Spurtability= The statistic

by Batumshakalaka on Jan 30, 2012 1:25 PM PST

Hey guys, I've been taking a statistics class recently and just got into variability and standard deviation. I'm not going to get deep into it, but basically variability is how much a data set can vary. Say, a player is averaging 15 points per game. However, his scores vary wildly from well over twenty to single digits. This would show high variability, or "spurtability". A player who ranged closer from 13-17 on average, would show low variability. Are there any NBA statistics out there to show streakiness or even, egads, "Spurtability"?

Obviously this wouldn't be an end all stat like PER, but it would be a good indicator for how consistent a player is. This would probably be done best with Standard deviation, which isn't the easiest statistic but can't be much worse than PER. You might even be able to combine this with the variability of minutes so that players with inconsistent minutes won't be penalized.

What do you guys think?

Poll

Which of these do you think would be most useful?

Variability of scoring

1 votes

Variability of scoring considering minutes played

4 votes

Variability of FG%

2 votes

Variability of FG% considering minutes played

2 votes

Variability of scoring and FG% but not minutes played

0 votes

Combine all three for total "Spurtability"

12 votes

You must be high, do you even know what standard deviation is?

10 votes

I'm so confused...

6 votes

Other as indicated in comments

1 votes

38 votes | Poll has closed

Tweet Comment 26 comments | 0 recs |

Do you like this story?

Comments

Ben Does one for the site

he calls it Shake.

but essentially its the actual change of scoring outputs across time. more of a visual way to look at it tho

i figure u could calculate pretty easily by getting a fat zip file of all game logs (they are around), and then using some excel stuff.

not sure if there is one on baskebtall reference etc.

Rip City Baby...People have no idea what is coming.

Follow my twitter @PDXBlazersFTW. Lots of random Blazer Posts from links I find around the blogosphere.

by lanepete on Jan 30, 2012 1:44 PM PST reply actions

sounds a bit like Ben's "shake chart"

by poorwebguy on Jan 30, 2012 1:44 PM PST reply actions

http://www.blazersedge.com/2010/2/18/1316184/shake-charts-measuring-in

Blazersedge.com || CBS Sports NBA || Twitter

by Ben Golliver on Jan 30, 2012 1:44 PM PST reply actions

Ah of course, how did I miss that

Now I can actually understand it as well. I’m sure you’ve heard it before, but that is a pretty amazing visualization of consistency. It seems odd that this isn’t a more common measurement outside of the “eye test”. Then again, maybe some of the stat head scouts/GMs use this and try to keep the rest of the league from knowing about it.

by Batumshakalaka on Jan 30, 2012 2:02 PM PST up reply actions

this sounds like ben's shake chart he does or something.

"I told somebody to stop crying," Pendergraph said after the game. "Actually, I told them all to stop crying."

by cornplant on Jan 30, 2012 2:13 PM PST up reply actions

this sounds like Ben's shake chart

just throwing this out there for the umpteenth time, lol.

"If you can do a half-assed job of anything, you're a one-eyed man in a kingdom of the blind."

by thankyouforblaze on Jan 30, 2012 2:59 PM PST up reply actions 1 recs

Doesn't this sound quite a bit like the shake chart thing that Ben does? Everybody!

Blunts and cupcakes...Bumble Bee's secret weapon.

by Wotan on Jan 30, 2012 4:19 PM PST up reply actions

look what i started!

"I told somebody to stop crying," Pendergraph said after the game. "Actually, I told them all to stop crying."

by cornplant on Jan 31, 2012 1:16 AM PST up reply actions

eh, I think Lanepete started it with the first comment of the post

"If you can do a half-assed job of anything, you're a one-eyed man in a kingdom of the blind."

by thankyouforblaze on Jan 31, 2012 9:41 AM PST up reply actions

i know. i was being even more silly than i was before

but you don’t get it.

"I told somebody to stop crying," Pendergraph said after the game. "Actually, I told them all to stop crying."

by cornplant on Jan 31, 2012 5:45 PM PST up reply actions

What ? you mean that this is like Ben's shake chart?

