Dave Berri on the Bucks moves

[gobrewersbucks]

I thought we made some pretty good moves.
What is he talkin about?
Lue and Allen are good.
Jefferson is good.
I think Bogut got a lot better last year so it will now go up from hear for Bogut.
I question what hes talkin bout.

[Epicurus]

Games are won on the basis of scoring more than your opponent.

A differential between points scored and points allowed over the season is very correlated with wins.

Using regression analysis with box score variables on can determine a model( wieghted variable added) for that differentiation for all teams through studied years.

Those regression weights per box score variable then can be used to determine player performance (as related to winning).

This determination (WinsProduced per minute x. minutes played) tells of a player's relative contribution to wins.

Combining player's contribtuion and voila estimation of wins. If very close, then high predictive validity.

Model can then be used to predict the following season. And again these predictions can be after the season be compared to actual wins. Not 100%, why? The predictions are ceteris paribus (all things being equal) and all things are rarely equal: Players health might be different; skill levels may go up or down between seasons (not as usual as in other major sports); coaching changes; different personnel mix, etc. But again, NBA performance is much more predictable than in other major sports.

Much more predictive than observations and impressions of fans, subjectivity, wishful thnking, incomplete data base. But not perfect. Unfortunately the model has worked very well for analysing and predicting the Bucks.

Jefferson for two years has not been good. Allen has never been good. Lue gets closer, but is still a below average performer. Bogut improved around 9% last season (can't remember exactly, but have it somewhere), mostly, as europa has mentioned, the improvement in the second half of season was around or over 15% (from the entire previous season).

Jefferson returns to performance of three seasons ago + Bogut improve another 10% or so + Redd return to pre-injury two seasons ago = a top three that could drive a playoff team or better. Big assumptions.

[Dexmor]

Here is why he has no credibility. His first sentence was obviously Jefferson is not what he used to be. This past year was his best individual season. He played in 82 games. He is at prime age of 27 or 28. I honestly can not think of any reason for him to say that.

[REDDzone]

http://www.wagesofwins.com/JeffersonCareer08.html

The game is more than scoring points, dexmor.

[Epicurus]

Here is why he has no credibility. His first sentence was obviously Jefferson is not what he used to be. This past year was his best individual season. He played in 82 games. He is at prime age of 27 or 28. I honestly can not think of any reason for him to say that.

He scored more points, but all the other stuff leading to team wins went down and even his shooting percentage was lower than his career average. Now can he get back to where he was three seasons ago? Pehaps, he is not that old and I think maybe he didn't fully recover from his injuries the previous season. But Prof Berri is very credible regarding Jefferson the last two seasons. One of Berri's points for sometime is that fans and even gms get carried too far away with points scored (and not even the efficieny of doing so).

[paul]

I don't know anyone who thought Ruben was moved for purely basketball reasons, and i don't know anyone who thought that Des was a better pure player than Ruben. Ruben was moved because you don't want guys like Ruben on your team, it's a shame because he was our best player that season and a super hussle guy, but he was a bad dude. It doesn't take an economist to figure that move out.

I don't know how accurate Berri is or how often he is proven correct, but i do know that reading stuff like that makes basketball a lot less fun for me. That's fine if that's your thing, but I don't think anyone should take what he says as pure gospel, for as complete as his analysis is it still largely ignores extremely important areas of the game - as all statistical analysis does in one way or another.

[Epicurus]

The test of statistical analysis is not its completeness (has it considered all factors possible), but its precitiveness. I really doubt that any observer's, even trained ones, impressions are comprehensive either. What it appears Berri has done is use box score information in a way that produces verifiable conclusions about player's relative contributions to team wins. Pure gospel? Nope, because again things need be equal and never are and factors exists to modify estimations. But again all attempts to analyze, including most definitely fan impressions, leaves leaves unturned. I prefer the objective to the subjective, particularly the untrained subjective.

[Sigra]

What Berri predicted for Bulls for last season? Season before that they were good playoff team with potential to be better. People predicted them to be best at east last year. What Berri said? Did he predicted their doom? I mean that was almost the same team as one before that. Hinrich, Duhon, Gordon, Deng, Nocioni, Wallace etc. What great Berri said?

