Thursday, July 17, 2008

Great for the Sixers. Great for Brand? -- Part III

The Sixers are clearly a better team with Elton Brand than they were without him. In my last two posts, I've investigated the related question of whether Elton Brand made the right decision (in terms of being on the better team) by choosing the Sixers over the Clippers.

What I've found so far is that both PER and Adjusted +/- indicate that Brand made the correct choice in signing with the Sixers. To finish my trilogy of posts on the subject, today I'm going to see what Wins Produced (WP48, to be precise) has to say about this analysis.

And related to Wins Produced, Professor Berri put up an article the other day on how Brand's presence will help the Sixers. In addition to being as interesting as all his posts, his post made it very easy for me to find all the Sixers' WP48 information in one location. Thanks, Professor!

(He also put one up that details how the Clippers acquisition of Camby helps make up for the loss of Brand, but how the Clippers still aren't a particularly good team.)

Professor Berri also put up a post in May that looked at the Clippers' season, and that article was my source for the Clippers' WP48 information. I got my Baron Davis numbers from here.

We begin, as always, with the point guards. As with PER (but quite different from adjusted +/-), WP48 ranked Baron Davis (.177) and Andre Miller (.167) as very comparable point guards, but with Davis having the slight edge. At shooting guard position, Cutino Mobley (.029) held a slight advantage over Willie Green (-0.019). Andre Igoudala (.172) was, once again, ranked substantially ahead of Al Thornton (-0.081) (Corey Maggette, no longer with the Clippers, had a WP48 of .132). Finally, Chris Kaman (.233) was ranked ahead of Samuel Dalembert (.197), with each player having the highest WP48 on their respective teams.

As a result, based on WP48 the Clippers have an advantage at three positions while the Sixers have an advantage at just one position. However, the Sixers hold an overall edge in WP48 of .517 to .358. Thus, while a position by position comparison favors the Clippers based on WP48, the overall strength of the "other" starting four (excluding Brand) favors the Sixers.

If you move Thaddeus Young (.099) into the starting line-up as the SF and shift Andre Igoudala to the SG as seems likely to happen this year, then all of a sudden the Sixers and Clippers are back to having an advantage at two positions each. And the Sixers overall WP48 increases to .635 (increasing their advantage from .159 to .277. Again, even if Thornton improves, there's no reason to believe that he'll improve more than Young does. And I also don't think there's any reason to believe that Eric Gordon will have a better rookie season than Thornton did (or than Willie Green did last season). So if you move Young into the Sixers' starting line-up, I'd call the comparison a clear win for the Sixers.

So what have we learned?

Based on the advanced statistical metrics, it looks like the Sixers' "other" starters (excluding Brand) are better than the Clippers' "other" starters. With PER, the Clippers and Sixers each hold an advantage at two positions. With adjusted +/-, the Sixers hold an advantage at three positions. With WP48, the Clippers hold an advantage at three positions (although the positional advantage disappears when Young is inserted into the Sixers' starting line-up as seems likely to happen this season). In all three cases, the overall strength of the Sixers "other" starters is stronger than the overall strength of the Clippers "other" starters.

The verdict: Elton Brand made the right call. Welcome to Philadelphia, Mr. Brand.

(One final note: I did the comparisons using all three metrics comparing Igoudala to Thornton as the small forwards and Green to Mobley as the shooting guards because that seemed like the most accurate way to do it. Considering how interchangeable the wing positions are at times, I could have compared Igoudala to Mobley and Green to Thornton. For PER and
adjusted +/-, this change would have made no difference. For WP48, this change would have benefited the Sixers. Igoudala was better than Mobley and Thornton on all three metrics. Green was better than Mobley on two out of the three metrics, and Green was also better than Thornton on all three metrics. Considering the relatively low regard in which I view Green as a player, I found it a little shocking that he "outscored" Thornton on all three metrics, especially when you consider how well-regarded Thornton appears to be among Clippers fans. Statistical measures aren't the be-all and end-all (and Thornton was just a rookie), but if I were a Clippers fan I think these results would make me seriously question my belief in Al Thornton as a savior of the franchise.)

