"NBA Statistics Treatise"

[rlee]

http://www.basketballinsiders.com/an-nb ... -treatise/

[Voyaging]

Site is automatically playing two videos and I can't find the second one to pause it.

Really great article though. Some incredible resources I wasn't aware of. I really like Inpredictable. What do you guys think about the Win % added metric? Seems really interesting, just very limited. I kind of like the "Kitchen Sink" one.

Also, I noticed a bunch of the pages like the individual TPA pages on NBAMath don't work. Any idea what's up there?

[mtamada]

Cool compilation and reviews. But Dowsett is oddly dismissive of 82games.com; is he unaware of the history behind it?

[Crow]

We talked about that a bit on twitter. I tried to explain its value. I think he recognizes he may have been a bit quick on that take.

[Nate]

... What do you guys think about the Win % added metric? Seems really interesting, just very limited. I kind of like the "Kitchen Sink" one.
...

The part of 'analytics' that people always seem to gloss over is that it's important to know what the question is.

Mostly, I think win % added is a trivia stat. It doesn't really matter whether that 3-pointer gets hit in the first minute or the 47th.

Micheal Beuoy (the guy who does inpredictable) did do some articles for fivethirtyeight, but they were NFL related.

[Voyaging]

Cool compilation and reviews. But Dowsett is oddly dismissive of 82games.com; is he unaware of the history behind it?

What is the history behind it?

Mostly, I think win % added is a trivia stat. It doesn't really matter whether that 3-pointer gets hit in the first minute or the 47th

I've been thinking a lot about it and I'm really not sure. It's a really tricky subject, whether those high win % affecting shots really are more valuable.

[Mike G]

82games was the original public plus-minus site.
They claimed to have 'adjusted' PM, but I never saw it.

In a blowout, that last-minute shot isn't worth much. But some really believe it's worth everything in a close game.

[Nate]

...
I've been thinking a lot about it and I'm really not sure. It's a really tricky subject, whether those high win % affecting shots really are more valuable.

They are more valuable in terms of wins, but they're not very rich in information about how good the players are.

[Mike G]

Explain please how 2 or 3 points made vs missed are more/less valuable, or likely to produce a win, at any point in the game.

[Crow]

82games was the original public plus-minus site.
They claimed to have 'adjusted' PM, but I never saw it.

It was and is in the commentary section only. http://www.82games.com/articles.htm

[Nate]

Explain please how 2 or 3 points made vs missed are more/less valuable, or likely to produce a win, at any point in the game.

Suppose your team is up 20 points with 5 seconds to go. Your chance of winning isn't going to change much if you make or miss a shot. This is a situation where 2 or 3 points have very little value.

Suppose, instead, that your team is down by 1 with 5 seconds to go. Then the difference between making and missing a shot is pretty close to the difference between winning and losing.

[Mike G]

I didn't ask about drastically different game scenarios, nail biter or blowout.

In either type of game, how is a last minute bucket more or less valuable than one in the first minute?

[Nate]

I didn't ask about drastically different game scenarios, nail biter or blowout.

In either type of game, how is a last minute bucket more or less valuable than one in the first minute?

That's the whole point: First minute shots (which consistently have a small impact on expected wins) are a different game situation than last minute shots (which tend to either not matter or determine the outcome of the game.)

There's some probabilistic or statistical thinking involved here. It's perfectly sensible to say "the whole game happens, and all the points count the same." But, you can also build some kind of model and then see how the various in-game events shift the win expectation according to that model. Then you can credit marginal win percentage to box score events (and in turn to the players.)

[rlee]

A buzzer beater that puts a team ahead doesn't leave time for the opposing team to offset

[rlee]

From Todd Whitehead (@crumpledjumper):


https://twitter.com/CrumpledJumper/stat ... 1689932801