[LISTEN2JAZZ]

Sigra, of course he thought the Bulls would be great last year; everyone did. But remember he always says things like "if player X stays the same as last year, and player Y improves..." so he can never be wrong.

While looking for Bulls stuff I found more Bucks stuff.

http://dberri.wordpress.com/2008/02/05/ ... e-and-now/

So Larry Harris - the team’s general manager - claims the Bucks are very talented. But the team’s record indicates otherwise. And when we turn to Wins Produced, it becomes hard to buy the story Harris is telling.

Table One: The Milwaukee Bucks at the Midpoint of 2007-08

Table One reports two forecasts of the Bucks. The first assumes that each player - except for Yi Jianlian - performs as he did last year. That projection indicates that the Bucks would win 28 games this season. The second projection assumes what we saw the first half of this season will be seen in the second half. That projection indicates that the Bucks will win 25 games in 2007-08.

What do these projections tell us? Charlie Villanueva and Bobby Simmons are playing a bit worse than they have in the past. Desmond Mason is playing a bit better. And everyone else is playing just about as well this year as he did last year. In sum, we should not be surprised - given what these players have done in the past - the Bucks are not winning.

As I noted when I made my original projection, the Bucks only have three above average players: Redd, Bogut, and Mo Williams. All the other players on this roster are below average. So this team may be the most talented team Harris has seen in five years (although the 2003-04 team was the most productive of the bunch), but this team is still not very good.

Summarizing the Story

The Bucks are not performing well this season. That is an objective fact. Why the Bucks are not succeeding, though, is something people might question. Harris would like us to believe that he has assembled a “talented” team. But for that to be true, we would have to believe that last year’s team - the team that won only 28 games - was also “talented.” Clearly the Bucks were not very good in 2006-07. And given that virtually all these players are performing this season as well as they did last year, I don’t think we can conclude this is a talented collection in 2007-08.

So why does Harris tell us this team is talented? A cynic would note that Harris has two choices:

1. The team is losing because he chose the wrong players.

2. He chose the right players, but the team is losing for some other reason.

Given this choice, Harris is naturally going to gravitate to #2. Although one can think that Harris is motivated by self-interest to state that his team is talented when it’s not, it’s also possible that he sincerely believes this to be true. It’s possible that he believes that the current Bucks, if they tried very hard (or received the right coaching, or had the right attitude, etc…), this team could be successful.

The data in basketball, though, tells a different story. The numbers posted by basketball players - relative to what we see in football and baseball - are quite consistent from season to season. In sum, what you see is often what you get. Given this reality, to expect the current collection of players in Milwaukee to start winning consistently is unrealistic. Three slightly above average players does not a great team make. And all the coaching, attitude adjustments, etc… are probably not going to make much difference.

The solution is of course obvious. Milwaukee needs more productive players. In other words, for Milwaukee to start winning consistently, Harris is going to have to find even more talented players to come to Milwaukee next year. If that does happen, the Bucks will indeed start winning. But until that happens, the Bucks will remain stopped.

[Ruzious]

I read the whole thing. Basically Berri says we'll suck and suck big time.

He really doesn't mince words or temper comments. Here he goes at JA:

A possibility is to shift 2008 lottery pick - Joe Alexander - from small forward to power forward. Alexander’s college production, though, doesn’t compare favorably to an average small forward taken in the draft in recent years. Therefore - given that he probably won’t be a productive small forward — moving him to power forward doesn’t seem like such a good idea.

There's some value to his analysis, but the above comments show the flaw. Any system that doesn't combine statistical analysis with scouting is incomplete - at best. He's completely disregarded the scouting part. If he did, there would be adjustments for the fact that Alexander was a late bloomer - physically and skill-wise - who showed great improvement towards the end of the season - as the schedule got tougher and more meaningful.

[europa]

That Ruben thing was spot on. We had a net loss of 10 wins by replacing Ruben with Desmond. I truly do believe that.