Wednesday, July 16, 2008

Great for the Sixers. Great for Brand? -- Part II

In my previous post, I compared the Sixers' and Clippers' "other" starting four (apart from Elton Brand) based on PER to try and see whether Brand made a mistake in deciding to come to the Sixers if he was just concerned about being a part of the better basketball team. Based strictly on the PER numbers, the decision to sign with the Sixers appeared to be the correct one.

Today I'll be checking to see if Adjusted +/- gives a different take on the situation.

You can find adjusted +/- stats at Sixers stats are here; Clippers stats are here. Baron Davis's stats from the Warriors come from here.

Looking at the point guards, Baron Davis (6.29) has by far the highest +/- of any of the eight "starting" players, while Andre Miller (-8.70) has by far the lowest. Compare that result to PER, which ranks Davis ahead of Miller, but only slightly. Also in contrast to PER, but this time in a good way for the Sixers, adjusted +/- ranks Samuel Dalembert (0.73) ahead of Chris Kaman (-2.23)

As expected with the shooting guards, Cutino Mobley (-3.63) and Willie Green (-0.75) both fall in the negative range, although surprisingly (at least to me) Green isn't in the realm of the terrible. On the other wing, Andre Igoudala (4.95) unsurprisingly trounces Al Thornton (-5.57). Corey Maggette (3.88) had a very respectable adjusted +/- (just as with PER), but he's no longer with the Clippers.

So based on adjusted +/-, the Sixers hold an advantage at three of the four positions. In addition, the Sixers' adjusted +/- total is -3.77 -- bad, but still better than the Clippers' overall total of -5.14.

As discussed with PER, this total ignores the likelihood that the Sixers slide Igoudala to the SG position and insert Thaddeus Young into the starting line-up at SF. Based on adjusted +/-, this change makes a huge difference in favor of the Sixers since Young had an extravagantly high adjusted +/- of 12.33 (not a typo; he had the third best adjusted +/- in the league last season). Putting Young in the starting line-up would leave the Sixers with their advantage at three positions, but it would increase the Sixers' overall total to 9.31-- increasing their advantage over the Clippers from 1.37 to 14.45.

I can't imagine Young will have such a high adjusted +/- next season. But even assuming he regresses towards an average performance and Thornton improves (as Clipper fans expect), it still seems like the Sixers come out ahead in this analysis. And I don't think anyone should expect Eric Gordon to put up better numbers this year than Cutino Mobley did last year, so I really don't see the addition of Gordon making much of a difference to this analysis either. I guess it's possible that if you just say Thornton improves and Green remains in the starting line-up than the Clippers can overcome the 1.37 adjusted +/- difference that exists between the starting line-ups based on what happened last season. But then again, we could make any assumption we want about how people will perform next year (after all, anything can happen) if we want to reach a certain result. So I think that we should just stick with the numbers from last season for now.

I will say that I was shocked to see that Miller did so poorly in adjusted +/-. He's generally considered a very good, if not quite top tier, point guard. Watching him play, he always seems to be doing positive things. In terms of the other advanced stats, he ranks well in PER, and I know that he also does well under Wins Produced (which I'll look at tomorrow). Hmm. Definitely something to think about.

(One side note-- Reggie Evans had an adjusted +/- of 2.75 last season, third best on the team. He'll most likely get reduced minutes to some extent with the arrival of Brand, but if he gets the vast majority of the back-up big man minutes (in a three man rotation of Brand, Dalembert, and Evans) then the Sixers will be very solid up front according to adjusted +/-.)

So what have we learned so far about whether Brand made a good basketball decision to sign with the Sixers over the Clippers?

According to PER, the Sixers and Clippers each have an advantage at two positions, but the Sixers are stronger overall.

Adjusted +/- leans even more in the Sixers favor, giving the Sixers an advantage at three positions, as well as an overall advantage.

Check back tomorrow to see what Wins Produced has to say on the matter.

Tuesday, July 15, 2008

Great for the Sixers. Great for Brand? -- Part I

There's no question that the Sixers are a much better team with Elton Brand than without him. Somewhat lost in the hubbub around Brand's decision to come to the Sixers is whether the move is good for Brand in a basketball sense-- that is, are Brand's chances of winning with the Sixers better than they would have been if he had stayed with the revamped Clippers.