…oh

Rip City Baby...People have no idea what is coming.

Follow my twitter @PDXBlazersFTW. Lots of random Blazer Posts from links I find around the blogosphere.

by lanepete on Feb 1, 2012 9:23 AM PST up reply actions

When attempting to quantify the variability of something, the best place to start would seem to be the extremely well established statistical measure known as the variance.

by jksnake99 on Jan 30, 2012 2:13 PM PST reply actions

I also recommend looking into the concepts of the standard deviation and coefficient of variation.

by jksnake99 on Jan 30, 2012 2:15 PM PST up reply actions 1 recs

A standard error would even be useful at some point when dealing with your full sample..

by your face on Jan 30, 2012 2:16 PM PST up reply actions

if you want spurt, then don't be mean.

i like it here, there, everywhere.

by Name's Ash on Jan 31, 2012 8:05 AM PST up reply actions

If this somehow ties into "shower presence" I'm gonna get really grossed out

"What began as a credible protest against bank bailouts, crony capitalism and the like has, in large measure, been hijacked by crazies and criminals,"

by 92wastheyear on Jan 30, 2012 2:14 PM PST reply actions

huh???

"If you can do a half-assed job of anything, you're a one-eyed man in a kingdom of the blind."

by thankyouforblaze on Jan 30, 2012 3:00 PM PST up reply actions

Wait, didn't YOU just tie it into "shower presence" with that comment?

Phase 1: Collect underpants
Phase 2: ???
Phase 3: Profit!

by HailOden! on Jan 31, 2012 10:14 AM PST up reply actions

Spurtability is the natural conclusion of the Shake Chart

IE. it’s end result

I'm a grown man stuck inside the mind of a 3 year old

by The Arkitect on Jan 30, 2012 2:14 PM PST reply actions

Yeah I already made that joke

"What began as a credible protest against bank bailouts, crony capitalism and the like has, in large measure, been hijacked by crazies and criminals,"

by 92wastheyear on Jan 30, 2012 2:14 PM PST up reply actions

If you have to point it out

you weren’t doing it rite. :)

Honor Alaa Abdelnaby.
First in the NBA. At least alphabetically

by OhOhOden on Jan 31, 2012 9:56 AM PST up reply actions

Ark's joke was waaaaaay better

Phase 1: Collect underpants
Phase 2: ???
Phase 3: Profit!

by HailOden! on Jan 31, 2012 10:15 AM PST up reply actions

Still isn't perfect.

May need to be a statistic viewed in 3D. That would be cool.

by FlyingOutlaw on Jan 31, 2012 12:21 AM PST reply actions

I think just variance or standard deviation of both TS% and pts/FGA would give you the idea you want

That would be really easy to do…for someone who’s good with Excel (I’m not).

Phase 1: Collect underpants
Phase 2: ???
Phase 3: Profit!

by HailOden! on Jan 31, 2012 10:18 AM PST reply actions

Play 2k11 much?

OSU '06
GForce Crash Wallace FTW!

by TyboOSU on Feb 1, 2012 3:26 AM PST via mobile reply actions

Actually, just saw this link on the sidebar:

http://www.sbnation.com/nba/2012/2/2/2766177/lebron-james-volatility-scorers-hook

However, both this and Ben’s shake are incredibly suspect from a mathematical standpoint. First, the idea of a consistent player being one who scores exactly on their average is extremely wrongheaded. This would mean that when they’ve scored more than average, they’re likely to performer poorer, and thus hit their average. A better model is to use something like a Poisson distribution. A player who’s performance followed a poisson distribution, especially on a per minute basis, should be the model of a perfectly consistent player.

However, if we take the Poisson distribution as a model, then you can’t use standard deviation/average as a metric, as that article does. The standard deviation of a Poisson distribution is the square root of lambda, so the expected average will be proportional to 1/sqrt(lambda), making better players seem more consistent. Instead, using the variance should give you a better way to compare volatility between players.

by jnewhouse on Feb 2, 2012 7:09 PM PST reply actions