Hi everyone, Long time, no post. There's no real reason for my extended absence, other than just not having time to fully commit to this board. I've been just as die hard about the W's as I was when I stopped posting, but I just haven't had the time to fully commit to this board. Anyway, with all the hullabaloo (yeah, I said it) over Baron's departure and Maggette's signing, I've started cruising the boards again to see what everyone thinks. I've got a couple of thoughts: 1) I personally would have signed Baron to a 5 year $65 mil deal if I were Mullin. But, I'm admittedly naive, and I continue to be disappointed by players who have a steep drop off in production after their contract years. I think I'm up to "fool my 947 times..." and the shame is definitely on me. Nevertheless, I'll still be surprised if Baron has the drop off in production next year that Mullin apparently thinks he'll have. 2) For everyone who thinks Mullin overpaid for Maggette, I respectfully disagree for two reasons. First, the W's were prepared to pay a big "Screw You, Clippers!" premium on the contract they offered Brand. Brand isn't worth the $92 mil the W's offered, he's not even really worth the $85 he got from Philly, I think the Clippers had him priced about right at $72 mil. Yet the W's were willing to give it to him to get him away from the Clippers. The consensus seems to be that Maggette is a good value at $40 mil, but not at $50. Even if Maggette's contract is for $50 mil (I can't believe we still don't have exact numbers yet), that's no more than a $10 mil premium. To ensure that the Clippers lose their two best players from last year, I'm fine with that. Second of all, Maggette (19.43) has a higher PER than Jason Richardson (18.48), and Richardson has three years and $40 mil left on his contract. I understand that $50 mil seems like a lot, but when you look around the league, it's very, very reasonable given his production. 3) Speaking of production...when you look at the PERs for the top shooting guards and small forwards in the league, Maggette is essentially tied with Brandon Roy for the 12th best PER among all shooting guards and small forwards. Only the following shooting guards had a higher PER last year: Ginobili, Bryant, Wade, Martin Iverson, and Roy (one one-hundredth of a point higher than Maggette). Only the following small forwards had a higher PER last year: James, Anthony, Butler, and Pierce. That's production, I don't care what any detractor says. Thinking about Maggette's PER is what leads me to the point of this post. When I was looking at PERs, I got curious as to which team's starting five had the highest cumulative PER. I was expecting there to be some correlation between the PER of a team's starting five and the team's total amount of wins. What I found blew me away. Not only is there a correlation, the correlation is so strong I'm amazed Hollinger hasn't brought it up yet. I tried to upload the excel spreadsheet I created (btw, I'm profoundly aware of my own dorkishness), but apparently I'm not permitted to upload that type of file. If you have excel and want to play around with it (which I highly recommend, everyone I've shown it too spends at least half an hour staring at it in disbelief), I'd be happy to email you a copy. Unfortunately, and not for a lack of trying, I can't figure out a good way to give you the numbers as part of this post. If someone knows, let me know and I'll repost. Anyway, here's the gist of what the numbers below mean: -Team: self-explanatory -Total PER: I added the PERs of the starting five together for each team (more accurately, the five players who were the primary starters for their teams) -PER Rank: I ranked each team from highest to lowest cumulative PER (I did this so I could more easily see how relatively high/low a team's PER was compared to the rest of the league when I was looking at the teams in order of their win total.) -Win Total: self-explanatory -Win Rank: Like PER rank, teams were ranked by the number of wins they had (so Boston and Detroit were both ranked #1, since they each had the most wins {59}, the Lakers were ranked 3rd because they had the 3rd most wins, and so on.) -PER/Wins Difference: Here's the money column (and where it gets tricky, but it's less complicated than it looks). This simply shows the correlation between where a team's PER ranked among the other teams compared to it's Win Total rank (So, since Boston had the 7th best cumulative PER and the best record, their PER/Wins Difference was +6, i.e. they overachieved, given their PER; New Orleans, on the other hand, had the 5th best PER and 5th best Win Total, so their PER/Wins Difference was 0, i.e. they finished with a Win Total that was in line with what their PER was, so they neither overachieved or underachieved; on the...third...hand, Milwaukee had the 18th best PER and 24th best record, so their PER/Wins Difference was -6, i.e. they underachieved because they won fewer games than what their cumulative PER would have suggested.) Now, here's the good stuff (as Dimitri Martin would say: Findings) 1. Last year, seven teams had a starting five with a combined PER above 90, they all won 51 or more games. 2. Of the 30 teams, 20 had a difference between their PER/Wins Difference of 3 or less. TWENTY TEAMS!!! There were five teams who significantly overachieved: Houston and Portland (who had the two absurdly long win streaks), and Boston, Detroit, and Philly (As Hollinger points out, playing in the Eastern Conference gives teams an additional 2-4 wins a year due to the weaker strength of schedule. If you take four wins from Boston and Detroit, their PER/Win Difference falls back in line, as they would have the same PER, but their adjusted Win Totals would drop them to a tie for 4th place, which would mean they'd fall in line with the rest of the league.) At the bottom of the PER/Wins Difference, you had five other teams: Utah and Dallas (i.e. the anti-Boston/Detroit, these teams' win totals suffered from playing in the tougher Western Conference), Miami (massive injury problems, their PER would be much lower if I included guys like Chris Quinn in their PER calculation), Milwaukee and Charlotte (other than having some injury problems and a continually shuffling line up, I don't see why they underachieved as badly as they did). 3. The Warriors had the 9th highest cumulative PER at 88.70 and the 12th highest win total (so their PER/Win Difference was -3, i.e. they underachieved). 4. The W's are essentially replacing Baron Davis (19.87 PER) with Corey Maggette (19.43 PER) in their starting line up. If everyone maintains their PERs, they should be just as good as they were last year. If Biedrins and Ellis (each of whom have increased their PER 2-3 points the last two years), are able to up their production a little, that could push the W's cumulative PER over 90, which would mean they'd be in line to win 50+ games if these numbers hold true. I really want to create similar spreadsheets for the previous years to see if this PER/Win total correlation holds up, but since PER stats for previous years are only accessible by going to each player's profile, it'd take more time than I'm able to spare to do. (I've ordered the teams by Total PER) Team Total PER PER Rank Wins Wins Rank PER/Wins Difference LA Lakers 99.35 1 57 3 -2 San Antonio 94.09 2 56 4 -2 Utah 93.76 3 54 8 -5 Dallas 92.92 4 51 10 -6 New Orleans 92.84 5 56 5 0 Phoenix 92.28 6 55 6 0 Boston 90.42 7 59 1 6 Detroit 88.98 8 59 2 6 GSW 88.70 9 48 12 -3 Orlando 86.99 10 52 9 1 Denver 86.91 11 50 11 0 Washington 85.85 12 43 14 -2 Toronto 84.78 13 41 15 -2 Cleveland 83.45 14 45 13 1 Charlotte 83.21 15 32 23 -8 Houston 82.28 16 55 7 9 Sacramento 81.67 17 38 18 -1 Milwaukee 80.79 18 26 24 -6 Atlanta 80.23 19 37 19 0 Indiana 78.58 20 36 20 0 Miami 78.54 21 15 30 -9 New York 78.25 22 23 25 -3 New Jersey 77.24 23 34 21 2 Philadelphia 77.14 24 40 17 7 Chicago 76.95 25 33 22 3 Portland 74.49 26 41 16 10 Memphis 73.66 27 22 27 0 LA Clippers 72.39 28 23 26 2 Minnesota 71.97 29 22 28 1 Oklahoma 71.69 30 20 29 1
I dont think the signing is a bad one, but I did have one serious question....dont Ellis and Maggette both need the ball to operate???
Hollinger? HAHA, nice work man. I admit I didn't read through it all. I'll just say, "if you say so!" 50 wins sounds good to me!
Great post wtwalker, and I agree, Maggette will be exciting and fun to watch for the Warriors in the Nelson system.
