Predict the Wizards record in 2010-2011

[Nivek]

doc: I agree with your thoughts on Wins Produced. I like Berri personally -- I've exchanged many emails with him and talked with him on the phone. But, I don't think his metric does all that he thinks it does. Which I've told him. We're at an agree to disagree point.

Berri makes many, many, many excellent points with his research, however. They're similar to points others have made at APBRmetrics, but they're still good. Among my favorite Berri findings: that stars don't drive attendance at home, they drive attendance on the road. What drives attendance at home? Winning. Work stoppages (strikes and lockouts) actually have not hurt professional sports leagues. Money and playing in the NBA are divvied up not by activities that cause winning, but by what Dan Rosenbaum calls "glory stats" -- per game points, rebounds and assists. Almost all NBA salary can be explained by points per game.

Despite Berri's claims to the contrary, I think his metric overvalues rebounding. Specifically, the defensive rebounder gets credit for a defensive stop, when defensive stops are almost always caused by teamwork. He sorta gets around this with a defensive adjustment based on team stats. A lot of folks don't like the team adjustment because with a team adjustment you can make almost any stat correlate with winning. I've done the trick myself in a spreadsheet -- I calculated team level PER, Wins Produced, NBA Efficiency and PPA (my own metric I've used off and on) and found that with an adjustment for team defense, they ALL have a .95 or better correlation with winning. And, although efficiency is extremely important, I think he over-punishes missed shots.

Kevin Pelton has an excellent critique of Wins Produced in which he points out that while WP may do reasonably well at the team level, it's not proven that it's a good way to evaluate individual players. You can google Pelton's comments up over at APBRmetrics. I also think he wrote something up for Basketball Prospectus.

[GilArenas88]

Put me down for 42 wins. We're young but have a lot of room to grow, along with a few backcourt vets.

[Illuminaire]

Thanks for those posts, Doc and "Nivek." You saved me the trouble of a long, rambling discourse on the many reasons why Wins Produced drives me crazy. =p

[Nivek]

Quick point -- I did a study of my own looking at the relationship between various stat categories and playing time. My findings echoed Berri's, and Rosenbaum's -- playing time is decided overwhelmingly by per minute scoring. For the Wizards under Eddie Jordan, playing time could be explained by per minute scoring and per minute steals. The Suns under D'Antoni were interesting because their playing time correlated with efg -- players who shot better from the floor played more. For the Wizards, there was a slight negative correlation between efg and playing time, meaning that players who shot better were more likely to get less playing time under Eddie. The numbers there were probably picking up something else, but it was still interesting.

League-wide, here are the categories that had negative correlations with playing time: fouls (-0.60), offensive rebounds (-0.21), blocks (-0.09), and total rebounds (-0.06). All big man stats.

If you go by the positive correlations, here's what coaches "liked" -- scoring (0.58 correlation), assists (0.36), and turnovers (0.31). Interestingly, 3pt shots (made and attempted) had a weak correlation (.14 for made 3FG), although the latest research shows that 3pt shooting is valuable.

[Joe_Wiz]

I usually do this based on what I've seen of the players in the past, and that generally means what I've seen them do in a Wizards uniform. That means this year what I know is pretty much a jacked shot. Of course, that won't stop me from having an opinion.

There is an enormous amount of uncertainty this year, especially around three of our players: Arenas, Howard, and McGee. Each of these has the potential to play at a very high level -- or to give us nothing or worse than nothing. I'm going to go with two analyses and corresponding predictions: realistic (unofficial) and optimistic (which I will use for my official prediction).

Realistic Analysis:

Back court: Wall will be good, but he is a rookie, has limited range, and commits too many turnovers. He will be a net positive, but will not turn the team around in one year. It's impossible to predict what we'll get from Arenas, but here's my attempt. He'll give us much more over the course of a season than he has since his injury, and he'll be good overall, but both on and off the court he'll still disappoint those of us who still want to believe in him. Hinrich will be a solid anchor. Young will be marginally improved, but will still be Nick Young.

SF: This position will be a major weakness until Howard gets back. He might come back late, will have to shake the rust off, and the team will have to adjust to his presence.

Big men: Blatche will be good, but double teams will limit his offensive effectiveness much of the time. McGee's offense will be improved, mainly because of ally-oops. His defense, especially team defense, will continue to be a work in progress. the remaining big men will be unexceptional, at best (THIS season).

Realistic Prediction: 33 wins, with no more than 15 coming in the 1st 1/2 of the season.

Optimistic Analysis (take the realistic analysis and modify as follows):

Arenas will return to elite offensive form, though fewer minutes and having the ball in his hands less will make for less gaudy stats. He will be somewhat better on defense. There will still be nagging injury and judgment problems, but nothing serious.

