BP: More Per-Minute Analysis

One of the arguments statistical analysts have made in favor of per-minute numbers is the ability to identify players stuck on the bench as breakout candidates when they get more minutes. A quintessential example is Milwaukee Bucks guard Michael Redd, who backed up Ray Allen early in his career before becoming an All-Star the year after Allen was dealt to Seattle. Andrei Kirilenko and Zach Randolph are other examples of players touted by analysts well before they became mainstream stars.

How well does good per-minute performance portend future success? To test that, I turned to the “Diamond Rating,” invented a few years ago by a former colleague of mine, Kevin Broom of RealGM.com.

Go to the article.

Mike Kurylo

Mike Kurylo is the founder and editor of KnickerBlogger.net. His book on the 2012 Knicks, "We’ll Always Have Linsanity," is on sale now. Follow him on twitter (@KnickerBlogger).

32 thoughts to “BP: More Per-Minute Analysis”

1. Mike says:

Looks to me the results are very mixed and a person could probably do a better job just watching players actually play than guessing based soley on statistics. Thats probably why teams have scouts instead of computer geeks. :)

2. Lol, Mike, those are fighting words…. ;-)

The reason you keep stats is because you can’t watch all the games. And even if you could, you can’t really make sense of them without the numbers, especially when you throw human bias into it.

And most NBA teams use stat geeks these days…

3. caleb says:

It’s not laid out very well – maybe because I have very little feel for what a good “diamond rating” is – is 13.4 good? 6.8? 26.2?

From what I can tell, it’s a pretty good predictor of young players – but also looks like going to a lot of trouble, instead of just looking at normal per-minute stats and weeding out, say, those over age 25 and those playing less than 500 minutes. The extra math doesn’t seem to add much, except confusion.

4. Mike – what a great idea. Let’s throw out all scientific evidence. The earth now goes around the sun, radon is harmless (and in fact doesn’t exist), and we can all be millionaires playing the lotto. All because that’s what I observe. Visual evidence is much better for everything.

Once someone suggested that people be well versed in both observational and statistical evaluation. But we killed him. There’s no room for thought like that.

5. Mike says:

Owen no offense but when it comes to drafting or trading for a player I dont know a single team that trusts a computer statistician over actually seeing the guy play either in person or vis film with their own eyes.

Also These stats show how often the break out players were called rightly and wrongly but what about all the players that broke out that are not even listed?

Sorry I will trust my own eyes over a computer every day. Its stuff like this that had Hollinger saying Jackie Butler was around the 6th best center in the nba before the last season ahead of Curry and Shaq. Its stuff like this that implied Allen Houston was just a good defender as Bruce Bowen years ago.

Thats why they call this stuff FANTASY

6. Mike says:

Mike K did I say what you implied I said? No I did not so why throw out a 100% dishonest red herring?

Do you think you are a great scout based on computer stats? Well if you do then make your case to NBA teams and have them hire you to replace their head scouts.

7. “a person could probably do a better job just watching players actually play than guessing based soley on statistics.”

Probably, and if not they might give someone like Mike Olowokandi a big deal. If the people that only watched the game were so good at identifying talent, then there wouldn’t be people trying to do it better with statistics.

8. Mike says:

The people that gave the Candy man a big contract didnt look at his statistics? Wow thats shocking. How did you come by this inside information?

If you are such a great scout using just statistics once again I think a NBA team should hire you to replace their head scouts. Think of how much money they could save in travel costs and video tapes. They could let you work at home and just pay for your internet connection. :)

9. “I dont know a single team that trusts a computer statistician over actually seeing the guy play either in person or vis film with their own eyes.”

Nobody trusts only statistics versus visual, and I don’t think anyone in the statistical community believes that either. (It’s certainly not the prevailing point of view). However there are a good amount of teams that do use statistics in their evaluation. I can think of at least: Dallas, Houston, Boston, Denver, and Cleveland.

10. Mr. Black says:

Statistics can be used to prove just about anything. 90% of people surveyed agree with that.

11. Frank says:

Wow it got testy in here in a hurry. Overall, though, I’m not sure what the linked article adds to anything. All it really proves is that their statistical system is a complete crapshoot. Doesn’t mean that statistical analysis is invalid, just that THAT one is. Not to mention I think he was being kind to himself that there are 11 hits out of the 25– I only counted 5 there that are actually good players — Zach Randolph, Kirilenko, Biedrins, Jefferson, and David West. Maybe I can give him Varejao but that I think should fall into the “mixed results” category. That would give roughly 6 hits, 11 mixed, and 6 busts. Sounds like a roughly normal distribution. You could probably pick 25 NBA players at random and come out with the same success.

