Power Rankings #6: A change is going to come

You'd thought I'd forgotten about the rankings before leaving for the holidays, right? I just decided to wait until the Christmas Eve break and give you an extra special post. Oh, and we'll have some Christmas surprises too. Case in point:




Yes, I did update the Visual Rankings just in time for Xmas.

A third of the season is almost in the books and I'm guessing a lot of you feel like you've got a handle on most of the NBA's teams. Aside from Miami and Indiana, the East is pitiful. On the other hand, the West is top loaded -- murderous and filled with contenders. Portland is better than you thought. Houston is schizophrenic. San Antonio does not care about the regular season. Milwaukee, Utah, Orlando and Philly seem to be checking out.

But ttrades are coming. Prepare for the final rankings before the some of the big dominoes start to fall (Rudy Gay was something like a 3-for-1 in a second-hand store; it counts, but no one really cares).

Let's get to the rankings:

BoxScore Geeks Power Rankings as of 12/24/13

BoxScore Geeks Power Ranking as of 12/23/13
Rank Change Team Point Margin Sched Rk Sched PM Orank Drank
1 Up 3 Indiana Pacers 9.7 26 0.9 12 1
2 Up 5 Los Angeles Clippers 7.4 6 -0.7 4 7
3 Down 1 Oklahoma City Thunder 7.2 4 -1.0 7 3
4 Down 3 San Antonio Spurs 5.7 7 -0.5 3 4
5 Down 2 Miami Heat 5.6 30 2.5 1 22
6 Up 6 Phoenix Suns 4.5 25 0.8 9 14
7 Down 2 Portland Trail Blazers 4.4 20 0.2 2 16
8 Up 5 Atlanta Hawks 4.0 22 0.4 8 20
9 Down 1 Minnesota Timberwolves 2.0 3 -1.1 10 9
10 Down 4 Houston Rockets 2.0 19 0.2 6 12
11 Down 1 Golden State Warriors 1.4 28 1.0 15 6
12 Down 1 Dallas Mavericks 1.2 13 -0.2 5 24
13 Down 4 Denver Nuggets 0.3 10 -0.3 21 8
14 Up 2 Toronto Raptors -0.2 15 -0.1 18 11
15 Up 6 Detroit Pistons -0.2 5 -0.9 14 13
16 Down 2 New Orleans Pelicans -0.4 23 0.5 11 23
17 Up 3 Charlotte Bobcats -1.1 27 0.9 28 5
18   Washington Wizards -1.2 29 2.1 20 19
19 Down 4 Boston Celtics -1.5 9 -0.4 23 10
20 Down 3 Los Angeles Lakers -1.8 1 -1.9 22 15
21 Up 2 Brooklyn Nets -1.9 11 -0.3 16 29
22 Down 3 Chicago Bulls -2.3 17 0.0 29 2
23 Up 2 New York Knicks -2.6 8 -0.4 17 25
24 Down 2 Orlando Magic -3.3 14 -0.2 24 21
25 Up 4 Sacramento Kings -3.7 12 -0.2 13 28
26 Up 2 Memphis Grizzlies -4.1 24 0.6 19 26
27 Down 1 Utah Jazz -4.7 2 -1.8 25 27
28 Up 2 Milwaukee Bucks -4.8 16 0.0 30 17
29 Down 5 Cleveland Cavaliers -5.7 18 0.1 27 18
30 Down 3 Philadelphia 76ers -6.3 21 0.3 26 30


Those are our Power Rankings through a 33% of the season (and it includes pace and schedule adjusted offense and defense). Did you enjoy the Christmas colors? Let's talk highlights (ICYMI, all the ratings went up as a visual post ):