The one place that this question has been most frequently raised (the only place, really, that I've come across it) is on the Clipper fan blogs, notably Clips Nation. From the perspective of many Clippers fans, Brand went to a situation with the Sixers that is decidedly worse than the one he would have had in L.A.

In particular, I noticed that Clipper fans were saying that the "other" starting four on the Clippers was clearly better than the "other" starting four on the Sixers. I wasn't quite clear on how they were coming to this conclusion, so I asked for clarification in the comments to this thread. I noted that going solely by Player Efficiency Rating, the Sixers and Clippers each seemed to have better players at two of the four positions, but overall the Sixers scored higher. I got some interesting and thoughtful responses from some of their posters. I didn't necessarily agree with their conclusions, but they definitely were well thought out and had plenty of merit.

So, first, I suggest you go click on the links above and read the exchange there (particularly the responses from Citizen Zhiv and Clipper Steve). Second, I thought I'd reproduce my side of the argument related to PER here, and expand it to also include the results based on looking at Adjusted +/- (Part II) and Wins Produced.

So, setting aside Elton Brand, do the Sixers or Clippers have the better starting line-up? To start with, I'll assume that the Sixers start the same players at the other four positions as last season-- Andre Miller, Willie Green, Andre Igoudala, and Samuel Dalembert. I'll also assume that the Clippers start the same players as last year, only with the recently signed Baron Davis at point guard.

(Note- these next three paragraphs are copied from my comments at Clips Nation. I've put changes in brackets.)
Based on PER [(found here; average PER for the league is 15.00)], Baron Davis (19.87) is slightly ahead of Andre Miller (18.51), Willie Green (12.91) and Cutino Mobley (11.56) are both pretty bad, but Green ranks slightly higher, Andre Igoudala (19.05) is far ahead of any Clipper SF (Maggette was at 19.43, but he’s not a Clipper anymore; Al Thornton was at 12.71), and Chris Kaman (17.62) gets the nod over Samuel Dalembert (15.62).

That gives PER totals of Clippers 61.76 and Sixers 66.09. So according to PER, the Sixers and Clippers each have an advantage at two positions, but the Sixers have the overall advantage. If you assume the Sixers will start Thaddeus Young (16.58) at SF this year and play Igoudala at SG, then the Sixers and Clippers still each have an advantage at two positions, but the Sixers’ overall advantage increases by 3.67 PER.

If [Clippers fans are] assuming that Thornton is going to improve, I don’t see any reason to assume that he’ll improve more than Young. If [Clipper fans are] assuming that Eric Gordon will make the difference (i.e. he’ll play SG and be better than Willie Green), then that doesn’t take into account the plan for the Sixers to switch Igoudala to that position (or my belief that Gordon is unlikely to do any better than Thornton’s rookie PER which was less than Green’s PER last season).

So, all in all, PER indicates that the Sixers have the stronger starting line-up, although it's a close call in many ways.

I think the response of the Clippers fans (and clink on this link to read their full comments if you haven't already) can be summed up as (1) PER is not the be-all, end-all metric for player evaluation (agreed!), (2) the difference between Kaman/Davis and Dalembert/Miller is actually much greater than PER indicates, (3) Thornton is much better than PER indicates, and (4) the Clippers have more (and better) shot creators, and that makes a big difference in close games and playoff games.

The first three responses are all basically subsets of the same argument that PER has flaws. I don't disagree with this point, but I'm not entirely convinced that their views on the relative abilities of the players involved is the correct one (notably, I'm not nearly as impressed with Thornton as they are). As for the fourth response, I'd say-- yes, the Clippers have better creators overall, but I think this advantage is maybe a tad overblown in the importance given to it relative to all the other aspects of a basketball game.

Am I right or are they? Who knows. You can let me know what you think in the comments.

And what does Adjusted +/- say about this debate? Check out my next post to find out.

(And since we're talking Clippers, I just saw in this post at Clips Nation that Marcus Camby has been traded to the Clippers from the Nuggets. And he was traded for basically nothing-- a trade exception and the ability to swap 2nd round picks. Wow, what a steal.)