And Mikes, I don’t think either of you is really saying that visual evidence or stats are used in exclusion to the other, so you’re probably both saying the same thing. I think we should just find a BETTER analysis model than the one in the article, which to my mind is not a useful one at all.

12. teddd says:

Mike, anybody that uses JUST stats to make their scouting decisions is a fool. Even statheads would agree with that. But anybody that dismisses all stats as being useless is equally foolish for the reasons that people have tried…hopelessly it seems… to point out to you. (you can’t see every game so it avoids biases created by seeing that player play a good/bad game etc…).

Of course you of all people should know about preconceived biases, right…? ;)

13. “Wow it got testy in here in a hurry.”

Yeah. I had a bad day at work, so I was less diplomatic than normal. But even looking back, I really didn’t think much of the original comment, especially the “computer geek” crack.

“Not to mention I think he was being kind to himself that there are 11 hits out of the 25? I only counted 5 there that are actually good players ? Zach Randolph, Kirilenko, Biedrins, Jefferson, and David West. Maybe I can give him Varejao but that I think should fall into the ?mixed results? category. That would give roughly 6 hits, 11 mixed, and 6 busts.”

I think it’s how you define hit. If a player is getting less than 20 minutes, and is being highly productive, then going from a bench player to a starter is a major jump. Remember in the NBA there is little variation year to year, and there isn’t much turnovers in “All Stars”. It’s not like baseball where you have essentially 15 starting positions (hitters + rotation + closer) and there is a fair amount of variance (some guys will just have career years, while others regress a bit).

14. teddd, it seems to me like you’re saying the exact same thing that Mike did three posts above you, yet somehow also slamming him for that.

15. Gregson says:

How could you ever get a handle on player performance without stats? No one is suggesting we gouge out scouts eyes…

16. Z says:

“you can?t see every game so it avoids biases created by seeing that player play a good/bad game etc?”

I’ve read Mike K.’s analysises (what is the plural of analysis?) for a few months now and I am still skeptical of the usefulness of compiling such data (though I understand it better!). It seems to me that statmeisters have been able to tell that Davis Lee will be a valuable player in this league since early in his first season. Far earlier than the rest of the league (since most casual fans of the NBA have no idea who Lee is). But without stats, the people (like Mike with no K) that watch Lee play regularly have also known exactly what the stat guys have seen: that the team plays better with Lee than without Lee. Stats simply color between the lines of the conversation. They are useful for comparing and contrasting what we’ve seen, and more importantly what we haven’t seen.

This particular study predicted Randolph, JO, Mike Redd, and AK for success in the league before they were all stars. Lately there have been a good number of players selected 29th and higher that have become All Stars in the NBA: Arenas, Manu, Parker, Boozer, Okur, Josh Howard, Rashard Lewis, etc…

Is there statistical evidence that any would become All Stars after their rookie years? Is there similar statistical evidence that the likes of Kwame Brown, Joe Smith, Andrew Bogut, and other high draft picks that haven’t become All Stars would (or still will?)?

Was it ever apparent that Redd would be anything more than a great shooter. Has he in fact let down some prognosticators with his one dimensional game?

I’m sure the answers to these questions are in the archives somewhere, but since the conversation has gone this way, I thought I’d think out loud…

17. Z says:

“there isn?t much turnovers in ?All Stars?. It?s not like baseball where you have essentially 15 starting positions (hitters + rotation + closer) and there is a fair amount of variance (some guys will just have career years, while others regress a bit).”

It’s true– the All Star team in the NBA is kind of like a fraternity. Once you’re in, you’re in for life. In baseball there are lots of one time All Stars not only because people have career years (or actually career first-half-of-a-year) but because every team is required to be represented in the All Star game.

18. Peter says:

The fact that current ‘experts’ do things in a certain way (e.g. trusting scouts more than statistics) does not mean it is right. Experts are frequently wrong. When a mathematics professor finally analyzed the game of blackjack and figured out the winning strategy, it turned out that none of the ‘experts’ had it even remotely right (for example, it is frequently correct to stand pat with a 12, something that no expert at the time ever did).

19. Jesse says:

In response to the RealGM article cited by bucketz —

First of all, it’s great that the Knicks are 2-1 but let’s wait a bit before declaring them a playoff team. The Denver team they barely beat lost by about 40 to the Celtics the next night. They still can’t guard a 3-point shooter to save their lives.

