Baseball Prospectus - Chat with Kevin Pelton

Chat: Kevin Pelton

Welcome to Baseball Prospectus' Thursday April 09, 2009 1:00 PM ET chat session with Kevin Pelton.

Check in on what\'s going on with NBA action with Kevin Pelton of BasketballProspectus.com.

Kevin Pelton: Hey everyone. Thanks for joining me to chat today as the NBA regular season winds down and we prepare for the playoffs. If you've got some questions about the conclusion of the NCAA Final Four, feel free to share those too. I'm also preparing for the WNBA Draft at 3 p.m. Eastern today (coverage on storm.wnba.com), so bear with me if I have to drop out for a second and definitely share any questions you have there! Let's begin ...

Cole (Portland): What is the best/fave NBA blog?

Kevin Pelton: You mean besides BP Unfiltered? Ball Don't Lie and True Hoop.

Oh, and hi, folks from Blazer's Edge. Thanks for joining us and of course I have nothing but love for the great work Dave and Ben do covering my new "hometown" team.

tominhawaii (Honolulu): I think advanced stats are destroying half the fun of being a fan. I can't say a player made a clutch shot or a is a good fit for a team without a "stathead" telling me I'm wrong and using a new stat to "prove" his point. Do you think there will be new stats to account for team chemistry, player roles, and other intangibles, or a compromise between stats and intangibles, or an apocalyptic future where stats become the only word in all conversations?

Kevin Pelton: Anyone who thinks the numbers truly prove points in basketball is not a true APBRmetrician. To paraphrase the great Bill James, the numbers suggest more than they prove. I agree that the zealots can make things less fun, but the great thing about basketball statistics is most of the time there is plenty of room for interpretation.

That said, when there are numbers easily available that do paint a pretty clear picture one way or the other, to be upset when they contradict your personal opinion is to be cutting off a chance to improve your understanding of the game. I don't think that's especially fun.

Village Idiot (Norway): Fans of teams often think that specific officials are out to get them. Is their any data showing a strong corelation to teams winning or losing more irregularly often when a particular referee works the game?

Kevin Pelton: I'm sure that kind of information is out there on some Web site (the Vegas-related ones tend to keep a much closer eye on refs than the statistical analysis community). Statsheet.org does have some referee tendency data: http://statsheet.com/nba/referees

Teams also tend to track this information, and I'm sure there are some referees with whom they have a better or worse record. The sample sizes are pretty small over the course of a season, though, and odds are it's mostly noise.

SaberTJ (Cleveland, OH): Hi Kevin, Will the Cavs even lose a game in the playoffs before the Eastern Conference Finals? I know they will have a road game or two against Atlanta or Miami, but I just can't see them having a tough enough time until they face Boston or Orlando.

Kevin Pelton: Yeah, they'll lose a game. Remember that last year we thought Boston had an easy road to the Eastern Conference Finals -- and certainly a favorable matchup in the first round -- yet the Celtics needed the maximum 14 games to get there. I could see a Cleveland-Miami matchup taking a while to play out along the same lines as Boston-Cleveland last year, though odds are the Cavaliers won't need more than about 10 games over the first two rounds -- which is a huge advantage over whoever survives the likely Boston-Orlando showdown.

Cole (Portland): How cool is Bill Simmons to hang out with?

Kevin Pelton: To answer that would be overstating the amount we interacted dramatically, but he lends some valuable coolness to a gathering of stat geeks, that's for sure.

Jacob (Portland): How do the databases used by teams like the Rockets compare to publically available numbers on Basketball Prospectus, Basketball Reference, etc?

Kevin Pelton: Well, I've never seen what the Rockets are doing, but I would imagine they are very different. Our focus is on providing a picture of player value for fans, but that's just one element of what teams that are really strong in analytics are considering. A lot of their numbers relate to player tendencies, both for scouting opponents and for getting a better handle on how a newcomer would fit into their system.

Sam (Bellingham, WA): It seems like alot of the advanced stats coming out are being used to predict behaviors (what player x will do in situation y) more than being used to compare players like with baseball stats. Is this the current trend? And how do you see the evolution of basketball stats continuing?