  • #1 Indiana Pacers Neutral Site Win:80.3% (Up 3,#12 in offense, #1 in defense): The Pacers have annihilated their last three opponents (Houston, Boston, and Brooklyn on the road on a back-to-back) and the model loves that. It has no way of knowing that Brooklyn was without their best player (Lopez), which makes it less shiny. Indiana's defense remains the best in the league, but their offense is starting to make some noise. The big story for this team continues to be the outstanding play of their top 6 in minutes played (all 0.138 WP48 or better) and their remarkable health. Continue to use pen to mark them down for the ECF if they stay healthy.
  • #2 Los Angeles Clippers Neutral Site Win:74.4% (Up 5,#4 in offense, #7 in defense): Through a third of the season, Chris Paul is the Wins Produced MVP, with an ungodly 0.367 WP48, which would be the highest by a point guard not named Magic or Stockton. He's damn good. As much as I gave Doc crap about their defense, I have to give him kudos. If you consider the amount of injuries this team has right now? Definitely a contender. Seems like the issue is always injuries and streakiness with this group. They're on a tear right now, and that got them to the 2 spot. They need to pray for it to hit in April. They hold the belt, but can they make a run at the title?
  • #3 Oklahoma City Thunder Neutral Site Win:73.9% (Down 1,#7 in offense, #3 in defense): OKC is now 16-2 over their last eighteen, with their only losses in Rip City and in Gay-less (and thus improved) Toronto -- both on the second night of a road back to back. Quietly OKC has rounded back to form. But it's a slightly different form; KD is still KD (0.308 WP48) and Russell and Ibaka are recovering and slightly diminished, respectively (0.101 WP48 and 0.177 WP48). The big story is the emergence of Jackson, Lamb, and Adams as significant contributors. The talent is there, it comes down to the rotation. Do you trust Scotty Brooks? Do you believe OKC can win the title with the worst coach in the Western Playoffs?
  • #4 San Antonio Spurs Neutral Site Win:69.2% (Down 3,#3 in offense, #4 in defense): My process for handicapping a San Antonio game is to always to ask: is Kawhi playing? If so, then they can win (if not see the OKC game). Every time San Antonio rests their supposed big three everyone is surprised by their other guys either winning or coming close. They shouldn't be. I would pick the Spurs minus Tim, Tony, and Manu to win over 52 games. They have six players playing at a star level (>0.180 WP48): Kawhi, Manu, Tiago, Tim, Patty, Marco, and Tony. They have zero to prove, but the other night at Oracle? Signature win. For them, it remains all about banner number five.
  • #5 Miami Heat Neutral Site Win:69% (Down 2,#1 in offense, #22 in defense): The Heat are exactly where they want to be at this point in the season. Did you know 21-6 is their best record after 27 games as the Heatles? And thanks to their split with Indiana they keep the possibility of a one seed open. LeBron at 0.384 WP48 is hands down still the best player in the world. Yes, Miami has some depth issues -- which will always keep their regular season record down -- but once the rotations get shorter, this team is still the two-time defending champ. They may also have a solution to the depth issue deep on their bench (Mr. Oden come on down). Miami has the luxury of being able to take its sweet time.
  • #6 Phoenix Suns Neutral Site Win: 65.5% (Up 6,#9 in offense, #14 in defense): Hey, Phoenix is back in the top 10. They're 8-1 over their last nine and continue to look like a legit playoff spoiler out West. They're a solid eight deep, young, and cheap. Excellent team for the future. I think this year is about growth and experience for them.
  • #7 Portland Trail Blazers Neutral Site Win:65.1% (Down 2,#2 in offense, #16 in defense): Portland continues to win even if their point differential is below their win total (translation: expect a correction). Of note is that their rotation boasts eight guys over 300 minutes played, with everyone else not cracking 60 minutes. This is Tom Thibodeau territory. This probably means that they're a weaker playoff team than regular season team because there is really no way to shorten up the rotation.
  • #8 Atlanta Hawks Neutral Site Win:63.9% (Up 5,#8 in offense, #20 in defense): With an impressive 4-2 run (losses on a back-to-back in NYC and an OT loss in Miami), the Hawks finally look like they're putting it together. They're not even at full strength (at 0.170 WP48, Gustavo Ayon will help). I like this team and I think they should resist the urge to blow it up. The Hawks should just hold for that lottery pick from Brooklyn, which is looking very high indeed. At this rate, the 2014 and 2015 picks from Brooklyn might be lottery picks both.
  • #9 Minnesota Timberwolves Neutral Site Win:57.1% (Down 1,#10 in offense, #9 in defense): As much as the model likes this Wolves team, I remain skeptical until they can put together a run. Kevin Love is a top five guy and the best power forward in the NBA (0.333 WP48), but their depth remains terrible and they haven't won four consecutive games this season. I can't take this team seriously.
  • #10 Houston Rockets Neutral Site Win:56.9% (Down 4,#6 in offense, #12 in defense): The Houston team we have now is a bit of a schizophrenic mess. D12 looks like himself again. Daryl needs to pull the trigger on an Asik deal and stop playing brinkmanship games with potential trade partners. I understand trying to drive up the price, but a few down weeks in the West could have you on the road in round 1 or out of the playoffs. Houston has all the horses to contend now and I wouldn't waste the opportunity. I also would not have minded seeing Asik in Celtic green (Bass is a great fit for playing next to Dwight and Lee would be the best wing defender on the Rockets). I do fear he might wind up a Laker (Pau, Hill, and a pick swap for Lin, Asik, and Montiejunas -- who says no?)
  • #11 Golden State Warriors Neutral Site Win:55.1% (Down 1,#15 in offense, #6 in defense): Golden State looks to be in the second tier in the West right now. Andre being back should be a substantial help for this team though. After some subpar play to start the year, David Lee has been on fire to get himself back to a respectable .131 WP48. The typical Warrior injury crisis has now moved to big man depth. Basically, if you have size and are alive, Golden State might have room in the rotation for you and this is with Bogut healthy. Cross those fingers warrior fans because we need you in the playoffs.
  • #12 Dallas Mavericks Neutral Site Win:54.1% (Down 1,#5 in offense, #24 in defense): The good news for Dallas is that Brandan Wright is back and he's been fantastic (0.420 WP48 in four games). The bad news is that their defense is still piss-poor. Back-to-back losses at home to Toronto and in Phoenix could come back to haunt them. I like this team but they could end up with a great record and the 9 or 10 seed. They need to hope that Brandan is the answer for them or that Cuban can find an answer in the trade market.
