The Dean Oliver Interview

Baseball is in the midst of a revolution of sorts. No it’s not about steroids or home run records, but rather the wave of statistical analysis that is hitting the league like a Pedro Martinez high and tight fastball. The grandfather of this uprising could be Branch Rickey, the former Dodgers GM. More than half a century ago, he knew there were flaws in measuring a player’s value with stats like BA, RBI, and even fielding percentage. Even though the baseball world ignored these simple findings, other people did not. A small group of people asked: “what are the best tools we have to evaluate baseball players?”

The fathers of the revolution are guys like Bill James, Pete Thorn, and John Palmer, who have given birth to the modern day soldiers. Current GM’s like Billy Beane, J.P. Ricciardi, Paul Depodesta, and Theo Epstein value statistical analysis over observational appraisal. Columnists like Rob Neyer are part of the mainstream and have forced the old regime to pick up new terms like OPS to adapt. Voros McCracken, a lawyer turned baseball consultant, turned a world of baseball thinkers on their head with his theory that (with few exceptions) pitchers have little effect on a batted ball in play. The movement has hit the streets with a new generation of writers from all walks of life preaching the new credo.

What does this have to do with Dean Oliver? His book is part of a parallel movement in basketball. Dean’s goal with his book is to better understand basketball through statistics. Where current statistics don’t give enough information, Dean creates his own. Whether you’re a casual fan or a failled GM looking for a way to improve your way of thinking, Dean Oliver’s book should be of interest to you.

“Basketball On Paper” tries to answer the tough questions. Is Team A really good defensively or does their slow tempo give that illusion? Read the book. Which team was the best offensive team of all time? Read the book. Can my team benefit from playing a risky strategy or should they tone it down? Read the book. How valuable is Iverson with his low FG%? Read the book.

The fires of same revolt are slowly starting to kindle in basketball. No there are no Billy Beane’s in the front offices of the NBA, yet. Nor is Bill Walton scrambling to learn what PER means. Jon Hollinger’s “Basketball Prospectus”, columns like Kevin Pelton’s Page 23, discussion groups like the APBR, and Dean Oliver’s “Basketball on Paper” will change the way people think about basketball, like their sabermetric cousins are currently doing in baseball.

And now, on to the interview:

“Basketball on Paper”:

KB: Since writing your book “Basketball on Paper,” what have you been working on since & what can we expect from you in the future?

I’ve actually been working on marketing the book and convincing NBA teams of the value of this kind of work — any of the type of statistical work that is logical and reasonable to help a team. I have been talking to a number of teams, the league office, and to media. I’ve done a couple of limited studies relating to work I’ve done for years for the Seattle Sonics. I’ve written a couple things and I am outlining another book on statistical approaches to basketball strategy.

KB: What were some of the responses to your book?

I did a radio show down in Tampa and Doc Rivers came up to me after hearing it and really wanted to talk to me and get a copy of the book. That was nice. I sent the book on to Bill James, who in an email to me said it was an excellent book. He is planning an endorsement. I know that some readers get concerned by seeing formulas or numbers of any sort, which is a bit of a shame since they really are there to support stories and lessons of how the game works and how good talent is.

KB: What is missing or what is next in the literature on basketball statistics?

I think what is missing is the same large audience that baseball stats have. There isn’t a lot of money in a lot of the basketball writing at this point. And it’s a difference in the games. Baseball is slow enough that fans can really talk about what’s going on, pull in numbers. Basketball numbers are tremendous, they’re plentiful, and they’re insightful, but the time to use them or to make adjustments based on them is just not there. So fans look for the excitement and coaches make a lot of decisions based on gut feel. I’ve introduced some rules of thumb for hoops that can be applied on the fly, but it’s going to take a little while to catch on.

What is next in line for research is really a translation system for players from one league to the NBA. College stats have the best history, but high school and international leagues have significance now. We’re just catching up on this and it is not an easy problem. I’ve figured a few things out, but I have too many rules like “The Pitino Rule” that are just annoying and not very general. (Basically, Pitino’s system causes some of his players to be overvalued by stats.)

KB: When I read your book, “Basketball on Paper” I couldn’t help but think of it as the “Win Shares” of basketball, because of :

  • It’s ability to try and understand defense, where traditional methods are lacking
  • The clarity in which advanced statistical ideas are presented
  • The author’s humor

Do you see yourself making a book similar to “The New Bill James’ Historical Abstract”, either by making a recount of basketball by historical periods, or by ranking players by position?

