Which Position Should You Build Around For Playoff Success?
Posted by Neil Paine on May 13, 2010
Here at Basketball-Reference, we recently gathered a new database of player positions that goes further than merely listing guards, forwards, and centers. It isn't live on the player pages yet (that's coming soon), but I've worked with it in several posts over the past few weeks, and today I'm going to use it to see which position has traditionally been the strongest for teams that are successful in the postseason.
Here's the study I ran: First, I calculated Win Shares and John Hollinger's Game Score per game (which I used rather than PER because I can easily make it a per-game metric) for both the regular-season and the playoffs. Then I sorted each team by all four categories, and noted the position of the team's best player in each (players had to play more than half of the team's scheduled games to qualify for the team lead in GmSc/G). Finally, I tallied up the total playoff wins, losses, and championships for teams whose best player was at each position -- so for instance, teams whose best regular-season player by WS was a PG have won 7 championships since 1952, etc.
Here are the results for 1952-2009:
| 1952-2009 | RS-BestbyWS | RS-BestbyGmSc | PO-BestbyWS | PO-BestbyGmSc | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Position | # | W | L | Champs | # | W | L | Champs | # | W | L | Champs | # | W | L | Champs |
| PG | 110 | 533 | 521 | 7 | 118 | 585 | 543 | 12 | 127 | 662 | 600 | 9 | 129 | 587 | 572 | 9 |
| SG | 120 | 541 | 545 | 8 | 113 | 522 | 524 | 8 | 129 | 584 | 550 | 15 | 128 | 549 | 565 | 9 |
| SF | 121 | 550 | 574 | 5 | 135 | 573 | 614 | 6 | 120 | 507 | 548 | 9 | 122 | 547 | 562 | 9 |
| PF | 153 | 641 | 682 | 11 | 161 | 651 | 722 | 8 | 144 | 580 | 639 | 7 | 137 | 586 | 629 | 7 |
| C | 174 | 804 | 747 | 27 | 151 | 738 | 666 | 24 | 158 | 736 | 732 | 18 | 162 | 800 | 741 | 24 |
As you can see, teams whose best players were centers have dominated the championship count by any metric, although teams whose best playoff Win Shares-earner was a SG also stand out from a championship point of view. Also, it should be noted that centers dominate the raw totals because, in NBA history, more playoff teams have had centers as their focal point than any other position, and by a pretty wide margin. So here's the same table, but with percentages instead of totals:
| 1952-2009 | RS-BestbyWS | RS-BestbyGmSc | PO-BestbyWS | PO-BestbyGmSc | ||||
|---|---|---|---|---|---|---|---|---|
| Position | Wpct | Champ% | Wpct | Champ% | Wpct | Champ% | Wpct | Champ% |
| PG | 0.506 | 0.064 | 0.519 | 0.102 | 0.525 | 0.071 | 0.506 | 0.070 |
| SG | 0.498 | 0.067 | 0.499 | 0.071 | 0.515 | 0.116 | 0.493 | 0.070 |
| SF | 0.489 | 0.041 | 0.483 | 0.044 | 0.481 | 0.075 | 0.493 | 0.074 |
| PF | 0.485 | 0.072 | 0.474 | 0.050 | 0.476 | 0.049 | 0.482 | 0.051 |
| C | 0.518 | 0.155 | 0.526 | 0.159 | 0.501 | 0.114 | 0.519 | 0.148 |
Now, the picture becomes clearer: in order to be successful in the playoffs, you traditionally need to be built around either a dominant center or a great point guard. That said, the game has not stayed static over the past 57 years -- rules have changed, styles of play have come into and out of vogue, the playoffs have become longer and more drawn out, etc. So I thought I'd also break down these charts using just data since the playoffs expanded to 16 teams in 1984:
| 1984-2009 | RS-BestbyWS | RS-BestbyGmSc | PO-BestbyWS | PO-BestbyGmSc | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Position | # | W | L | Champs | # | W | L | Champs | # | W | L | Champs | # | W | L | Champs |
| PG | 77 | 372 | 366 | 4 | 80 | 402 | 374 | 6 | 87 | 480 | 428 | 5 | 85 | 395 | 387 | 5 |