  • #13 Denver Nuggets Neutral Site Win:51.2% (Down 4,#21 in offense, #8 in defense): Ty Lawson has been hurt and sick and this has affected his play lately (36% FG in December is not of the good). Combine that with injuries and limited minutes for Faried and we get a slide in both offense and overall rating. Denver is a team to watch, as they have some interesting pieces and contracts to trade (Faried, Chandler, McGee, and even Gallo). I expect them to make a move simply because I don't think their coach likes how this team is put together.
  • #14 Toronto Raptors Neutral Site Win:49.4% (Up 2,#18 in offense, #11 in defense): Toronto is 5-3 since the Rudy Gay trade, with two losses to San Antonio and an overtime loss to Charlotte. Both stats and reality seem to be against poor Rudy Gay (0.000 WP48 in Toronto). Wasn't that trade supposed to mean Masai was tanking? Isn't it hard to replace inefficient high volume shooters? I love the media -- really I do -- but the Gay trade was all about making the Raptors better right now, not worse. Toronto is the best team in the Atlantic at this point. It remains to be seen if that's what they want. They have a move to make and that move will give us a bit more clarity about their intentions. A Lowry trade probably means packing it in. A DeRozan trade might mean making a run. I do expect Masai to win any trade he makes though. NBA GMs: stop taking his calls!
  • #15 Detroit Pistons Neutral Site Win:49.4% (Up 6,#14 in offense, #13 in defense): The Pistons are a five hundred team on the strength of their franchise center Drummond (0.323 WP48 at 20 is just sick). There isn't anything else there right now. The Josh Smith/Greg Monroe experiment has not worked. I'd advise this team to trade for assets, but Dumars does not have a stellar track record on that front.
  • #16 New Orleans Pelicans Neutral Site Win:48.7% (Down 2,#11 in offense, #23 in defense): In 2066 minutes last year, Al-Farouq Aminu had a 0.217 WP48, and he's at 0.268 WP48 this year. He looks like a legit star. Throw in Davis looking like a future top five player and the rest of the pieces around them, and you have to like this Pelicans team going forward. Even though their defense is very, very poor. If I was running the team, I would take a good, long look at an Asik trade.
  • #17 Charlotte Bobcats Neutral Site Win:46.1% (Up 3,#28 in offense, #5 in defense): I was asked to summarize the Bobcats recently for a friend. My answer was the Bobcats are a well-disciplined, well coached, strong defensive team that's fun to watch, but couldn't sink a jump shot in the ocean. Did you know Josh McRoberts leads the team in assists? Not a good sign. If they had any offense at all they'd be pulling the East's version of Portland this year. Even shooting threes won't help them, as they shoot the worst percentage in the league. They're definitely a team in need of a trade.
  • #18 Washington Wizards Neutral Site Win:45.8% (No Change,#20 in offense, #19 in defense): Three straight road wins for the Wizards sounds good right? Did I mention it's against the Atlantic? The Wizards are talented, but their coaching remains atrocious. They should be able to coast to 45 wins, but they'll struggle to get there.
  • #19 Boston Celtics Neutral Site Win:44.7% (Down 4,#23 in offense, #10 in defense): I was convinced that I wouldn't care about the Celtics this year, yet here I am, completely invested. The Celtics will be making some trades soon. My hope is that they keep Rondo (who is one of the few guys in the league who's traded haymakers with LeBron in the playoffs and won) and who I think still has his best basketball in front of him. I also think they should do what they need to do to get a starting center. If only one of those were available....
  • #20 Los Angeles Lakers Neutral Site Win:43.8% (Down 3,#22 in offense, #15 in defense): We've learned that, with a hobbled Kobe, the Lakers are a lottery team, and without him they're a fun, scrappy, borderline playoff team. Here's some detail on Kobe's injury: the words Yao Ming should scare the hell out of any Lakers fans. I would shut down Kobe for the year and explore all my trade options while trying to maximize my lottery options. 
  • #21 Brooklyn Nets Neutral Site Win:43.2% (Up 2,#16 in offense, #29 in defense): Before the season started, my model was worried about injuries completely decimating the older players on this team. Now, on top of that ongoing concern, their best player and perhaps lowest injury risk (Lopez) is out for the year. This team feels done to me. Completely and utterly done. I really don't know how you can salvage this season other than by completely blowing it up.
  • #22 Chicago Bulls Neutral Site Win:42% (Down 3,#29 in offense, #2 in defense): At this point, only Noah, Butler, and Rose should own any property in Chicago. Everyone else on this team should be renting on short term leases with no penalties for cancellation. I expect everyone to be on the block.
  • #23 New York Knicks Neutral Site Win:40.8% (Up 2,#17 in offense, #25 in defense): Lost amongst the disaster of a season the Knicks are having is that Carmelo Anthony is having the best year of his career (0.167 WP48). This cann't offset that fact that this team has played Andrea Barganani -- our all-time wins produced most harmful player -- the second most minutes on the roster. At times the Knicks roll out lineups that couldn't stop me from scoring in the post.
  • #24 Orlando Magic Neutral Site Win:38.5% (Down 2,#24 in offense, #21 in defense): If you're tanking and you know it clap your hands. If you're tanking and you know it, and you're not afraid to show it; if you're tanking and you know it clap your hands.
  • #25 Sacramento Kings Neutral Site Win:37.1% (Up 4,#13 in offense, #28 in defense): Since the Kings traded for Rudy Gay they are 2-5, while his previous team is 5-3. The lesson continues to be to never trade for Rudy Gay. Boogie is finally showing signs of life (0.133 WP48) and it might go for naught.
  • #26 Memphis Grizzlies Neutral Site Win:35.8% (Up 2,#19 in offense, #26 in defense): The Grizzlies continue to be a casualty ward but that has provided opportunities for some players. Recently signed James Johnson from the D-league is posting a 0.325 WP48 through 4 games. This bears watching.
  • #27 Utah Jazz Neutral Site Win:33.7% (Down 1,#25 in offense, #27 in defense): Utah has been a better team recently since Trey Burke started playing more minutes. They've also played the toughest schedule in the league. Given that and the youth of this squad, I suspect that Utah will continue to move up slowly through the bottom of the ranks. The over for this team (27.5 wins) is still very much in play.
  • #28 Milwaukee Bucks Neutral Site Win:33.5% (Up 2,#30 in offense, #17 in defense): As of right now Giannis (.197 WP48) is your Wins Produced rookie of the year. Milwaukkee: keep playing him and Henson all the minutes.
  • #29 Cleveland Cavaliers Neutral Site Win:30.8% (Down 5,#27 in offense, #18 in defense): 1-4 over their last five with their one win at home in overtime over Milwaukee. The Cavs suck. They really, really suck. Honestly? Varajao and Thompson have played well (0.201 and 0.130 WP48 respectively), but everyone else should be considered expendable or tradable. Everyone else.
  • #30 Philadelphia 76ers Neutral Site Win:28.8% (Down 3,#26 in offense, #30 in defense): This team is better across the board when Michael Carter-Williams plays. They're also totally making a trade soon.