I would very likely do such a book in combination with people like John Grasso, Bijan Bayne, Harvey Pollack, and my contacts in the NBA. The statistical library to evaluate the older players is incomplete, so video and news stories would be very important to make some accurate representations. These guys could definitely help. And I wouldn’t want to do what Elliott Kalb did with his book recently either, which is subjectively rank guys without a lot of link to objective evidence. You mention the recount of basketball by historical periods and I have done a lot of that. Total Basketball does some work along these lines, but doesn’t really talk about why Wilt scored 50 ppg in 1962, for instance, which is important. I mention those things — pace, less judiciousness in shot selection, lack of double-teams — in Basketball on Paper and would like to assemble more of that in a book like this.

Is this something that is right around the corner? Probably not. The group of interested basketball historians is growing, but it isn’t large enough to pay for the massive research involved. I hate having to say it that way, but basketball writers do need to make a living.

Stats:

KB: On ESPN’s MLB main stat page, they have a “sabermetric” stat OPS on the main page, and a sabermetric page with such stats as isolated power and runs created. On ESPN’s NBA page, the closest they have to advanced metrics is points per shot and adjusted field goal percentage. How long do you think it will take before the major sports web sites will post stats like your off. Rtg., stop%, or Hollinger’s PER?

We’re getting there. The WNBA will put out possession stats and points per possession stats this year. The WNBA is also quite receptive to some of the defensive work we did with them a couple years ago and we’re hoping to do that again this year. Issuing monthly reports on what individuals forced the most misses would be nice small pieces of information that help that league and intrigue the bigger brother, the NBA.

The NBA does now have something it calls “efficiency”, which is just a sum of the good things minus the bad things, something I’ve heard referred to as “plus/minus” or is almost Tendex or Bob Bellotti’s Points Created. Going to the more advanced individual numbers is going to take a while. Putting team numbers out there first is huge. Then, in a couple years, we can start talking about other stuff. What we don’t want is another NFL-like passer rating stat that everyone jokes about as the lead stat.

KB: How would you define who is the best rebounder in the league? Would you simply look at REB/48min, or are there other considerations (incl. FG% and pace)?

Primarily, rebounders can be evaluated by the percentage of available rebounds they get. With 10 guys on the court, 10% would be average. I think Rodman was up around 20%. This accounts for pace and FG%. What it doesn’t account for is whether it really makes the team a good rebounding team, which is a modifying factor. A guy who gets 15% of available rebounds on a bad rebounding team is not as useful as one who gets 15% on a good rebounding team (because he is competing against his own teammates for boards). But that is relatively small in terms of importance.

KB: Is there a way to translate a player’s college statistics into his success at the professional level?

As I said above, this is one of the most important projects on the front burner. I’m getting better at it by studying the cases that just don’t work the first time and figuring out why.

KB: What has surprised you about the NBA season this year so far?

In terms of wins and losses, there aren’t a lot of surprises. I’m a little surprised at how successful the Nuggets are. I thought they’d be a lot better, but not this much better. I have been surprised by the coaching turnover. Doc Rivers was out quick, though the losing streak made it less surprising. Byron Scott’s firing and the discord between Ainge and Jim O’Brien — those really surprised me. I also was surprised at how cheaply Portland got Shareef Abdur-Rahim.

The Knicks:

KB: What do you think of the job Isaiah Thomas has done with the Knicks this year?

He wasn’t shy. Being decisive is usually a good thing, but his history in charge generally concerns me. So, picking up Marbury is a nice thing. Marbury is a very good piece to have. I just don’t know if the baggage of the other players, plus the losses were worth it if he’s trying to build a championship team. It’s more important that Isiah know how to follow up the moves he has made than just standing on what he’s got. So he gets a big Incomplete at this point.

KB: If you had to, where would you rank Stephon Marbury among today’s point guards?

Jason Kidd is solidly above him. Otherwise, he is in a class of points who are quite good (Nash, Parker, Billups, Cassell, Payton, and I’m sure I’m forgetting someone). Marbury is the best on the offensive side, but his defense lacks. Intangibly, he’s worn out his welcome a few times now and that matters.