Flip will cobble together a rotation at SF (or 3 guard offenses) that will work OK until Howard gets back. Once he gets back, he will be very good, and Flip will have a wide variety of options as to who to play behind him that will be remarkably effective based on match ups (e.g., neither Hinrich nor Booker is a suitable starting SF, but both can be nice back ups against the right opponents).

McGee's improvement will be noticeable in conditioning (and asthma treatment) and shot selection (mainly dunks), and other marginal improvements will add up. One big man not mentioned above (my bet is on Yi, but it could be Armstrong or someone else) will be a reliable backup, rebounding and defending credibly for 20+ minutes a game without crippling the offense either.

Optimistic (and official) Prediction: 42 wins. They will still get off to a shaky start, but not a disastrous one. In the second half, they'll be pretty good, winning around 25 of their last 41 games.

[Joe_Wiz]

This is not part of my official prediction, but I never really gave voice to my inner pessimist. If Arenas and Howard due to injury or other reasons are unable to contribute in a positive way to this team this year, and if McGee does not noticeably improve, this team could be very, very bad. Add an injury that keeps Blatche or Wall off the court for any length of time, and we could be talking historically bad. The realist in me says that any of these things alone is not terribly unlikely, but the combination is very unlikely (albeit possible).

[verbal8]

This is not part of my official prediction, but I never really gave voice to my inner pessimist. If Arenas and Howard due to injury or other reasons are unable to contribute in a positive way to this team this year, and if McGee does not noticeably improve, this team could be very, very bad. Add an injury that keeps Blatche or Wall off the court for any length of time, and we could be talking historically bad. The realist in me says that any of these things alone is not terribly unlikely, but the combination is very unlikely (albeit possible).

Unfortunately this is the Wizards/Bullets you are talking about

I don't think things will go that badly, but I don't expect the team to get a lot of lucky breaks either.

[Joe_Wiz]

Man, I knew I hadn't been following the Wizards as carefully the last year or so as usual, but I actually completely forgot about the existence of Al Thornton until I saw him mentioned in another thread. Wow, that's embarrassing.

That should probably add marginally to my win totals*, but my predictions were not exactly based on a sophisticated metric, and his presence would effect my realistic (and unofficial) prediction more than my optimistic official one, so I'll stand pat.

* It's not that I think he's that great, but he's one more serviceable solid player on a team with a lot of young or otherwise very much unproven players.

[The Consiglieri]

I'm with Illuminaire. My feeling is that the Wiz have so many players that will be heavily in the rotation that have little experience as starters and little background to extrapolate stats on that it's damn near impossible to actually accurately project anything.

Wall: Zero NBA games.

Areanas: Hasn't played a full season in 3 or 4 years, i forget at this point.

Howard/Thornton/whatever: Never a consistent starter.

Blatche: 30+ games as a regular, never a season long starter.

McGee: Never a season long starter.

The bench is similar.

How the hell do you project anything w/the kind of limited information available? Projections for Blatche's performance preceeding his elevation to the starting lineup were probably nowhere near what he actually did.

It's just really hard to know. Like all fanbases, i think the average fan posting here skews to a more favorable result except for the chicken littles (myself being a perenial member of that fun contingent), which skews expectations heading into any season. Few want to start a season with no hope. How else to explain 2/3's of the redskins fan base expecting year to year for the past 18 years a season of 8 or more wins? I like to set a floor and a ceiling. My floor for this team is about 24 wins. Last years team won at a clip of around there after the vets were shipped out, i figure w/some bad breaks and injuries, its a 20-25 win team in terms of a floor. Now if everything turns out well? I could see 38-42 wins. A big 3 of Arenas, Wall, and Blatche all playing quite well, add McGee growing as a player, and Howard healthy with some decent options off the bench? Maybe we shock the world and grab the 7 or 8 seed. I actually think the interests are served more by landing another top pick in the draft and going from there, though if the draft is as weak as some argue, maybe success is more in our long term interests. I just know that we were finally lucky, for once, in May, and instead of a team loaded with entitled, mediocre vets, we have a young and hungry group. Much more interesting, even w/a ton of growing pains.

[Illuminaire]

Quick point -- I did a study of my own looking at the relationship between various stat categories and playing time. My findings echoed Berri's, and Rosenbaum's -- playing time is decided overwhelmingly by per minute scoring. For the Wizards under Eddie Jordan, playing time could be explained by per minute scoring and per minute steals. The Suns under D'Antoni were interesting because their playing time correlated with efg -- players who shot better from the floor played more. For the Wizards, there was a slight negative correlation between efg and playing time, meaning that players who shot better were more likely to get less playing time under Eddie. The numbers there were probably picking up something else, but it was still interesting.

