Big Plays as Black (or Gray) Swans

[PatsFanInVa]

It's not just that - it's his extreme posture that no financial modelling at all can work (because of the fat tails) that drives "professionals" crazy. It's an extreme "I'm right and all the rest of the world is completely wrong" belief system.

More “You’re wrong and don’t know it, I’m wrong but I know it.”

I get that he’d piss off any professional, moreso the closer they are to the practice of probability. Hey, he managed to piss me off on about 3 different counts, and the most sophisticated investing I do is occasionally reallocating my defined contribution plan.

But credit where it’s due; it’s more of a “You’re all charlatans” system.

 

[Actual Pats Fan]

It's an extreme "I'm right and all the rest of the world is completely wrong" belief system.

...aaaaand we have the mantra of Tom Brokaw's "Greatest Generation"

No disrespect to WWII vets - but please give me a break

 

[PatsFanInVa]

...aaaaand we have the mantra of Tom Brokaw's "Greatest Generation"

No disrespect to WWII vets - but please give me a break

::cough cough:: or logo wars ::cough cough::

Couldn't resist

 

[PatsFanSince74]

Consistent with what I've thought and posted for a while.
All sports, especially at a high level, are an interaction between the forces of skill and randomness.
Skill takes many of the forms mentioned in this thread.
Randomness enters in the form of officials' calls, bounces of the ball, a gust of wind during an FGA, etc., etc.
That's what Belichick means, I think, when he says that "we control what we can control." Hightower used skill to strip sack Ryan, but the ball still had to bounce where Branch could recover it, etc.
That's why what the Patriots have accomplished is so remarkable.

 

[Joey007]

What do gay swans have to do with football?

 

[SlowGettingUp]

All sports, especially at a high level, are an interaction between the forces of skill and randomness.

This is actually quantifiable to some degree. 538 assigns "Elo ratings" to each NFL team. These originated in chess, where an adult beginner might have a rating of around 1000 and the highest rated players are around 2800, with an "expert" at 2000. Each hundred points reflects a 65% chance of winning a game against a player ranked 100 points lower.

The highest rating ever awarded was 1849 to the Pats 2007 just before they lost the super bowl that year:

The Complete History Of The NFL

The current Pats team has a 1751 rating (highest in the league), 60 points higher than the Eagles, making them a 58-42 favorite:

2017 NFL Predictions

So the interesting concept here is that the highest NFL Elo ever was 1849 - compare that to the 2800 in human chess (and perhaps 3100 for computer chess). The higher maximum represents less randomness.

 

[neali]

Randomness enters in the form of officials' calls, bounces of the ball, a gust of wind during an FGA, etc., etc.

Sometimes, you have to intentionally introduce your own random behavior.

 

[PatsFanSince74]

This is actually quantifiable to some degree. 538 assigns "Elo ratings" to each NFL team. These originated in chess, where an adult beginner might have a rating of around 1000 and the highest rated players are around 2800, with an "expert" at 2000. Each hundred points reflects a 65% chance of winning a game against a player ranked 100 points lower.

The highest rating ever awarded was 1849 to the Pats 2007 just before they lost the super bowl that year:

The Complete History Of The NFL

The current Pats team has a 1751 rating (highest in the league), 60 points higher than the Eagles, making them a 58-42 favorite:

2017 NFL Predictions

So the interesting concept here is that the highest NFL Elo ever was 1849 - compare that to the 2800 in human chess (and perhaps 3100 for computer chess). The higher maximum represents less randomness.

That's very interesting. I'd heard of Elo ratings, but, to be honest, had never focused on them. Thank you. I found the Wiki article very helpful in understanding Professor Elo's model (Elo rating system - Wikipedia).

I just eyeballed the chart in the NFL History link and it looks like the Pats were 100 points higher than the Giants in 2007, which suggests 65-35 odds of winning a game. Each game played between two teams is, of course, a separate event, independent of any other games played between the teams, but the Pats did win one of the two games played.

It's all built around "expected outcomes" (in Zero Sum games), so that means there's a distribution around that 58--42. So, as was demonstrated in 2008, there is a risk that relying on the "58" this year or the "65" in 2008 puts one at risk of falling prey to the "Flaw of Averages."

But, I guess that's the point you are making. The Elo model only suggests how much randomness the "skill" level of any team can take out of the equation. In the end ref's calls, bouncing balls and helmet catches can still determine any particular game's outcome.

Thanks again.

 

[PatsFanSince74]

So I'm reading Taleb's The Black Swan and just finished Fooled by Randomness.

Everybody can weigh in, but I'm especially interested in those who've read his thinking.

It's striking to me that the Pats continually win very big games by very small margins. Our SB wins (and the two losses) are great examples.

To the extent that BB talks about the Pats' "philosophy" he has emphasized the role of the big play: Defending against it, usually -- and more left unspoken, seeking to optimize our opportunities to make them.

If you think about it, they're as important on offense as defense, especially when you're doing one of those 90 yard marches in 45 seconds on the clock.

Now, there are a lot of little things we do from the point of view of moving normal bell-curve events toward our side of the distribution: Minimize penalties, emphasize execution, continued/continuing focus on situational football so that every player is alert and aware to both responsibilities (do your job) and opportunities (the big play - see M. Butler.)

