Wednesday, June 10, 2009

Rainfall

You live in an area with a moderate climate. A lot of times the sun is out and the weather is warm, but there are some rainy days. There's not really much to complain about, however, as the precipitation is about average. You don't consider it to be the best place in the world to live, sometimes you watch the rain come down and wish you lived in Los Angeles, but you figure you could certainly do worse as well.

For 15 years, the rainfall has been pretty much as expected. There were a few years that it rained less than usual and you had a lot of sunny weather, 1998 for example, but otherwise it's been pretty typical, and some years were wetter than normal.

Last fall, there was a lot more rain than usual. This mostly continued throughout the winter. You've been a bit concerned about the extra precipitation, hoping that that doesn't mean it will continue throughout the spring and the summer. You have some big parties and events that you are planning to host, and you hope that the weather won't put a damper on their sucess.

Suddenly in April and May, there has been some great weather. There have been twice as many sunny days as usual, and you have lots of opportunities to get out and enjoy the sunshine. Everything you planned went off without a hitch.

You are surprised by this, and wonder if it means anything. You remember that rainy days aren't always evenly spread out - sometimes they come in groups and sometimes they don't appear for quite some time. What is more significant, the extended track record or the last couple of months? Is there any reason to expect things will change in the future?

What do you conclude?

A) This is a temporary weather variance, and the weather will likely return to normal.

B) The weather is really clutch in the playoffs

16 comments:

Navin Vaswani (@eyebleaf) said...

B.

Justin said...

B is hilarious, but A, A, A, a million times A.

The weather in Washington this April and May was especially fine. So fine, in fact, that I said that it couldn't possibly stay that way because the weather was actually too nice to hold up...and then the monsoon came.

The weather in Detroit's pretty good right now too, certainly better than it's been. There's a bit of a difference there, though, in that it's just nice, not impossibly nice. Folks up there should enjoy it for what it is, though.

Down Goes Brown said...

I wouldn't know. I'm a Leaf fan, I've lived in a rain forest since 2004.

Vic Ferrari said...

CG:

Have you read this article on hitting streakiness in baseball?

http://www.stat.berkeley.edu/~aldous/157/Papers/albert_streaky.pdf

I think it's up your street. Tyler at MC79hockey did something remarkably similar to the BlackSTAT for measuring streakiness with even strength goalscoring for a few Oiler players.

I haven't seen anything done for all regular players for a season though. Nor for regular goaltenders with EVsave%. I'm of the same mind as you, I suspect that NHL goalies are just about exactly as streaky as random chance expects them to be, but I don't know.

Anonymous said...

Are you implying Ozzy has absolutely no control over his own performance?

Berger said...

How about if you live somewhere and notice that for some reason every summer it rains for an hour every afternoon like clockwork. How large a sample do you need before you claim that there is a monsoon season in the summer where it rains every afternoon?

Osgood has had a quite a career, and quite a career built on solid post-season goaltending. Your analogy is silly because obviously there is no connection between the weather and the stanley cup finals. However, there clearly is a connection between the stanley cup finals and performance of a hockey player. Namely if one performs well then he gets to go to them.

As the guy above me said "Are you implying Ozzy has absolutely no control over his own performance?"

JLikens said...

Stephen:

The problem is that, aside from the last two postseasons, Chris Osgood has not been significantly better in the playoffs than in the regular season.

So, if I've understood your point correctly, to say that "every summer it rains for an hour every afternoon like clockwork," and analogize that to Osgood's career, is just not accurate.

FatMan said...

A


Christ Osgood is t3h CLUTCH!


And Marc-Andre Fleury just wins. Except he ain't as CLUTCHHHH as Christ Osgood.

The Contrarian Goaltender said...

Again, I'm fascinated by this perception that Osgood has consistently raised his game in the playoffs. His playoff save percentage has been below league average more often than not.

Osgood, regular season:

Save percentage: .906
League average: .905

Osgood, playoffs:

Save percentage: .916
League average: .914

That looks to me like the same goalie. If he has "control over his own performance", he doesn't appear to be controlling it any differently in the playoffs than in the regular season.

This year he has obviously been much better in the postseason, but again, sometimes you get sunshine after the rain...

Anonymous said...

"Again, I'm fascinated by this perception that Osgood has consistently raised his game in the playoffs. His playoff save percentage has been below league average more often than not.

Osgood, regular season:

Save percentage: .906
League average: .905

Osgood, playoffs:

Save percentage: .916
League average: .914

That looks to me like the same goalie. If he has "control over his own performance", he doesn't appear to be controlling it any differently in the playoffs than in the regular season."

Another erroneous and misleading way of presenting things.

First the sample size of the goalie pool in the regular season is at least twice the size of the sample size used for the playoffs, likely bigger than that though depending on the cutoffs.