KB: In your opinion, what player (current or former) is Nazr Mohammed most similar to? In the best case scenario what player could he aspire to become?

Curious that you ask that. Curious that the Knicks really valued him. One of my techniques for looking for players is to look at successful players in the league and see what guys are modestly similar to them who may not be getting much time. Mohammed came out modestly similar to a couple prominent big men in a study I did for the Sonics. I think Garnett or Duncan was one. But I have sooo many numbers that suggest he just isn’t going to end up good that I discarded him as a legit possibility. I did all that right before the trade, so I found it very coincidental. Best case: he might be Jamaal Magloire. He just can’t consistently score inside.

KB: Is Kurt Thomas worth $30 over 4 years, or could the Knicks have gotten similar production out of a cheaper and younger player?

Kurt Thomas is serviceable. He’s not a bad player. But he’s just not someone I see as a significant component of a championship team. It’s a risk to go after a younger player, but I would have taken that risk.

One final question:
KB: What do you like best about the Knicks?

Weeellllll, I do tend to root for Lenny Wilkens. He may be old-school and not care about the kind of work that I do, but he’s always seemed like a decent guy.

The Best & Worst Offenses & Defenses

I dragged this lake looking for corpses
Dusted for prints, pried up the floorboards
Pieces of planes and black box recorders
Don’t lie

— “Private Eye”
Alkaline Trio

If you wanted to know quickly which teams had the best & worst offense & defense in the league, you might go to ESPN.com NBA stats page. You can click on the team-by-team comparison link and you’d see which teams have scored and given up the most points. You can sort the teams by a few different stats, and for this example you might use points scored per game and opponents points scored per game. So according to this you might conclude the best NBA offenses are:

Rank	Type	Pts
1	SAC	104.6
2	DAL	103.3
3	SEA	97.8
4	LAL	97.6
5	MIL	97.1
6	DEN	97
7	MEM	97
8	LAC	96.5
9	MIN	95.5
10	ORL	94.9
11	BOS	94.6
12	GS	94.4
13	PHO	93.6
14	CLE	92
15	POR	92
16	NO	91.9
17	NYK	91
18	NJ	90.7
19	WAS	90.6
20	SAS	90.1
21	IND	89.8
22	ATL	89.7
23	CHI	89.6
24	PHI	89.6
25	DET	89.6
26	UTA	88.9
27	HOU	88.3
28	MIA	88
29	TOR	84.9

And the best defenses:

Rank	Type	Pts
1	SAS	83.9
2	IND	84.9
3	HOU	85.1
4	NJ	86.5
5	DET	87
6	TOR	87.3
7	MIA	89.4
8	MIN	90
9	UTA	90.6
10	PHI	91.2
11	NO	91.8
12	NYK	92.6
13	POR	93.6
14	LAL	94
15	MEM	94.4
16	GS	95.2
17	ATL	95.2
18	DEN	95.3
19	CLE	95.3
20	MIL	95.4
21	CHI	95.7
22	WAS	96.2
23	BOS	96.6
24	PHO	97.2
25	SAC	97.3
26	SEA	98.8
27	LAC	98.8
28	DAL	99.8
29	ORL	100.7

But if you’ve ever seen an NBA game, you know some teams’ offenses are faster than others. This is because some teams like to run down the shot clock because they want to slow the tempo, where others like to take the first or second available open shot, because their offense is good enough to get an efficient shot off quickly. A team like Dallas or Sacramento would have more chances to score than let’s say Miami or Toronto. So the team that scores the most points in a game may not have the best offense in the league, because they simply might have the fastest offense in the league.

Well some smart people have already thought about this. They said, maybe instead of looking at how many points are scored per game, we should see how many points are scored per possession (or 100 possessions). Unfortunately the people at ESPN have never thought to ask this question, or at least have not thought it important enough to put it on their stat page. Luckily the good people at 82games.com have. You can’t sort the teams all nice like the ESPN page, and they don’t even have a page where all the teams are displayed. They only have the information by team pages. But with some patience, a good spreadsheet program, and 10 minutes of free time, you can do some quick analysis yourself.