How does the model think all this plays out over 82 games?

Season Simulation as of 12/23/13
NBA Rank Playoff Seed Team Win Proj Vegas
Playoff Odds Top 4 Seed
1 Seed
Avg Seed
1 E1 Indiana Pacers 66 53.5 100% 100% 97% 1.03
5 E2 Miami Heat 57 60.0 100% 100% 3% 2.01
8 E3 Atlanta Hawks 48 40.0 100% 96%   3.16
15 E4 Detroit Pistons 40 41.0 93% 37%   5.50
14 E5 Toronto Raptors 39 36.5 88% 21%   6.17
18 E6 Washington Wizards 39 42.0 86% 20%   6.27
17 E7 Charlotte Bobcats 38 27.5 84% 16%   6.59
19 E8 Boston Celtics 37 27.5 71% 8%   7.43
21 E9 Brooklyn Nets 33 52.5 25% 1%   9.76
22 E10 Chicago Bulls 33 56.5 26% 1%   9.80
23 E11 New York Knicks 32 49.5 20% 0%   10.13
24 E12 Orlando Magic 29 24.5 5%     11.81
29 E13 Cleveland Cavaliers 28 40.5 3%     12.36
30 E14 Philadelphia 76ers 25 16.5 0%     13.72
28 E15 Milwaukee Bucks 23 28.5       14.27
3 W1 Oklahoma City Thunder 61 50.5 100% 100% 57% 1.65
4 W2 San Antonio Spurs 59 55.5 100% 97% 21% 2.49
2 W3 Los Angeles Clippers 58 57.0 100% 97% 17% 2.64
7 W4 Portland Trail Blazers 55 38.5 100% 86% 5% 3.53
6 W5 Phoenix Suns 50 21.5 98% 16%   5.48
10 W6 Houston Rockets 47 54.5 84% 3%   6.82
9 W7 Minnesota Timberwolves 45 41.0 66% 1%   7.79
12 W8 Dallas Mavericks 45 44.0 63% 1%   7.88
11 W9 Golden State Warriors 44 49.5 62% 1%   7.94
13 W10 Denver Nuggets 41 47.0 18%     9.81
16 W11 New Orleans Pelicans 38 40.0 7%     10.79
20 W12 Los Angeles Lakers 36 33.5 2%     11.49
26 W13 Memphis Grizzlies 31 49.0       13.18
25 W14 Sacramento Kings 27 31.5       14.07
27 W15 Utah Jazz 26 27.5       14.45

A small note here: I'm working on updating the playoff model over the holidays. For now it remains unchanged.

Merry Christmas to all, and to all a good game night.

Minor flaw in your season simulation: division winners are guaranteed a top-4 seed. Someone from the Atlantic's gotta win it, so your 'Top 4 seed' odds are off.
Wow. The Hawks are sitting pretty long-term with their current draft-and-stashes and the Brooklyn picks.


Stand by both of those. Rockets will be better. Lakers are all about Jordan Hill.
Is there any realistic trade that could create a third team with a shot at winning the East (no matter how slim)?

So basically, you just cannot be wrong.

The projection in this article gives the Rockets the 9th best record but "they WILL be better". When ? Next year ?

And you're telling me that it's because of a guy who plays 20mpg that your Laker's pre-season projection is 11 (!!) games off ?

Oh, and your Pacers pre-season review is just hilarious.
That's a little harsh!

And your two quoted tweets are positives for Arturo... and I quote:

> Now reading Bill Simmons thinking the Lakers made themselves better by signing Kaman, Young and Farmar

Of those three, only Farmar is adding much. The other one laughs at the projection Houston will be eighth. We'll see where that falls. That hardly equates to "So basically, you just cannot be wrong. " By all means, make that argument with detail, but otherwise it just reads like an angry rant.

Sidenote: I never get why internet people are so angry and aggressive. I really don't. It is always so personal - would you talk to someone in real life that way? /Sidenote.

There are massive flaws in any projection system, and that is why we watch sport - what makes it so compelling - and finding the flaws and fixing them is what is required.

I program, and the goal is to iterate around a solid base - not start with perfection but strive towards it. A system that is imperfect but IMPROVING is better than no system.
I've got a suggestion: can you put the projections from here: http://www.boxscoregeeks.com/articles/the-pacers-the-overreaction onto the team page here: http://www.boxscoregeeks.com/teams/ind with over/unders? By that I mean projected minutes vs actual, and WP48 actual versus predicted.