| SG | 88 | 448 | 423 | 7 | 82 | 422 | 392 | 8 | 92 | 462 | 408 | 12 | 95 | 473 | 445 | 9 |
| SF | 70 | 303 | 325 | 2 | 80 | 324 | 363 | 2 | 62 | 228 | 272 | 2 | 68 | 270 | 301 | 2 |
| PF | 96 | 445 | 455 | 6 | 100 | 454 | 480 | 5 | 97 | 408 | 453 | 4 | 95 | 446 | 464 | 5 |
| C | 85 | 386 | 385 | 7 | 74 | 352 | 345 | 5 | 78 | 376 | 393 | 3 | 73 | 370 | 357 | 5 |
And again, the percentages:
| 1984-2009 | RS-BestbyWS | RS-BestbyGmSc | PO-BestbyWS | PO-BestbyGmSc | ||||
|---|---|---|---|---|---|---|---|---|
| Position | Wpct | Champ% | Wpct | Champ% | Wpct | Champ% | Wpct | Champ% |
| PG | 0.504 | 0.052 | 0.518 | 0.075 | 0.529 | 0.057 | 0.505 | 0.059 |
| SG | 0.514 | 0.080 | 0.518 | 0.098 | 0.531 | 0.130 | 0.515 | 0.095 |
| SF | 0.482 | 0.029 | 0.472 | 0.025 | 0.456 | 0.032 | 0.473 | 0.029 |
| PF | 0.494 | 0.063 | 0.486 | 0.050 | 0.474 | 0.041 | 0.490 | 0.053 |
| C | 0.501 | 0.082 | 0.505 | 0.068 | 0.489 | 0.038 | 0.509 | 0.068 |
Now we start to see the Jordan Effect come into play, as SGs emerge as the dominant position of the modern era. After MJ retired, though, PFs began to assert their dominance for championship teams:
| Year | Team | W | L | WPct | RS-BestbyWS | RS-BestbyGmSc | PO-BestbyWS | PO-BestbyGmSc |
|---|---|---|---|---|---|---|---|---|
| 1999 | SAS | 15 | 2 | 0.882 | PF | PF | PF | PF |
| 2000 | LAL | 15 | 8 | 0.652 | C | C | C | C |
| 2001 | LAL | 15 | 1 | 0.938 | C | C | SG | C |
| 2002 | LAL | 15 | 4 | 0.789 | C | C | C | C |
| 2003 | SAS | 16 | 8 | 0.667 | PF | PF | PF | PF |
| 2004 | DET | 16 | 7 | 0.696 | PG | PG | PG | SG |
| 2005 | SAS | 16 | 7 | 0.696 | PF | PF | SG | PF |
| 2006 | MIA | 16 | 7 | 0.696 | SG | SG | SG | SG |
| 2007 | SAS | 16 | 4 | 0.800 | PF | PF | PF | PF |
| 2008 | BOS | 16 | 10 | 0.615 | PF | PF | PF | PF |
| 2009 | LAL | 16 | 7 | 0.696 | PF | SG | SG | SG |
The only constant? Teams built around small forwards don't tend to be as successful as other positions. Here are the only NBA champs whose best player was (or might have been) a SF:
| Year | Team | W | L | WPct | RS-BestbyWS | RS-BestbyGmSc | PO-BestbyWS | PO-BestbyGmSc |
|---|---|---|---|---|---|---|---|---|
| 1956 | PHW | 7 | 3 | 0.700 | C | C | SF | SF |
| 1958 | STL | 8 | 3 | 0.727 | PF | PF | SF | SF |
| 1968 | BOS | 12 | 7 | 0.632 | PF | C | SF | SF |
| 1969 | BOS | 12 | 6 | 0.667 | PF | SF | SF | SF |
| 1974 | BOS | 12 | 6 | 0.667 | SF | SF | SF | SF |
| 1975 | GSW | 12 | 5 | 0.706 | SF | SF | SF | SF |
| 1981 | BOS | 12 | 5 | 0.706 | SF | SF | SF | SF |
| 1984 | BOS | 15 | 8 | 0.652 | SF | SF | SF | SF |
| 1986 | BOS | 15 | 3 | 0.833 | SF | SF | SF | SF |
And Larry Bird & John Havlicek alone make up 2/3 of those cases.
So, of the remaining teams in the playoffs, we have one built around either a C (Amare Stoudemire, regular-season) or a SG (Jason Richardson, playoffs) in the Suns; one built around a SG (Kobe Bryant) or a PF (Pau Gasol) in the Lakers; one built around a center (Dwight Howard) or a PG (Jameer Nelson) in Orlando; one built around a point guard (Rajon Rondo) in Boston; & finally one built around a SF (LeBron James) in Cleveland... And based on these past trends, maybe it shouldn't have been such a surprise that the lone team relying on a SF is on the verge of being eliminated -- since 1952, SFs have been one of the 2 positions (PF being the other) least likely to give you a title, and since 1984 they've been the least likely of all, with no SF having led his team to a title in the last 23 years.
May 13th, 2010 at 1:28 pm
Couldn't you just have looked at career win shares of the top 50 players and found a similar breakdown by position.
May 13th, 2010 at 1:43 pm
A couple things stand out:
Jordan massively skews things towards SG, to the point that removing him puts the SG in the same boat as the SF.