wtwalker, glad to see you again. This stuff is excellent. Very academic! For those that don't know what PER is, it's the Player Efficiency Rating metric popularized by John Hollinger of ESPN Insider. The formula goes as follows: Well... actually, the formula is very long and ugly looking so I ain't posting it. It's probably best served if you plug in such a formula in spreadsheets or some computer program. The idea is that the formula measures a player's performance per minute and based on his positive stats (points, steals, rebounds, etc) and negative stats (turnovers, poor ft%, poor fg%) against the league average. It also factors in team production, but it's mainly designed to amalgamate the differences in production between a running team and a halfcourt team. I've never been a huge fan of PER's simply because you need some kind of "veteran" sample size to more accurately gauge very young teams and also you can't really read young players that don't get playing time or get to venture outside their bit roles (Take for example how Joe Johnson got stuck on Boston's bench for a while or Michael Redd was a second string player behind Ray Allen). Big men can also suddenly turn around like Mehmet Okur or Jermaine O'neil. Then there's guys like Erick Dampier/Jerome James who could be good, but are only good in contract years. It's still an interesting find by Wtwalker because I've always had my hunches that the Warriors were going to be decent in terms of production. Whether this production equates to wins is yet to be seen. I've always believed that basketball is more about chemistry. Certain guys thrive better if they've got the right role players around them. Also, there's a lot of defensive areas that aren't recorded as stats such as the ability to individually shut down a scorer. I figure a guy like Bruce Bowen does not get many steals or blocks, but he noticeably makes it tough for perimeter scorers to do anything. Plus, he's got help defenders to seal off any passing lanes. There's no real statistic that can accurately reflect how well a team limits another team's offense and how that player factors the most into his team's limiting the other team's best perimeter or inside scorer. It's like one of those things an analyst would have to observe because there's too many scoring statistics and not enough measurements for hustle and effort on D (the intangibles). Plus, there's no measure of clutchness in important situations. If a guy scores 15 ppg and hits the game winning shot over the other team's best defender on multiple occasions, I think he's more valuable than a 20ppg game guy who fails hitting the wide open shots when you need him to hit. So obviously consistency and clutchness factor into a team's overall success (especially for those that have a scoring type of role in the rotation and for those that need to close out the 4th quarter). The PER is still a nice measurement and something that will at least look at the Warriors in a different way. Assuming Monta Ellis improves as a playmaker and our bench depth is not in question (we're also losing Pietrus and possibly Barnes/Azubuike), we could be in line to do about the same or slightly worse or maybe even overachieve if there's some real chemistry and intangibles going on. I for one believe that Mags will help the team in the areas most needed. One of which is to pound the enemy into the free throw bonus and improve inside finishing ability. Our D may be about the same or worse (depending on how you view size versus quickness and losing Baron Davis at point, but gaining more size at shooting guard with SJAX). On the other hand, this might not be a play to win season. This year was supposed to be about developing rookies and Nelson could be committed to doing so at the cost of wins. Who knows. It's both a strange and exciting season. Anything could happen much like the Blazers. BTW walk, I'll look at the excel spreadsheet. I love those things! I use it to win in fantasy basketball (factoring in the lame position restrictions - PF-C's, PG-SG's and SG-SF's for the win!). Although I failed because Yao and Bosh's health failed. 14 team league and you lose your best rebounder, scorer, fg% anchor, shotblocker, it's just pure doom. I think I might have had Allen Iverson and was surprised his body didn't break down. He's like the only pleasant surprise that would have fooled the PER in games played and maybe fg%. Also, players like Earl Watson really screw me up because Seattle's point guard rotation was so whacked. And guess what, I dumped him after I couldn't believe he was coming off the bench. He then goes into the starting rotation until he gets injured.
not a huge fan of PER myself. It's a stat nerd's tool to try and justify a players value, when watching a game or two will tell you HOW a player achieves his stats.
I think a mix of both is in order because sometimes observations can be really biased. Take for example watching Mike Dunleavy Jr. Guys like TV commentator/former NBA player Jim Barnett love his game. I just didn't see it very often and I'm sure many others didn't see it either. The guy couldn't shoot. In the end, it's about who can put the bucket in the hoop, control the pace of the game, and dismantle the other team's best defensive efforts while being able to play defense against the same attack. Guys like Dun could be good on other teams, but the Warriors simply weren't built for both Jrich and Dunleavy. It was a team that had little chemistry or a style of play. PER is definitely helpful in fantasy ball and I'm sure all the top fantasy ball players use it, in addition with throwing in their best hunches. BTW, Dmitri Martin is an acquired taste, but he's really clever, original, and extremely talented. The most interesting and innovative people I know are f#$#cking nerds and I really admire them.
Here is a list of rankings in PER http://insider.espn.go.com/nba/hollinger/s...er%2fstatistics Its hard to put a lot of weight into any formula. For example Calderon had the 26th best PER, and Ford had the 27th best PER. However Toronto felt one pointguard was a much more effective fit for the team. Andre Miller has a better PER than TMac, however Tmac has less playmakers around him when Yao was out and had to create a lot of the offense for Houston, which doesn't show in PER. Marbury had a better PER than Azuibuike. Battier was ranked 230th in PER, however a lot of people consider him a great glue guy. (Luke Walton, Tim Thomas, and Ricky Davis had better PERs) Posey was 214th in PER. Sam Cassell, Vujacic, Nate Robinson, had a better PER than S Jax and two spots below SJax is Stromile Swift. However if one takes into account things like if they spend more energy on defense, the chemistry they create, if they create better shots for teammates, are they boxing out well, etc.
Very nice work wtwalker. I'd say, I'd like to believe you. My one skepticism is on that basis that somehow all of these starters will have to keep the same PER as they had last season with Baron Davis. Baron Davis was the boss (as Pietrus called him). I am not sure if each player can keep up that same production without the boss around. Also in a more general term, not having a legit PG (based on Ellis' previous performances) may be something to bring down some of these PERs.