So here we are, the best offenses in the NBA, by points per 100 possessions (pPts):

Rank	Type	Pts	Poss	pPts	OldRank	DRANK
1	SAC	104.6	91	114	1	+0
2	DAL	103.3	91	112	2	+0
3	SEA	97.8	89	109	3	+0
4	MIN	95.5	87	109	9	+5
5	LAL	97.6	90	108	4	-1
6	MIL	97.1	90	107	5	-1
7	MEM	97	90	107	6	-1
8	LAC	96.5	90	107	8	+0
9	POR	92	86	107	14	+5
10	DEN	97	91	106	7	-3
11	ORL	94.9	89	106	10	-1
12	GS	94.4	89	106	12	+0
13	NO	91.9	88	104	16	+3
14	IND	89.8	86	104	21	+7
15	UTA	88.9	85	104	26	+11
16	BOS	94.6	91	103	11	-5
17	PHO	93.6	91	103	13	-4
18	NJ	90.7	88	103	18	+0
19	SAS	90.1	87	103	20	+1
20	DET	89.6	86	103	23	+3
21	HOU	88.3	85	103	27	+6
22	MIA	88	85	103	28	+6
23	CLE	92	90	102	15	-8
24	NYK	91	89	102	17	-7
25	PHI	89.6	87	102	24	-1
26	WAS	90.6	90	100	19	-7
27	ATL	89.7	90	99	22	-5
28	CHI	89.6	90	99	25	-3
29	TOR	84.9	85	99	29	+0

The first team that jumps into the top 5, that wasn’t there before, is the Timberwolves at #4. Even though they only score 95.5 points a game, they would score 109 points if given 100 possessions. They just play at a slow pace of only 87 offensive possessions per game. (The average pace is 88.4 possessions/game). Portland also jumps into the top 10, while Utah & Indy move into the middle from the bottom 10. The big losers are Cleveland, Washington, and my own beloved Knicks, who drop to 6th worst.

Now for the defenses:

Rank	Type	Pts	Poss	pPts	OldRank	Diff
1	SAS	83.9	87	96	1	+0
2	IND	84.9	86	98	2	+0
3	NJ	86.5	88	98	4	+1
4	HOU	85.1	85	100	3	-1
5	DET	87	86	100	5	+0
6	TOR	87.3	85	102	6	+0
7	MIN	90	88	102	8	+1
8	LAL	94	90	103	14	+6
9	PHI	91.2	87	104	10	+1
10	NO	91.8	87	104	11	+1
11	NYK	92.6	89	104	12	+1
12	MEM	94.4	90	104	15	+3
13	DEN	95.3	91	104	18	+5
14	MIA	89.4	85	105	7	-7
15	MIL	95.4	90	105	20	+5
16	UTA	90.6	85	106	9	-7
17	GS	95.2	89	106	16	-1
18	ATL	95.2	89	106	17	-1
19	CLE	95.3	89	106	19	+0
20	CHI	95.7	90	106	21	+1
21	WAS	96.2	90	106	22	+1
22	BOS	96.6	91	106	23	+1
23	SAC	97.3	91	106	25	+2
24	PHO	97.2	90	107	24	+0
25	POR	93.6	86	109	13	-12
26	DAL	99.8	91	109	28	+2
27	SEA	98.8	89	110	26	-1
28	LAC	98.8	90	110	27	-1
29	ORL	100.7	89	112	29	+0

Not much difference here among the top teams. The Lakers move into the top 10. Meanwhile Miami and Utah drop out of the top 10 into the middle of the pack, while Portland takes the biggest dip to the 5th worst defense in the league.

Unfortunately it’s hard to judge the Knicks current team using these tools. The team has experienced a large turnover in their roster. These stats are for the Knicks over the entire year, which includes players like Van Horn, Charlie Ward, as well as Marbury and Penny Hardaway. However we can use it to learn a bit about tonight’s opponents, the Kings. The Kings are the team’s top scoring offense both by points per game, and by points per possession. Their 91 possessions/game tell you that they have one of the league’s fastest paced offense. Their defense is lacking, 7th worst in the league.

So then you might ask, if they have a great offense and a poor defense, shouldn’t they be an average team? Not necessarily. If anything they might be a little bit better, since their Pythagorean expected wins says they should have about 1 or 2 more wins, based on their points for/against. However, you can look at the offensive numbers above, and realize that they are a LOT better offensively than most teams. The difference between them and the #5 Lakers, is 6 points (per hundred possessions). This is the same difference between the #5 Lakers and #23 Cleveland.

So the Kings’ offense is very good, which is why they have the best record in the league.