If you consider Duncan a C, which is an understandable position, that dries up the PF category pretty dramatically (I think Gasol played most of his minutes at the C last season too).
May 13th, 2010 at 1:48 pm
I didn't look at the playoffs, but I did something similar a while back:
http://basketball-statistics.com/blog1/2009/07/19/at-which-positions-are-great-players-the-most-important/
and
http://basketball-statistics.com/blog1/2009/07/23/at-which-positions-are-bad-players-the-most-detrimental/
Using adjusted plus-minus, I found small forwards and centers to be the most important.
May 13th, 2010 at 1:57 pm
Great analysis. Quick question, though: are you dismissing Bird as atypical? Because you can really toss out every single superstar player with that same rationale. Maybe a fairer way to judge the teams would be to compare the winshares of the superstars' supporting casts. That way you can try to isolate what position is important in a championship while controlling for statistical superstar anomalies like Bird, MJ, Kobe/Shaq, Duncan, and LeBron. The 2004 Pistons seem like a good model for that, since their entire roster seemed to distribute itself in importance pretty evenly.
May 13th, 2010 at 2:22 pm
The post title sort of implies a choice that the data denies. Most championship teams (that aren't Pistons) have a superstar & a #2 star. If you can get a superstar, you grab him, right? And whether he wants to call himself a PF, Center, SG or statistical superstar anomaly, you build a team around him.
May 13th, 2010 at 2:27 pm
I think LeBron is his own category. He's not just some SF, he handles the ball like Jordan he just doesn't have that Pippen yet.
May 13th, 2010 at 3:59 pm
Is that 2nd chart right? Did Kobe Bryant really lead the Lakers in Win Shares the year they went 15-1 in the playoffs (arguably the most dominant playoff run ever). I thought Kobe just rode Shaq to those titles?
May 13th, 2010 at 4:00 pm
correction... I meand "2nd TO LAST chart"
May 13th, 2010 at 4:20 pm
Kobe was brilliant during that run. He not only led the Lakers, he led all players in the playoffs:
http://www.basketball-reference.com/play-index/tiny.cgi?id=jU1X2
Edging out Shaq by 0.1 WS. With 29/7/6 and outstanding defense, that was as good an all-around performance as we've ever seen from Kobe Bryant.
May 13th, 2010 at 4:40 pm
Not surprised the list is dominated by big men. Historically that has always been the case; MJ is the only real exception (Bird had Parish, Magic had Kareem). And let's stop pretending Gasol is a PF; he outplayed Dwight Howard head-to-head in the finals last year. I wouldn't conclude anything about Lebron based on this though; remember, before Jordan, shooting guard was the weakest position of them all.
May 13th, 2010 at 4:48 pm
The problem with using the the NBA's position system is that all that it really tells you nowadays is what position said player plays on defense. It says nothing as to whether that player is a slasher, spot up shooter, post up player, primary ballhandler, etc. LeBron James and Dwyane Wade are functionally drive-and-kick point guards because they are on ball the vast majority of the time outside the 3 point line.
May 13th, 2010 at 4:53 pm
Ummm... great idea but Lebron, Wade, Kobe...they're interchangeable. I mean they are all just perimeter players that can drive and shoot from the outside. Larry Bird and John Havlicek are more quntessential SFs and also Dr. J. So you have to lump LBJ with the guards.
May 13th, 2010 at 5:27 pm
I think these days positions tend to breakdown more into "point guard," "wings," and "big men" rather than the traditional guards/forwards/centers spots. Not sure how that translates into these numbers, but it seems like in today's league, the SG and SF position are pretty interchangeable, as are the PF and C positions.
May 13th, 2010 at 6:34 pm
So... Picking Oden over Durant was the RIGHT choice, then?
May 14th, 2010 at 12:48 am
"I think these days positions tend to breakdown more into "point guard," "wings," and "big men" rather than the traditional guards/forwards/centers spots. Not sure how that translates into these numbers, but it seems like in today's league, the SG and SF position are pretty interchangeable, as are the PF and C positions."
I really like this.
You have your on-ball players that create with penetration (PGs but some 2s and 3s like LeBron and Wade), your perimeter scorers (wings like Pierce, Dirk), and your back to the basket scorers (bigs like Shaq, Duncan). There are a fair amount of PFs that are functionally perimeter players nowadays (Bargnani, Gallo, Dirk) but I think that this is a much better way to look at offensive players than the tripartite G-F-C system.
May 14th, 2010 at 1:01 am
Also, based on my fairly frequent monitoring of win shares, I think that the most important offensive "position" today in the post-handchecking NBA is by far any player who gets to the free throw line 7+ times a game and shoots 70% from the line. This usually happens because a player is obscenely quick but can also happen if they are strong (Pierce) crafty (Martin) or a huge shooting threat (Durant).