Wtwalker back with a vengeance! I, like most otheres here apparently, don't really buy into these all-in-one stats like PER but I find it intensely fascinating that you nailed the overachieving/underachieving teams; Boston I always felt overachieved in the regular season, obviously Houston and Portland way overachieved, the Bucks were underachievers having a good amount of talent yet still being horrible in a crappy eastern conference. I think that pert is spot on so that does give the stat some credibility in my mind. As for the Maggette signing, it is growing on me more and more. I'm not expecting 48-50 wins but I wouldn't be surprised by it either. I find your J-Rich-Maggette comparison interesting. They're two similar-yet-different players. Both "power guard" types as CR2 would say, both pretty solid rebounders for their position, both not the best defenders around, and both 20 ppg scorers but you wouldn't want them to be your #1 option or franchise guy. J-Rich is an ace shooting it from 3, mid-range, hitting pull up jumpers, he finishes very well but has to be opportunistic taking it to the rim because he has a poor handle. Maggette does his scoring by playing extremely physically on offense and getting to the line, forcing his way to the paint and into the bodies of opposing players. While they are different their production is similar and I would think Maggette is significantly more efficient scoring the ball. Will we be that much worse if we lose Baron, move Monta into Baron's role then sign a J-Rich-caliber player? Like Mullin said, the team will be different but not necessarily worse depending on how they play together. We could have much better defense next year with Wright and Turiaf getting big minutes and possibly a Dooling type player in the backcourt. Ball movement as a team could be much better without Baron chucking so much and the team having more of a pass-first attitude. Half court offense could be a lot better with Mags and Monta able to drive to the lane and get to the foul line. Rebounding will be better I imagine with Turiaf/Wright/Hendrix(?) helping Beans down low.
Walker -- great post. The correlation between team PER ranking and win ranking seems just TOO close to be random. I really think your work adds some validity to these numbers. Hollinger should be paying you for your submission! -- AO
<div class='quotetop'>QUOTE (wtwalker77 @ Jul 17 2008, 03:26 AM) <{POST_SNAPBACK}></div><div class='quotemain'>Nevertheless, I'll still be surprised if Baron has the drop off in production next year that Mullin apparently thinks he'll have.</div> I don't think it's next year Mullin was worried about, I think it's more like years 3, 4, and 5. I think he would have given Baron a 3-year deal in a second, but Baron wanted the security. 5 years is just too much risk IMO. I don't want to see this franchise hamstrung with big, empty contracts again. We're not as good without Baron, but we're infinitely more flexible for the coming years.
Cool list, Walker, I would like to see that Excel file as well. Maybe try .zipping the file first? One thing I wonder about though is the accuracy of simply adding up the PERs of all the starters. That's only valid if all the players play the same minutes. If Player A plays 35 minutes per game and Player B plays 25 minutes per game, then Player A's contribution should be weighted higher. I think you should normalize that to the number of minutes in some way. I'm thinking of Biedrins for one example, who starts but only plays 27 minutes/game. Some teams give a lot more time to bench players. It might be useful to figure in, for example, the top 7 players in minutes on each team instead of only the starters. EDIT: Hmm, thinking about it and PER is already minute-adjusted...however I'm still not sure adding up PERs is accurate...
I'd like to see a stat of High PER players that move to losing teams. The stats I'd like to see are Win Differential, and their individual PER rankings.
I have to admit that even though I'm a big sabermetrics guy, for some reason PER never caught on with me quite as much. I prefer pure +/- or other forms of calculation for basketball. Still, I think it's a great post that at least gives me some optimism, but based on my own skepticism, I still need more convincing that our team will get to the 45+ win range.
<div class='quotetop'>QUOTE (Zhone @ Jul 18 2008, 11:01 PM) <{POST_SNAPBACK}></div><div class='quotemain'>I have to admit that even though I'm a big sabermetrics guy, for some reason PER never caught on with me quite as much. I prefer pure +/- or other forms of calculation for basketball. Still, I think it's a great post that at least gives me some optimism, but based on my own skepticism, I still need more convincing that our team will get to the 45+ win range.</div> I think PER is interesting and has a place, but I'm not sure where. Comparing players? Ranking players? I agree with CR2 that it's useful in Fantasy League, but for a particular team it may not be as useful. I hope Corey lives up to his PER on the Warriors, but still don't think Monta is a starting PG. What's his PER at that spot? Gawd, I hope we don't have to find out.