League-wide, here are the categories that had negative correlations with playing time: fouls (-0.60), offensive rebounds (-0.21), blocks (-0.09), and total rebounds (-0.06). All big man stats.

If you go by the positive correlations, here's what coaches "liked" -- scoring (0.58 correlation), assists (0.36), and turnovers (0.31). Interestingly, 3pt shots (made and attempted) had a weak correlation (.14 for made 3FG), although the latest research shows that 3pt shooting is valuable.

Berri's assessment of the perception bias that permeates professional basketball is an area where I do agree with him. It's a somewhat mind-boggling trend that even highly paid, intelligent professionals often fall into the same trap as everyday fans when it comes to overvaluing "glory stats."

Unfortunately, as you pointed out, Berri similarly overvalues his own favored stat, defensive rebounding. It is difficult for me to take other aspects of his work seriously as long as Winscore relies on a deeply flawed over-simplification of what creates defensive stops.

I do appreciate the different voice he adds to the overall discussion. He just reminds me of a good friend of mine who is similarly intelligent, insightful... and utterly, bullishly stubborn once he's decided on an idea, no matter how much evidence you bring up to the contrary.

[Chocolate City Jordanaire]

Kevin Pelton at Basketball Prospectus has updated season predictions based on his statistical projection system, which he calls SCHOENE. He plugs in the numbers and simulates the season 1,000 times. The average from those simulations says the Wizards will win 23.1 games and finish with the league's worst record. In those simulations, the Wizards and Clippers were the only teams to not make the playoffs even once.

Much as I respect Pelton's work from way back in Seattle days, if his simulation has neither the Clips nor the Wizards in EVER he's flat wrong.

Griffin will be a beast. Him with this year's Kaman? They will be good. Aminu will be good. Baron just has to pass and he will. Gordon played with Team USA. Rasual can play. Vinny is a good coach. Clips have a real good shot at an eigth seed. LAC was something like 19-23 last season before finishing weak.

As for the Wizard projection, it's not based on what I think will be a confluence of guys having bounceback years. It cannot account for Wall to McGee, or Yi in Flip's system. We all know Gil is the X factor. We remember Gil with Hughes working. How much better could that be with Wall? Hinrich is a factor on defense.

But then again, the Wizards have a real soft front line and I don't think Flip handles adversity well nor does he think out of the box. Season could get real ugly. So, about the Wizards more than the Clips his simulations could be right IMO.

[Chocolate City Jordanaire]

Also at Basketball Prospectus, Bradford Doolittle uses a different statistical prognostication method, which he used to simulate the upcoming season 10,000 times. In his system, the Wizards finish 28th on average with 30.0 wins. Doolittle predicts the Wizards as having a 1.0% chance of finishing .500, a 4.8% chance of making the playoffs and no chance of making the Finals or winning the title. In his 10,000 simulations, the Wizards' high in wins was 45, their low was 8.

Not a stats guy, but wonder what a normal distribution curve would look like. If the team has a 1% chance of 41 wins, that means it has (EDITED) a slightly greater than 1% chance of 40 wins, even greater if ever so slightly of 39 wins etc down to the mean of 30.

Where did Bradford cut off making the playoffs? Just seems to me 4.8% is low. I could see 10%.

[Nivek]

In a normal distribution, if there's a 1% chance of 41 wins (at the high end), there would be a slightly better odds for 40, slightly better still for 39 and so on. If the high end is 41 wins, the odds would favor something closer to 30 wins, which Doolittle's system predicts.

Doolittle didn't set a cut off for making the playoffs. He simulated the upcoming season 10,000 times. What that 4.8% means is that out of those 10,000 simulations, the Wizards made the playoffs 480 times.

Same thing for Pelton, although he simulated the season "only" 1,000 times. In all those simulations, the Wizards didn't make the playoffs.

I think you guys pegged some of the difficulties of statistical predictions for the upcoming season. Pelton's way of handling that is to use his system to generate historically similar players and use those similars to predict a player's performance for the upcoming season. I'm not saying he's right. It's something to think about, though.

[Hoopalotta]

Obviously, the Stat gurus don't have access to our storehouse of local knowledge.

Posted by: melodious_thunk | October 26, 2010 3:06 PM |

@KTV1

I know you're not one of the haters, but don't be among the last to figure out that the Wizards are built to be playoff team NOW. The ingredients are all there. They just need to have to have a little good fortune with their veterans' health while the youngsters continue to mature.

