardage minus penalties gives you their relative performance against each other.

You think their wins were lucky flukes?

I think this one was, yes.

How am I 'double counting'?

You ran into this when you had to ask yourself, "how do I balance these numbers against each other." You're counting the seahawks in game performance + sea hawks performance over time. I'm only counting the Seahawks performance once, in game.

[Ben Kenobi]

You're using league averages instead of the Seahawks and Broncos averages for god knows what reason

Even if we use your numbers, they don't spit out a 42-8 victory for the Seahawks. So yes, 3 standard deviations. Broncos' performance well below their norm (about 1 standard deviation), and Hawks performance about 2 standard deviations above their season numbers.

You're ignoring the points against figure

Again, when trying to calculate their expected points generated from yardage, this number is not helpful.

Yards per point is one metric of many.

Well, then. I think it's the least luck influenced stat involved. So when trying to find luck neutral results, it's the natural place to start.

The ****ING NFL describes net yards WITHOUT penalties.

Yes, and? Penalties count. When the NFL stops counting penalties then they won't matter to overall yardages when comparing team to team performance.

[Al B. Sure!]

The ratio of the relative in-game team abilities, ideally, should be the ratio of their respective in-game yardages.

No. It shouldn't be. Do I really need to explain how turnovers and special teams demolish your reasoning?

Based on the assumption that you would expect the team with greater yardage, on average, to do better than teams with less yardage.

Invalid assumption. Like I said. Turnovers and special teams. When the defense and special teams score touchdowns, you really throw a wrench in the idea that the team with the most yards will score the most points. Don't dare say that an INT or a kick return is somehow fluky luck. Some players and teams consistently outperform others in these areas.

You ran into this when you had to ask yourself, "how do I balance these numbers against each other." You're counting the seahawks in game performance + sea hawks performance over time. I'm only counting the Seahawks performance once, in game.

What? I'm using the Seahawks and Broncos average performance. If they played at their average level all season, both on offense and defense, by a single mere metric of hundreds that you have for some unknown reason elevated to singularly sacrosanct, the Seahawks would have been expected to win, given their yardage output in the Super Bowl.

[Al B. Sure!]

Again, when trying to calculate their expected points generated from yardage, this number is not helpful.

Wut? If one team has an elite defense and correspondingly opposing offenses need to rack up a ton of yards to score, how is that not helpful to consider?

If we want to calculate expected points, what sense does it make to discard the quality of the defensive team they're playing against?!

[Ben Kenobi]

No. It shouldn't be. Do I really need to explain how turnovers and special teams demolish your reasoning?

Turnovers are basically luck. Small sample size drastically changes the outcome.

Based on the assumption that you would expect the team with greater yardage, on average, to do better than teams with less yardage.

All else being equal, are you telling me that you would expect the team that puts up fewer yards to win more often than the team that puts up more?

Invalid assumption. Like I said. Turnovers and special teams.

All else being equal assumes neutral turnovers and neutral special teams.

When the defense and special teams score touchdowns, you really throw a wrench in the idea that the team with the most yards will score the most points. Don't dare say that an INT or a kick return is somehow fluky luck. Some players and teams consistently outperform others in these areas.

Yeah, a kick return for a TD and a pick-six are flukey luck based results.

What? I'm using the Seahawks and Broncos average performance.

Yes, but you're balancing the Seahawks performance against teams not the Broncos, and the Bronco's performance on teams not the Seahawks. This introduces systematic bias. I'm using just their in-game performance against each other.

If they played at their average level all season, both on offense and defense, by a single mere metric of hundreds that you have for some unknown reason elevated to singularly sacrosanct, the Seahawks would have been expected to win, given their yardage output in the Super Bowl.

Well, that's what we are assessing here - the expected result of their Super Bowl performance, taking luck out of the equation. Not their performance in week 1.