Second, the sample size of the goalies in the playoffs is presented completely wrong by CG. It is intended to be misleading, as he tries to prove "his argument" rather than find any sort of truth.

The "league average save percentage" in the regular season will no doubtedly be lower in the regular season. It factors in everybody. Where Osgood ranks compared to the rest of the league is irrelevant if your thesis has to do with him stepping it up in the playoffs because it is comparing different things.

The really misleading part about all of this is that the save percentage in the playoffs for the average goalie is almost .010 better! That indicates 2 things.
1) That the quality of competition is stronger. (No shit)
2) So is the "average" goaltender. (Check which teams had success in the playoffs over the past 15 years)

So again, considering that the competition in the playoffs is better than it is in the regular season, it would be fair to conclude that Osgood would have to step his game up just to replicate the same save percentage he put up in the regular season against weaker competition. The fact that he has not only done that, but improved it by over .010 is indisputable proof of this. Which leads to the next step.

Assuming that the top goalies in the league generally make the playoffs (which is usually true) then Osgood simply being average in comparison to that sample size would give some support for him even being a top 10 goalie in the league.

But of course when you have an agenda you can make the numbers say anything you want. This blog is a perfect example of that.

Ryan said...

Full disclosure: I don't believe in clutchness, at least not in the way it's usually framed.

The right answer is clearly A, but that's simply a consequence of the fact that you picked weather to be the analogy. What if, instead of talking about weather and climate, you were talking about an annoying roommate who was loud sometimes and quiet others. But this summer, he was unusually quiet. Was that "temporary roommate variance", or was there something else going on? Without more information, there's no way to say.

Your post makes the point that weather is random. That's great, but not very useful. You should instead be making the point that goaltending is like the weather. (And to your credit, that's the case you make with most of your other posts. I'm just picking on this one specifically.)

Jonathan said...

I think the point he is driving at is that a lot of times we look at outlying one-time performances, and use that as evidence that the player is hot/cold/clutch/choke artist. Not just for goalies, we think that about all athletes. I think in Jersey. Anytime a star player doesn't get a "the big hit in the big spot," the talk shows nail him on the radio for "not coming through in the clutch," just because that one at bat happened to fall in that 70+% of all at bats that ends with an out.

NY is HORRIBLE for overreacting to what just happened. Talk about the "error of recentness."

In reality, most of all "streaks/clutch performances" don't have any real cause. It's just typical variance.

FatMan said...

In reality, most of all "streaks/clutch performances" don't have any real cause. It's just typical variance.

And that is all that needs to be said. The key word here is "most", as sometimes players are actually clutch or do choke (though I'd wager the latter is several magnitudes of order higher than the former). Generally though, we see a player as having been "clutch" for a couple of years of their playoff careers, and who "choke" in other years, which may help us understand that this is actually just variability and randomness. Most of the time, a player who's clutch is simply a player who is great that gets enough opportunities and converts them at a solid clip.

For a recent non-hockey example, see "The Closer" Kobe Bryant's performance in Game 3 of this years playoffs. For those who didn't see it, the "clutchiest" guy in the league completely threw away a winnable game turning the ball over and missing crucial freethrows. To the layman, this seems like an impossibility, whereas to most of the viewers in this blog, that is just what is expected, that everyone fails sometimes, and that success in sports is measured by how often you succeed vs. fail, and not by lazy sportwriter cliches.

Berger said...

In terms of whether or not "clutchiness" exists or whatever, why does it have to be one way or another. Can't it both exist and sometimes be explained merely by stat variance in a small sample size?

Here's a specific example: A lot of times you can see young players dominate in the regular season but seem overwhelmed in the playoffs. Is it really that outrageous to claim that they don't have the experience or mental toughness yet to deal with the increased competition and are therefore not as "clutch?"

If that story is feasible, then it certainly seems feasible to me to claim that certain players are psychologically more able to deal with the pressure of important circumstances than others. Just because you cannot confidently prove it by examining statistics due to small sample sizes doesn't mean it isn't true.

Justin said...

That's a decent idea, Stephen, but the problem is that clutchiness (yes, that's a word as far as I'm concerned!) doesn't seem to be a repeatable skill from one season to the next. Someone who is really good in clutch situations one year can just as easily break down in such situations the next (and often will).

A good example in recent hockey history would be Marian Hossa. He got a reputation for sucking in the playoffs because of some crap games in Ottawa, looked fantastic for Pittsburgh last year, and then disappeared in the SCF this time around. It just doesn't seem to follow people from one year to the next, and that works against it as an actual skill.

Anonymous said...

"NY is HORRIBLE for overreacting to what just happened. Talk about the "error of recentness."

Yep, NYers, both fans and media are absolute morons as a whole. Its not surprising that outside of the Yankees the majority of their sports teams have historically sucked. Its also no wonder they do their best to turn average players into legends.