I submit to you that this preseason the Wizards won all three games in which they had at least 80% of their projected starters available for the full game. In other words, the team, as intelligently and painstakingly constructed by one Ernest Grunfeld, has been built with the attention to detail and precision of a Swiss watch - when their core players are suited up, a "W" for our Wizards is there for the taking. Granted, since injuries occur throughout a season, the team's young players will have to get accustomed to changing their mindset from that of guys who come off the bench to guys who are willing and able to start when called upon. I have no doubt that they will get better at this aspect of the game sooner rather than later and the Wizards will make the playoffs despite what the offkey haters spew from their acidic constitutions.

The official start of the Grunfeld Era is nearly upon us and Ernest Grunfeld is no longer working with one hand tied behind his back. Wizards Faithful, I ask: Are you ready?

In Grunfeld We Trust!!

[Illuminaire]

There is some truth there. Historically, players with a PER of 24 don't come back from 3+ years of injury woes and play at a high level. Boneheads who spent four years as a bench scrub very rarely transform into nigh-elite power forwards. Low bball-IQ pogo sticks usually bounce themselves right out of the league.

By all historical trends, we don't have much to feel good about. But you know what? I feel good. We drafted John Wall. We jettisoned most of our one-way players. We have a plan, we have leadership, we have a hungrier group of guys than before. Did I mention we have John Wall? I think at least two of those three chips fall in our favor. I think we avoid crippling, season-killing injuries. I think the Great Wall leads us to a winning record.

I base this on the most advanced of all metrics - hope. Man, it feels good to have hope for the Wiz again!

Put me down for 42-40, seventh seed. Book it.

[LyricalRico]

Put me down for 30-52. I just don't think the Gil/Wall backcourt will excel. I'm not saying it'll be a colossal failure and both guys will average single digits, but it won't be as explosive as some predict IMO. I'm also wary of Josh Howard playing a lot of games. Add in a lack of veteran inside presence on the boards, a bench that I think will go through scoring droughts, and I think that's a recipe for a losing season. This is a rebuilding team that I think will play hard, be in a lot of games, and in the end will take a step forward - but they still have a long way to go IMO.

(The only thing that could change that to 37 or 38 wins is Josh Howard coming back early and having a consistent impact while not getting injured again. I think he's the real wild card. I'm being consciously pessimistic in assuming he misses a lot of time, both on the front end and with other injuries.)

[1974onWizFan]

Reality intrudes a little (but not by much) and as a result I hope for 41 wins this year.

Its based based primarily on the preseason play, that while not showing much defensive chemistry, did show consistent offensive capability. Also, this team could be very deep into about the 10th to 12th spots on the team. Thus they could survive the inevitable injuries that occur every year and change the lineup to match up well with most teams.

If Flip and the coaching staff can get the players all on the same page, unlike last year, the Wizards will be fun to watch.

[Tyrone Messby]

41-41 Eighth Seed

[closg00]

Obviously, the Stat gurus don't have access to our storehouse of local knowledge.

Posted by: melodious_thunk | October 26, 2010 3:06 PM |

@KTV1

I know you're not one of the haters, but don't be among the last to figure out that the Wizards are built to be playoff team NOW. The ingredients are all there. They just need to have to have a little good fortune with their veterans' health while the youngsters continue to mature.

I submit to you that this preseason the Wizards won all three games in which they had at least 80% of their projected starters available for the full game. In other words, the team, as intelligently and painstakingly constructed by one Ernest Grunfeld, has been built with the attention to detail and precision of a Swiss watch - when their core players are suited up, a "W" for our Wizards is there for the taking. Granted, since injuries occur throughout a season, the team's young players will have to get accustomed to changing their mindset from that of guys who come off the bench to guys who are willing and able to start when called upon. I have no doubt that they will get better at this aspect of the game sooner rather than later and the Wizards will make the playoffs despite what the offkey haters spew from their acidic constitutions.

The official start of the Grunfeld Era is nearly upon us and Ernest Grunfeld is no longer working with one hand tied behind his back. Wizards Faithful, I ask: Are you ready?

In Grunfeld We Trust!!

Thunk/Rico has struck again

[LyricalRico]

Melodius is spot on about the longterm "Ernie era", but I disagree with him about playoffs this season. I like what Ernie did to go young and fill out a stop-gap roster for one season before the lockout and while they wait to see if they will be able to move Arenas. But unless Josh Howard is healthy and has a big impact, I don't see this team competing for the 8th seed. And that's not an indictment of Ernie, since playoffs weren't an expectation in the first place. I just think our success lies post-lockout.