That'd be a great way to see how accurate the projections were/are, and more importantly why.
"Since the Kings traded for Rudy Gay they are 2-5"

....which is the same winning percentage as before they got him

I'm not "angry" because they may be wrong, we all are very often. What makes me "angry" is that they NEVER say that they were wrong.

Before they made their pre season projections, they wrote an article saying that everyone is bad at it. But they never say that they are way worse than others.

And when you don't agree with them, it can only be because you're "smoking things". Because they cannot be wrong !

And just read their Pacers pre season projection article. Not only their analysis could not be more off, there is so much arrogancy in it!
I thought this was a fairly honest breakdown of what they got right and what they got wrong:
@ DG22, They did say that Rudy Gay was at ~0 WP. Which kinda makes them being exactly the same make a lot of sense. Not like they lost much in Salmons.

@Nick, It's their blog and they are giving us all this free insight and analysis. Let's just be productive and civil. The personal criticism aside I think a postmortem on the biggest misses would be great to analyze and read. That'd be an awesome article but wouldn't it make sense to wait till everything pans out before we point fingers. All I've read so far essentially is let's wait and see. Things don't look good for some of our projections but they're not implausible either. Most of the time, there are good reasons the projections are off, injuries, rotations, trades etc. But I think everyone knows they missed hard on Indy. I'm sure they're aware of it and they've admitted it.
@allintheeyes & GnoiWiaK

First phrase of their opinion of the Pacers : "I've watched the last several Pacers games and I have to say that in addition to their easy schedule, the refereeing has really been helpful for this team"

Their projection is off by 26 (!!!) games and who do they (primarly) blame ? Their schedule (they projected the Sixers would win 35 games because of the East tankfestival - how could Indiana only win only 40 !?) and the refs !

I mean come on.

Please use the word never with the greatest caution. We have weekly podcasts talking places we "missed" and disagree. Even our original predictions had the classic economist "on the one hand." And all of us here are huge fans of unexpected things happening in the NBA (Zach Randolph is one of my favs.)

In short, I think we do a good job tempering our predictions with analysis, and it's also worth noting that between the multiple authors, we don't agree on everything. We give lots of good content, we hope you like it. If you want to keep coming back and yelling at us though, well I suggest you find a different comment section for your anger.
@Nick Basketa, while the wording is bad, they're not wrong. The Pacers do get a lot of favorable calls. Hibbert on defense and George on offense. There's a good b-ball breakdown video were a ex-ref points out a good Hibbert no call and a reputation no call. This Verticality thing really just lets Hibbert foul people half the time. The Refs are also treating PG like a young Kobe Bryant (Which he pretty much is becoming): http://www.basketball-reference.com/play-index/pcm_finder.cgi?request=1&sum=0&p1=georgpa01&y1=2014&p2=bryanko01&y2=2000#per_game::none Their 4th year stats are almost identical, especially the free throws
Nick its just weird that of all the things this site got "wrong", your "aha" post consisted of two tweets that they got pretty much right. Nick Young and Kaman HAVE NOT helped LAL, and Farmar has slightly. That is 2/3 correct, and just one wrong, net score +1. The other was laughing at a prediction of an 8th placed Houston, who BTW are in 5th and IMPROVING.

If you want to construct an argument that this site "can't admit they are wrong", by all means do, because there is a LOT wrong with the preseason predictions (Utah, Phoenix, Pacers, Houston) but those two tweets are pretty much spot on.

Honestly, why is it so PERSONAL to you? It is like they came into your house and shaved your prized Poodle. This is a stats blog on basketball. It honestly can't get much more trivial (no offence to anyone intended - this topic is trivial in the truest sense - i.e. "of little importance").
Ok, let me explain why i'm hating so much.

Doing season projections is something extremely difficult. I remember a post you had before (or while i don't remember) posting all the team's reviews saying (and very rightly so) how difficult it is (and how wrong everybody was last year) because of LOTS of unpredectible factors (minutes, injuries, etc).

But your predictions are very different from most of those made by NBA "experts". While a grand majority of them uses their opinion to predict the wins each team will get, you use a stat you calculate yourself for each player, the Wins Produced.

So the logic would say, if a NBA "expert" is very wrong (10+W wrong) on one team that didn't suffer any crazy injury or other crazy PG explosion, it's because his opinion was wrong. And with the same logic, if you, BoxscoreGeeks are wrong, logically, it would be your formula's fault.

But you always blame the same factors the NBA "experts" blame. You never consider that maybe, your WP stat is actually bad at predicting how many wins a team will get the next year.

And on top of that, when one of your projection is way different that what the consensus says, you start bragging about how stupid is everyone while the year before the WoW projections were one of the worse across NBA "experts" (http://weaksideawareness.wordpress.com/2013/04/23/checking-2012-13-nba-win-predictions-projections/).

I study statistics and everyday i try to convince people that numbers are not wrong, it's their interpretation that is. But how do you want the advanced metrics movement to progress when you act so arrogantly. It's never "they may" or "i think", it's "they will". It's never "we'll be more wrong than others because our formula isn't perfect", it's "we'll be wrong because we can't predict injuries".