No offense to Berri's formulas, but you didn't need to be a math major to figure out that going from Patterson to Dez would be a gigantic downgrade. There were posts from some of us non-math majors in this forum last summer warning that would occur. I don't think the Bucks had a net loss of 10 wins because they didn't re-sign Ruben Patterson at SF. If they had a net loss of 10 wins it was because Harris signed Desmond Mason to replace him. So the mistake wasn't so much the absence of Patterson, it was the player chosen to replace him.

[Sigra]

Sigra, of course he thought the Bulls would be great last year; everyone did. But remember he always says things like "if player X stays the same as last year, and player Y improves..." so he can never be wrong.

Therefore his predictions have no value at all. For example, you can't say "we can't win 45 games next season because acording to Berri we didn't change our roster enough to do that". Just like Bulls fans, before last season, were not able to say "we will not miss playoffs because acordingto Berri we didn't change our roster enough to do that".

Berri ignores chemistry and coaching and that is main reason why he was so wrong about Bulls (chemistry) and why he could be wrong about Bucks (chemistry and coaching).

Because of all of that I see no reason to talk about Berri and his work.

[Rockmaninoff]

Sigra, of course he thought the Bulls would be great last year; everyone did. But remember he always says things like "if player X stays the same as last year, and player Y improves..." so he can never be wrong.

Therefore his predictions have no value at all. For example, you can't say "we can't win 45 games next season because acording to Berri we didn't change our roster enough to do that". Just like Bulls fans, before last season, were not able to say "we will not miss playoffs because acordingto Berri we didn't change our roster enough to do that".

Berri ignores chemistry and coaching and that is main reason why he was so wrong about Bulls (chemistry) and why he could be wrong about Bucks (chemistry and coaching).

Because of all of that I see no reason to talk about Berri and his work.

That's why it isn't the Gospel According to Dave Berri. But, any method that can accurately predict the outcome 95% of the time (estimated number from Rockmaninoff), shouldn't be disguarded, but should be used in conjunction with thoughts of chemistry and coaching.

Alexander is the new Ruben Patterson. Hopefully.

[LISTEN2JAZZ]

Sigra, of course he thought the Bulls would be great last year; everyone did. But remember he always says things like "if player X stays the same as last year, and player Y improves..." so he can never be wrong.

Therefore his predictions have no value at all.

That's ridiculous. What he says is right more often than any other writer I know of (certainly better than Hollinger, Bucher, Ford, Screamin A, etc). How can something that so accurately describes real life, and is more predictive than any of us have "no value at all"? It sounds as if you think someone would have to be right 100% of the time in order to have value.

For example, you can't say "we can't win 45 games next season because acording to Berri we didn't change our roster enough to do that". Just like Bulls fans, before last season, were not able to say "we will not miss playoffs because acordingto Berri we didn't change our roster enough to do that".

I would never say something as definite as "can't." But I could say "based on the moves we made so far we would be very unlikely to win 45 games, and I'd be right. Because such an event would require greater than normal improvement from a few of our players. It could happen and I hope it does. But then if it does, we'll be able to say "Bogut and Sessions (or whoever it is) improved more than normal, and that's why we won." It won't be a complete mystery unexplainable by logic.

Berri ignores chemistry and coaching and that is main reason why he was so wrong about Bulls (chemistry) and why he could be wrong about Bucks (chemistry and coaching).

That's exactly right. But other writers focus on chemistry and coaching and ignore turnovers, offensive rebounds, and missed shots, and in the end they are less predictive than Berri is. Maybe some of them got the Bulls right, but got 10 other teams wrong.

Because of all of that I see no reason to talk about Berri and his work.

Do you see reason to talk about any writer?

[paulpressey25]

Adamcz brings up a decent point. For the Bucks to do well next year, Bogut and Ramon are going to have to take a leap. Which is plausible. And we'll have to get better contributions from other guys (i.e. Joe Alexander contributing a lot more than Desmond did last season)

What I guess I don't know is how Berri factors a rising tide lifting all boats. In other words, if Sessions and Bogut play well, does that facilitate an easier time for Redd and RJ, hence their winscores would go up?

[fam3381]

Wins produced is a good thing to pay attention to, because it cuts through some of the BS that looking only at raw numbers can provide. And like PER, it aggregates all of a player's stats into one metric, which is helpful. Its predictive ability isn't flawless of course, as teams like the Pacers, Bulls, and Nuggets blew away their predicted wins from 06/07 because players simply didn't play like they did the previous year. I'm not convinced any metric can really predict performance one year to the next, though for older players it's easier to assume they are who they were in previous years.