But I'd posit that the bend-don't-break style (lots of zone, boo hoo), is like an application of Black Swan theory to football. To wit: You maintain an "anti-fragile" position in relation to exposure to the "black swan" (really a gray swan, because they happen every game, just with a lot of unpredictability); you put yourself in position to benefit from the "gray swan."

In between, you absorb the small losses (e.g., your defense can yield yards not points.)

These books are old enough that their risk management principles would be known by BB and company by now - and of course, the most likely scenario is that these principles were developed independently, since "Fooled by Randomness" was being written while we were starting our first run of SB dominance.

But regardless of whether there was cross-pollination... anybody else see the similarities?

(Clearly, I need this SB to get here already...)

Interesting.
Pls see my post immediately above in response to the post of @SlowGettingUp

 

[PatsFanSince74]

Sometimes, you have to intentionally introduce your own random behavior.

Even though anything done "intentionally" can not, by definition, be "random," your statement could be "True" (assuming you're not just joking, in which case a or a would have helped).

But, if you're serious, it's possible, though increasingly unlikely in the age of "All 22" tapes shared by teams going back several years and, more importantly, of the ability of Artificial Intelligence (or, the next best thing to AI, Bill Belichick working with Ernie Adams ) to analyze the tapes, for a team to introduce something that would be "random" to the opponent.

The purest example that comes to mind is the first time a "Fake Spike" was used in a game, by Dan Marino in 1994. Since that day, any team that isn't watching for a Fake Spike should be looking for a new Coach.

A close second would be the formation the Pats used when trailing the Ravens by 14 points in the Division game. But, even in that case, with good coaching on the Ravens' part, it would not have been "random" since, as Tom Brady said in a Presser after the game, "Maybe they should read the rule book." In other words, the Ravens should have been prepared for it.

Belichick has also used an Intentional Safety when he thought it was to his advantage and, at a meaningless moment in a meaningless game, had Doug Flutie attempt (and execute) a Drop Kick. But, nothing really comes close to the first use of Fake Spike on the "Intentionally Random" scale.

Onside Kicks outside of desperation time used to be unexpected enough that an opponent might (might!) have regarded them as Random, but now every team is prepared for an onside kick from the opening kickoff of the game or, as in a recent SB, the second half.

So, yes, it's possible but increasingly rare for a team to introduce the random on purpose.

 

[PatsFanInVa]

Interesting.
Pls see my post immediately above in response to the post of @SlowGettingUp

I think we're tracking pretty well, although I'm not sure how to assign these ratings, even in the regular season, never mind in the completely different structure of the post-season (cumulative wins vs. sudden death). Add to that the small sample size, even for the regular season, and I could imagine a world in which the whole past 15 years is unremarkable, and (as competing fans said, but not persistently enough), it's all a fluke. But putting on my common-sense hat, bullsh1t.

That said, there's no equivalent to unlucky bounces in chess, unless you count bad hair days, unfortuitous biorhythms, and the like. There's also no equivalent to the salary cap/free agency. Nobody takes this year's crop of grandmasters and glues electrodes on them to inhibit their executive cognitive functions, so that the JETE of the chess world - ranging from potzers to the developmentally disabled - can catch up a little.

And Cletus the Football Robot is still a fictional character, while Deep Blue beat Kasparov a generation ago now, and AlphaGo Zero just beat a go grandmaster, in part by making moves that go experts thought were huge blunders at the time. Sorry, this line was just fun, but only tangentially related.

Back to my first paragraph:

In 2008 it was trendy to say if they played SB 42 10 times, the Pats win 9. Well, no, they don't. Not that you can prove, not that you can measure. Maybe it's 5, maybe it's 1, maybe it's 0. Maybe the Giants were always going to win that game. We just don't know.

The only thing I can offer to back up this somewhat militant pose of uncertainty is the phrase so often spoken by BB, "It is what it is." Well, it was what it was. The Giants, both times, came from unimpressive regular seasons to win out in the "tournament" those 2 years.

I don't say we should go on a retrospective search for "why" as if it was a determined outcome (although of course, the coaching staff quite likely did). Performance at that level gets evaluated in New England, because on average, we go to every other super bowl in the present era. But we're somewhat blind to the reasons for each personnel decision, and I don't envy BB and company those decisions sometimes.

All that said, it certainly did prove to be the case that we weren't as good as the Giants in those two Super Bowls the only time that each was played/will ever be played. The goal is to score higher, and we did not.

What were the probabilities? I don't know, and neither does Professor Elo.

I would like to add the following:

 

[Clonamery]

What were the probabilities? I don't know, and neither does Professor Elo.

I would like to add the following:

One of these days you're gonna get it right
(Don't bring me down, no no no no no no no)

 

[Actual Pats Fan]

All that said, it certainly did prove to be the case that we weren't as good as the Giants in those two Super Bowls the only time that each was played/will ever be played. The goal is to score higher, and we did not.

What were the probabilities? I don't know, and neither does Professor Elo.

I would like to add the following:

When we went for it in 42 on 4-and-13 in the third quarter instead of a 49-yard FG attempt, I turned to stone.

("...Is Gostkowski ill or something?!?!?!...")("The game is being played INDOORS. ON TURF...")

When Wes didn't make the catch in 46, I

Oh you get the point. Let's please shut up about this until after Sunday, thanks.