PS. @motherwell

Simmons ranked Houston 8th leaguewide, not in the West. And their simulation in this post ranks them...9th (who's smoking what now?)
And if you're using their tools to say that Young and Kaman are bad, of course they are bad. But if you look at other tools, like win shares (i'm not an advocate of WS, i just prefer to look at many indicators, not just one), Young is 2nd in WS for the Lakers... Young+Kaman+Farmar=2.1wins. Which is positive. Which means they have helped :)

PS2 @GnoiXiaK

I'm not saying they are wrong. But why don't they start by saying "OK, we screwed up on this one. PG & Hib are awesome. [...] But i also think they had an easy schedule and are getting a lot of calls" ?
The wording is EVERYTHING and here it's more than just bad, it dishonnest.

PS3. Your model projected the Jazz would get more wins than the Pacers. And you haven't written something about how the model may (just may) be wrong...?
Looks like we have an infant getting irritated at xhildish natters ;)
completely agree Nickbasketa - I guess you just have to deal with people calling you an infant and saying you must be smoking something.

I know someone who works at the Suns as part of their growing analytics team; there is no way they look at WP. Funny enough, the Suns were labeled as 'The Tank' in the preview. I guess they must be smoking some stuff that Michael Beasley left behind...

To be constructive with my criticism, I think it's time you try to improve WP48 rather than defending it like crazy, it's riddled with statistical follies. This is a very well done website in terms of UI, but I think you guys are holding yourself back by putting all your eggs in the WP basket.

(prepares for insults, shit to be thrown, and be told how stupid I am)

@Nick, as I've pointed out multiple times here AND on wagesofwins.com, there are perfectly valid reasons why the model predicted the Pacers wrong (seeing as I'm the resident Pacers fan).

For the first 3 years of his career, Roy Hibbert was terrible. Awkward on offense, slow on defense. In his 4th year everything started to click, and he got a max deal. But in his 5th year he was playing with an injured wrist which destroyed his FG%. Only after the wrist healed (after the all-star break) did his performance start to improve, but that does not eliminate Oct-Feb. Now in his 6th year, he's playing like the guy we saw in the latter part of last year. But after 5 years of poor performance, the model would have to think that 3 months of good basketball was an outlier and he would return to (poor) form. That didn't happen and Roy went from a negative contributor to subpar to all-star.

The EXACT same story, minus the wrist injury, can be said of Lance Stephenson. Negative contributor to subpar to all-star this year. Now, those are two of your top 3 in minutes played, when the model was perfectly justified in thinking they would be bad players. Is it any wonder the projections were so far off?

The point in looking at advanced metrics is not to say "I was right, you were wrong!" but rather to understand WHY things happen the way they do. You learn a lot more from being wrong than you do from being right.

Glad someone else has picked up on what has been my same frustration with this website on their analysis of the Pacers. The statistical analysis part of the discussion and consistent logic applied to other teams (notably in the season preview) was seemingly thrown out the window for Indy for some unknown reason. The whole refs thing is even more annoying. Does anyone really think NBA referees don't hear about the Pacers from opposing coaches before every game? I'm sure someone from the opposing coaching staff is in the refs ear about Roy's supposedly illegal application of verticality before, and during, each game. At this point he is probably a focal point of each game, and would be more scrutinized than anyone else on the floor on defense. He won't be able to get away with calls unless ALL of the refs are just consistently bad, and they ALL need to be replaced. It is human nature that at least one call will be missed every now and then. Are the people criticizing the officiating calling for instant replay on each foul? Or going the NFL route and giving the head coach three challenges they can use on questionable non-calls? I don't want to see that.

Just frustrating as a fan of the Pacers to read about them on this website. I know, I know, just don't visit the website right? Well I'm sure that's not what the Box Score Geeks are trying to accomplish, because I do enjoy reading a different take on the NBA.
@Xavier, I'm also a resident Pacers fan and you know what? I said exactly the same things you did in my response to the Pacers predictions page. 1st post :) The model is too rigged and I think it weighs youth/rookie years too heavily. Rapid growth through offseasons are not unusual, especially for bigs. Remember how long Bogut was a "Bust" until he wasn't? Same for Hibbert. Also, perception matters. Star-treatment is real. After PG Exploded in the playoffs and became a household name, I knew the league was going to give him point. It's a star driven league and they love small market Super Stars. This is to take nothing away from him, he's got all the talent but now he gets the Kobe calls.

@Nick, I agree with your comment, I don't think people should be insulting you in anyway. Especially when it was more refined and civil.

@Everyone, let's be productive and, yes, nick's criticisms are productive.
Of course, I can't totally let BSG off the hook. This week OKC gets a qualifier in their paragraph saying they lost on the road twice on the 2nd of back-to-backs against good teams, aka scheduled losses. Why didn't the Pacers get that same qualifier when they lost to Portland and OKC on their 5 game road trip, with both losses coming on the 2nd of back-to-backs?

And they still get backhanded compliments on their wins, with Brook Lopez being out and their "remarkable health"? George Hill is playing through injury, and the Pacers are using CJ Watson and Lance Stephenson to limit his minutes. The Pacers also have a lot of faith in Luis Scola to suck up minutes from David West, to minimize any chance of injury. Paul George and Lance Stephenson are both 23, so you wouldn't predict major injuries anytime soon. I wouldn't say they have remarkable health, I'd say they're well constructed.
I will say this to all the Pacers fans: when we're hostile to BSG for their predictions and/or analysis, and we HAVE been hostile, it only sets them up to be confrontational in response. Don't like what they say? Form a coherent argument that doesn't call them haters, and don't imply that they are bad at what they do. The whole reason we come here is because they're damn good at it.
Agree with xavier. I am guilty as charged for the "hating" comments. I read the backhanded compliments and ref bias comments and lose my mind a little for some dumb reason. I apologize to the BSGs, keep up the good work.