The fact that RJ wasn't as good the past two years is borne out by looking at PER as well, and hopefully given his age and a willingness to get back to a more all-around game means he can improve towards previous levels. But he's never really been a superstar talent so we shouldn't expect that sort of impact now.

On the topic of Mr. Patterson, what was Ruben's win score in previous seasons vs. his year in Milwaukee? What was it last year for the Clips before he was cut? My take is the Ruben situation was an example of win score providing value in a descriptive sense, ie figuring out who played the best on a crappy 06/07 team. But I'd guess that neither winscore or any other system would have predicted Ruben's big season or whether it was going to be repeated. IIRC Mason was actually much less worse last year than would have been predicted, though still short of Ruben's previous season.

In any case, the key to me is putting all of these metrics in context. People like Berri and Hollinger are very smart and I always like to consider their respective statistical metrics, but at this stage their lives' work is inextricably linked to WP48 and PER respectively. They have every incentive to treat their stats as the be-all end-all because that's what their credibility depends on. Doesn't make them worthless by any stretch of the imagination, but it's always good to keep in mind.

And unfortunately for Berri, I don't think his numbers will ever go into the mainstream unless there's an easy-to-use database where wins produced data can be found (does anyone have one?). My understanding is that basketball-reference has no plans to add Berri's stuff to their database, though perhaps an ESPN competitor will. Not that PER has become the holy grail stat either, but at least PER can be checked daily in-season and for previous years on ESPN/BR.

[LISTEN2JAZZ]

What I guess I don't know is how Berri factors a rising tide lifting all boats. In other words, if Sessions and Bogut play well, does that facilitate an easier time for Redd and RJ, hence their winscores would go up?

Maybe Epi can correct, but I don't think he worries a lot about one player impacting another. I'm pretty sure his stance is that players switch teams all the time, and end up putting up winscores that are remarkably similar to the ones on their old teams. Chemistry doesn't seem to normally have a very big effect on production.

[Bernman]

The logic Berri employs in Richard Jefferson and the decline of the New Jersey Nets is incredibly circular. In 2004-05 and beyond he's trying to assert that the Nets started to decline because all of a sudden the guards were their only above average players. The primary role of the guards on your team, well at least your point guard, is to ensure that the rest of your team is above average. Kidd didn't appear to have a problem with that task in 01-02 when LUCIOUS HARRIS, TODD MACCULLOCH, AND KEITH VAN HORN were considered well above average players according to Berri's formula. All of a sudden the task became much more difficult for Kidd after his 2004 microfracture surgery and subsequent aging process. Berri seems to think that basketball is an individual and not team sport.

[europa]

Berri seems to think that basketball is an individual and not team sport.

Team composition and coaching are arguably the two most important variables in my opinion in terms of predicting a team's success or failure and neither to the best of my knowledge are utilized to any significant degree by people such as Berri or Hollinger. This isn't fantasy basketball where you just assemble a group of "talented" players and put them on the court and watch the wins start rolling in. You need to have the right blend of talent working together with the right head coach to achieve meaningful success.

That's not to say these metrics aren't worth discussing. I don't put much stock in PER mainly because I've seen how faulty it's been over the years. But I don't dismiss it entirely and I realize it has its place in terms of player analysis. I know less about WinScores to be honest so I don't have a strong opinion about Berri's work one way or the other at this time.

I do wish someone would develop a strong statistical tool that measures individual defensive performance. As we've seen with the Bucks, for example, you can have all the offensive talent in the world and it won't mean squat if you have players who can't or won't defend and head coaches who can't or are incapable of generating productive defensive results.

[Alice]

basketball games are won on talent,discipline,sweat, heart and hustle, not by some trumped up formula. yes you can look smart if it makes you feel better by crunching stats that have already happened, but nobody knows how all the variables will play out in future games. with a different coach, a different game plan and different teammates, nobody knows what impact jefferson will make. not to mention the rest of the team. thats why changes are made.