FTAs are an obscenely efficient way to get points, and if you have a player who is a decent passer and team defender who really gets to the line, they are an elite player in this league (Durant, Nowitzki, Bryant, James, Anthony, Wade).
The second most important type of player is your traditional big man. A ton of players that are really underrated in my eyes based on their win share totals are the players who have limited but efficient offensive games, get a ton of rebounds (an extremely underrated ability), and use their size to play effective defense (Howard, Bynum, Camby, Haywood).
May 14th, 2010 at 2:23 pm
I agree with Luke. Positions don't really mean anything. Any player can go anywhere and perform any role. LeBron handles the ball all the time, he could easily be called a point guard. Tim Duncan can call himself whatever he wants, but he's a post player. So are you more likely to succeed with a great post, wing, or point?
May 14th, 2010 at 3:25 pm
The great thing about this game is that a player can dominate at any position at any size
May 15th, 2010 at 2:50 am
pretty sure LeBron is a small forward. he may move around but he always starts out and ends there. the point is that no matter how great your superstar is, choosing a center or point guard is the best bet. there are exceptions, like mj and even dwade did most of the balling for that miami title, but for the most part a great big man or point guard win titles. the small forward spot, because its so interchangeable and versatile like everyone has said, makes it hard to form a team. look at pierce before allen and kg, look at lebron, look at durant they all need their superstar small forward and a great extra players. its easier if your big or point are your superstar because you can plug whatever you need in at the 2 guard and small forward spots (defensive stopper, 3 ball shooter, slasher, etc) and then that makes a team flow better.
May 15th, 2010 at 1:31 pm
Larry Bird, Elgin Baylor, Rick Barry, John Havlicek, Dr. J, Scottie Pippen, Paul Arizin... There have been more great SF's than SG's
And Barry won with the most underwhelming supporting cast in history
May 16th, 2010 at 1:34 am
LeBron handles the ball more than most point guards. He's not just a wing.
Defensively he's more valuable than most players in the league. Position means nothing in his case.
May 17th, 2010 at 11:11 am
Wow ... when are you guys thinking of updating the positions with your new data? I can't believe how I'm excited with that piece of info!
May 17th, 2010 at 11:46 am
I appreciate the attempt, but there's too much history and circumstance to judge based on position alone. You build around the best player available to you. Sam Bowie, Gregg Oden, Kwame Brown, Darko Milicic, Michael Oliwakandi, Bryant Reeves, Shawn Bradely, and Benoit Benjamin. That said, if you have a choice between a sure-thing center and a sure-thing shooting guard, it's not a bad idea to go with the center. But a maybe-thing center is no match for a sure-thing at any other position. In fact, a maybe-thing center will probably destroy your squad and completely waste your draft pick for that year.
May 17th, 2010 at 2:56 pm
The funny thing is, that logic is why Brook Lopez fell to #10, when he was projected to go 3rd. Then it turned out he can actually play; do you think Riley wishes he had the 2nd pick to do over again?
May 19th, 2010 at 12:37 pm
Riley? For sure. Would the Bulls trade Rose for Lopez? I wouldn't. It's hard to give dap to the best player on one of the worst teams of all time. You gotta be worth more than 12 wins.
It'd be cool to see a list of "surprise" awesome centers, who fell to a lower spot in the draft vs. the number of crap-ass centers then went higher in the draft than they should have. My bet is the crap-ass ones are legion. And the surprise ones are pretty few and in between.
I accidently left Hasheem Thabeet off my previous list of stinkers, too.
May 19th, 2010 at 8:36 pm
Who said anything about trading Rose for Lopez? The fact is that elite big men win championships; they don't do it alone of course, but center is the hardest position to fill; if you get a great one, you have a big advantage.
And NJ was so terrible this year due to injuries. They got healthy by the end of the season and started winning.
May 19th, 2010 at 11:03 pm
I was using Rose to elucidate my point about a sure-thing at an other position, versus a maybe-thing at the center position.
May 19th, 2010 at 11:25 pm
I wasn't really disagreeing with you; I just thought it was funny that Lopez fell for that reason and he can actually play
May 20th, 2010 at 11:04 am
I getcha, AYC!
July 16th, 2011 at 8:13 am
Didnt you also have a breakdown showing that teams had to have a solid Small forward more so than other positions?
so is the overall conclusion that although small forward is the least important position to build around you still need to get one?
November 10th, 2011 at 7:47 pm
Greetings. As a self confessed Sports looney your site is intriguing and literally what I have been looking for my sporting infatuation. I'm certainly going to book-mark it and recommend all my sports loving mates about it. Thanks.