I completely agree and totally understand how a model can be awfully wrong in some cases. The objective of all models is to be "right" in the big picture which means there will always be some outlayers.

But in the Pacers case, they projected that they would be win 10 LESS games than last year. What on earth did the Pacers do this summer that would make them lose 10W the next year !?

I mean i don't know much, but they have a great coach, average age last year : 25.7 (18th older) and didn't do anything major in the summer. You can't say their 2012 regular season was a fluke since they confirmed in the playoffs (YES taking the best team in the league to 6+ games two years in a raw MEANS SOMETHING). I mean, why would they win 10 less games the next year !?


Thx. Let's get insulted and have our messages erased (or changed) because we aren't bowing to their greatness
Lost in all of this bullshit is that we say we are wrong every fucking day on this blog.

If all we were interested in was being perceived as geniuses, we'd probably just remove old posts where we said things that turned out to be horribly wrong, and we'd also probably not post weekly updates to call attention to how wrong we were.

Does anyone remember me calling Brook Lopez the worst contract in all of basketball? And now we're calling him Brooklyn's best player this year? That's because we are interested in the truth, not in the size of our internet penises.


If you piss on my lawn, I will clean it up. This isn't your house and you don't get to come on here and say whatever the hell you want while acting like Dudley Dursley. I've let some of your posts remain and have deleted others. You might want to take a look back at what the common variable was in those posts that I have deleted.

If you want the freedom to post like an ass and never get censured or deleted, start your own blog.
Thank you Patrick for calling our conversation "bullshit". Classy as always.

My critisism is towards the WP (and its use for projections) and the way you write about it ("you disagree with us ? then you must be stupid/smoking smth"). Of course you answer on you being wrong about a contract. Spot on !

PS. Beside insults (which i've never used) i don't see why you would delete/change a message. And how do you want me to take a look at the messages you've deleted ?
> The model is too rigged

A model can't be "rigged" - because word means:

> situations where unfair advantages are given to one side of a conflict.

A model is designed to try to predict - it has no power to affect anything - it is supposed to MATCH reality, not INFLUENCE reality.

That misuse of that one word "rigged" seems to sum up nicely this whole thread. People seem to think the model here INFLUENCES what happens, and they take it personally because it has some Voodoo power to affect change. That view is so tinfoilhat-ish it is laughable, really.

It is a first attempt at prediction people - the prediction aren't going to change ANYTHING!

You seem to be confusing the WP model with the session prediction model. The WP numbers do, in fact, show the Pacers being better than expected. It shows where and how which players improved.

The part they got wrong about the season predictions were their guesses based on age, past performance, health, expected rotation, and projected minutes.

And, transparent as they are, they showed you what they were guessing on those projections.

If you are going to criticise so vehemently and voluminously, at least try to understandwhat you are talking about first, as a courtesy to the rest of us. The WP model actually showed where they guessed wrong, which doesn't fit your criticism at all.
@Nick, Like I said in my previous post, the model reasonably expected both Roy Hibbert and Lance Stephenson to be subpar players. They are both substantially above that mark. There was also expected regression from David West (reasonable) and George Hill (that was a ?). That's easily 10 wins over a season, considering how many minutes those guys play.
@Nick, disregard my last post, I misread what you wrote.

I think you misinterpreted what he is saying. Based on context, pretty sure Nick meant 'rigid' not 'rigged'


I think the criticism is driven by how confident they are in their predictions. It's not a prediction model, but they talk about it as if it's a fantastic prediction model.

Quickly scanning team previews reveals a tremendous amount of confidence. I think this is Nick's primary complains (please correct me if I'm wrong):

Phoenix Suns (keep in mind this is a team doing sophisticated quantitative analysis)

"Suns taking the season off from competitive basketball"
"It doesn't matter who's on this team now. They're not even trying to win."

Blazers - " It's possible they'll return to being a contender. It won't be any time in the near future though"

Jazz - "This team will win at least 41 games"

Ultimately, it would be great to see a true analysis of the model (predicted vs. actuals) at a team level, with out of sample testing. I understand that this would be time intensive and require assumptions (i.e. how would I have predicted minutes for that year). However understanding the variability of a model (and why) is tremendously important to making predictions, which the writers and readers both seem to enjoy.

Your definition of "insult" differs from mine. The posts of yours that I have deleted invariably involved you saying things in a manner that would get you punched in a face-to-face conversation.

So maybe you about 6'7" and 250lbs and just used to people letting it slide when you behave like that, or maybe you act tougher on the internet. Either way, those are the posts I've deleted.


We were very confident in many of our projections. We also admit how we're wrong a lot, and make fun of ourselves.

Many of you, however, are not happy when we admit that we're wrong, and try to elaborate why. Instead, I get the feeling that nothing less than "We were wrong because we're total morons and don't know wtf we are doing" will satisfy you. To those of you, I'd advise you to inhale very deeply and just hold your breath. I'm sure we'll get to that soon.
@Tyler Anyone can check this for themselves, as this is the WHOLE POINT of this site, by simply comparing the predicted WP48 to the (current) actual.

TL;DR I think its clear from the evidence why most of the predictions were wrong: injuries and random improvement / regression.

My team the Suns have seen The Morri, Green and Plumlee come from nowhere:

Marcus alone went from net negative to borderline star, and Plumlee? WTF? Who saw that coming? http://wagesofwins.com/2013/12/12/the-suddenly-amazing-morris-twins/ is a good explanation of this.