Second, the conclusion of the article — that “rebuilding is a four-letter word in New York and with Cablevision resources on hand, it is difficult to imagine how Isiah could have done much better” — seems ridiculous to me. What about, say, by letting expiring contracts actually expire in order to free up cap room to acquire really good players who don’t have major question makrs? The quote seems to imply that Dolan put a gun to Isiah’s head and forced him to spend money unwisely. Dolan basically gave Isiah carte blanche — and the Knicks have generally royally sucked under his tenure. Are we to believe that had Isiah elected to do a true “rebuilding” process rather than by making expensive trade after expensive trade, Dolan wouldn’t have let him?

20. “The quote seems to imply that Dolan put a gun to Isiah?s head and forced him to spend money unwisely.”

If you look at the way that Cablevision has operated over the past few years, and I mean the entire company, the implication of the comment is maybe not that crazy. Many people in the investing community, and many cablevision shareholders, believe the Dolan’s major goal lately has been to depress the value of the company in order to make it cheaper to LBO. There are a lot of ways to do this. For instance, goldplating. You build out a lot of infrastructure you dnn’t need right now. Cablevision has done a lot of that. This hides the true earning power of the company without reducing its real value.

I wouldn’t say the Dolans are deliberately mismanaging the Knicks. I don’t know what percentage of revenue the Rangers and Knicks account for. And I certainly don’t want to be the originator of a conspiracy theory.

But given that they have been trying to LBO the company for several years, the thought they might be deliberately mismanaging the company isn’t all that outlandish.

And by that logic, an LBO actually might be a great thing for Knicks fans. Once it’s taken private, it’s possible personnel moves might get more rational in a hurry. Wins in the NBA mean money….

21. Mike says:

The computer geek reference was based on the article and people that believe they can judge guys better based on stats alone rather than seeing them play. It was never said towards anyone commenting here on this board.

22. Kevin Pelton says:

And that assertion was made where in the article?

23. “What about, say, by letting expiring contracts actually expire in order to free up cap room to acquire really good players who don?t have major question marks?”

who would these be, just curious? is Rashard Lewis really a better addition than Zach Randolph? as I’ve said before here a bunch of time, cap space is decidedly overrated.

24. Ewing says:

Patrick Ewing is coming back to the Garden tonight. Ewing has been coaching Dwight Howard and there is no way Patrick is going to lose. Isiah is going down.

25. we need a game thread so I can point out that Mark Aguirre has been coaching Curry and Zach, and Aguirre actually won a couple of rings, unlike the good Mr. Rwing.

26. yeah, game thread, I am sort of excited to see what Curry does against Howard.

I thought that chatzy thing worked pretty well…

27. KAJB says:

The problem isn’t the use of stats, it’s how flawed the ones in use are. Until basketball stats catch up to baseball it’s better to just ignore most of them.

28. Brian M says:

“The problem isn?t the use of stats, it?s how flawed the ones in use are. Until basketball stats catch up to baseball it?s better to just ignore most of them.”

Care to expand on that a bit?

29. caleb says:

The problem with THIS stat (in Kevin’s article) is that I’m not sure what it’s really measuring. The best players within a particular group? Those most likely to improve? Those most likely to catch a coach’s attention and get more minutes?

What is a “breakout player,” exactly?

30. Kinda funny that Mike chastises KB for tossing out a red herring when he starts his comments in this thread with a red herring of his own — then repeats it.

No NBA stat analyst that I know of thinks that evaluating players purely by stats is a good idea. Not one. From what I can tell, ALL of us think that stats should be used in tandem with traditional scouting.

Some of us think that scouting can be improved by things like systematic tracking of additional statistical categories — something I’ve been doing for the past few years to get more information about defense.

As for the Diamond rating — I’m glad Kevin Pelton took a look at it. I’ve found it to be a useful tool. Not in identifying a list of guys who will become future All-Stars, but in coming up with a shopping list of guys who warrant further research — stuff like scouting, and additional statistical research.

31. Good defense by the Milwaukee Bucks drama boys, if only we could patch up the difference, this will be a Milwaukee good season. They are running, switching in defense, rebounding see those efforts. We don’t have yet the Bucks team down.

I which I could see some Bucks games live. I was looking for tickets all the good seats on ticketmaster were taken I had to check broker. And man you don?t want to do that especially for the Milwaukee Bucks. Thanks god there sites like Ticketwood which work as comparators here is the site
Bucks Tickets
http://www.ticketwood.com/nba/Milwaukee-Bucks-Tickets.php.

I like slam dunks that take me to the hoop my favorite play is the ally-hoop,
I like the pic n roll,i like the given goal its basketball yo, yo lets go!
Go Bucks Go!!!