The Pacers have seen Lance and Paul George EXPLODE:
http://www.boxscoregeeks.com/articles/the-pacers-the-overreaction vs http://www.boxscoregeeks.com/teams/ind?direction=desc&sort=per48_wins_produced which maybe someone could have been predicted, but how?

Blazers Mathews and Batum have been insane
http://www.boxscoregeeks.com/articles/the-trail-blazers-be-careful-what-you-wish-for vs www.boxscoregeeks.com/teams/por?direction=desc&sort=per48_wins_produced and Matthews especially is a MASSIVE shock at 27 years of age.

Brooklyn, another huge miss, has seen injuries yes, but it is more age striking cruelly and suddenly. Utah was AWFUL but now Trey Burkes is back and John Lucas III will no longer get 29 minutes a night (as he did 6 of the first 8 games), we should see that team improve as well, especially as the lack of a PG seemed to hurt a to of their guys (lots of underachievers).

Now, back to your over arching question, which is "... understanding the variability of a model (and why)", if you can model either of those two factors, injuries and sudden improvement and/or regression, I'd love to hear how.

Any model is going to make most of its mistakes at the edge, where things happen that are not "regularly scheduled" or expected. Improvement (Morri, Paul Georg and Stephenson), regression (Garnett and Pierce) and injuries (Brooklyn and Utah) are hard to predict, but maybe a confidence rating, based on age and say past minute variability, would be a nice addition to the predictions.
I dont know why people are so bothered by the confidence level found on this site. In Fact the team previous had a pessimist/realist/and optimist settings. People dont seem to bat an eye when the talking heads state for a fact that one team is great because of this that etc.

I for one am loving the new site and want more articles. The team reviews were great and seem to be going pretty much in line with what i expected. I thought Utah would do better but they seem to be coming around.

I think another variable that cant be quantified, one that is especially prevalent this year - that of teams straight up tanking. Hard to predict when some teams will just throw in the towel.

"I know someone who works at the Suns as part of their growing analytics team; there is no way they look at WP."

Well Ill be damned, its stuff like this that gets me riled up. So because someone people in "power" for lack of better word arent doing something then it is no good. This reeks of authoritarianism (they know better then me or you! Why bother start new businesses or come up with new products i mean they should already be in use. The basic fact of the matter is tons of teams are ran terribly or just by feeling. Just because they dont look at a particular model does not inherently make that model wrong.
I give up - I am trying to provide productive criticism to better the model, which I think can be improved. In fact, I think this site could be fantastic with the current UI and begin work on a new model. You clearly have a lot of knowledgeable readers that know a lot about basketball, but more importantly, statistics & modeling. This feedback could be very helpful, and I am guessing some readers would be willing to help build a new model for free. Instead I just get told these ridiculous reasons for why I am an idiot. (i.e. I "reek of authoritarianism" for citing an example of a team that uses sabermatricians that doesn't value WP48...).

Patrick, I can't believe how you talk to readers of your website like they are all children. "If you want the freedom to post like an ass and never get censured or deleted, start your own blog." If you tried to build on the criticism, you could have your own model and supporting website good enough to attract attention from major media companies/sports teams. I, and Nick, are both trying to start a good dialogue. Instead, we just get insulted.

I won't bother posting again.

Good luck on BSG - I think it's a great start (honestly).
If you're going to blast predictions without looking at where they went wrong you aren't advancing the conversation.

In the case of the Pacers, the biggest difference between last year and this year is the strength of their defense. More than half of the difference in performance this year is from the Pacers going from an excellent defensive team to a pretty ridiculous outlier (and if it keeps up it might be worthy of a post on best defenses ever). This is not something that *any* model got right - if you know of anyone out there saying the Pacers were going to go to the next level this year because their defense was going to go from great to otherworldly, do share it.

So I would not ding anyone here for that.

I would ding BSG for being pessimistic on Paul George and Lance Stephenson - their preview had PG with only a modest improvement, and LS regressing, despite both being only 23 years old. As we've seen again and again, players younger than 25 frequently improve substantially from year to year - and the BSG model was caught completely off-guard by this.

This is the same story we see in Phoenix, and to a lesser extent Portland - young, good players making the leap from good to great, bringing their teams with them.

Now to some extent this is unpredictable - we know that in any given year *some* young players are going to make the leap, but we do not know which ones (and if we did, there are millions of dollars to be made in NBA front offices!) The model should take this into account with higher 'best case' predictions for teams with several players primed to make the jump.

If I had to criticize the predictions, it would be that they only seem to take into account 'known unknowns' (the best and worst case scenarios look suspiciously like the 95% confidence interval of a binomial centered on each team's expected win percentage) and not enough unknown unknowns (like rookies stepping it up, backbreaking injuries, old players aging like milk, etc).
Waitung for the next article so that se can move on 😁
Wow... don't sign in for a few days and this is what I find! Anyway, enjoyed the article. For those of you whom are Pacers fans, yes, they are pushing the very limits of what refs will or won't call and it is why their defense is so great. All great defenses in NBA history have basically done that. I think the refs will eventually deal with it but I guess we'll see (the Bulls in the 90's got away with zone defense for years which is really too bad since it wasn't allowed at the time).
In basketball and in life, the boos always come from the cheap seats.

This blog is the best basketball analysis out there.

Sign in to write a comment.