Kevin Pelton: The complexity of understanding the NBA game and the fit between teammates makes those kind of head-to-head comparisons much more difficult. Tendencies help explain those interactions, and are an area where we can get useful numbers with less subjectivity.

I do think eventually we'll have more stats available to the public along the lines of those teams are tracking. And I think we'll get a better understanding of how to use adjusted plus-minus optimally to help us understand the value of those statistics.

Tim (California): Do you think the current generation of basketball statistics will still be useful in, say, 10 years? Or will improved data collection techniques, combined with another generation of analysis, give us much more advanced statistics compared to today?

Kevin Pelton: I think they'll still be useful to some extent. Offensively, I'm not sure we'll be a ton further along. However, if you look backwards 10 years to where we were (Dean Oliver was doing some interesting work, while Hollinger was just starting to hone PER plus-minus was only kept by teams and I was a junior in high school) it's obvious the immense progress that has been made. The pace has slowed a little over the last five years, as I wrote about recently at BP. Whether that changes depends on how many bright minds aren't snapped up by teams and how much the increased attention for statistical analysis translates into more people joining the community.

CMCWizard (Portland): Tom Ziller over at FreeDarko just posted an interesting way of evaluating players in the Z-system. My initial thought is that it is a great way to visually see where your players fall, and see graphically what the strengths and weaknesses of a given line-up may be. The main pitfall seems to be the objective nature of determining wether the player has "continuity" (blue dots) or that characteristic is a "isolated occurance" (red dots). How would you suggest to improve this "metric" to add a more objective approach? Or is this way of evaluating players and line-ups destined to subjective speculation?

Kevin Pelton: Yeah, Tom and Nathaniel did some very interesting work with the Z charts. I think it's a graphical way of representing something you can think about statistically. If you are able to translate each skill into a metric -- and a lot of those can be represented numerically -- and set a minimum threshold for calling it a strength (and potentially also a weakness, to add another layer to the analysis), it becomes quantifiable. Then you can overlay five players somehow to get a balanced lineup.

The one thing you'd probably want to add to that is a positional defense indicator. A lineup isn't going to work unless you've got guys who can guard each of the five positions.

Morgan (Melbourne, Australia): Who will end up in the 3rd spot in the western conference. There is currently a 3 way tie between the Spurs, Rockets, and Blazers. FYI, if you answer Portland everyone will continue to call you a "homer". But since you've already addressed them as your new hometown team, I'd say just go with your gut.

Kevin Pelton: I think the Rockets have to be the favorite right now based on the combination of schedule and holding the tiebreaker with Portland and potentially San Antonio as well. (Great breakdown of the tiebreaker scenarios on -- where else? -- Blazer's Edge this morning).

Joshua (Austin, TX): I think the NBA needs to push the D-League to one team per NBA squad, expand the draft to 4-5 rounds, and generally make it a true minor league system. You?

Kevin Pelton: I'm not sure about a true minor-league system. I don't think it's ideal to have all these late-round draft picks under team control for several years and I'm happy with the length of the draft right now. I do think we're eventually headed toward having one affiliate for each team. In Austin, the Toros and the Spurs are probably the model D-League/NBA relationship right now, with the Spurs regularly calling players up from their team to get a look at them in an NBA setting. The fact that Austin is running the same system helps make that easier. But technically those players are free to sign with any team, and I don't see that changing.

Hitchenstein ((Worcester)): Bodog is running a NBA futures promotion where your bet is refunded if the Lakers win the championship. If there wasn't a $50 limit, I bet the farm on the Cavs, right?

Kevin Pelton: Interesting. Depending on how much they're paying, it actually might make sense to bet on all three top East teams, knowing that the worst-case scenario is probably winning one of those bets and losing the other two. The fact that there are so few legitimate title contenders makes for some interesting possibilities.

Fund A Mental (PDX): What percentage of 3 pt baskets are made by left handed shooters? Is this percentage equal or lower than the percentage of left handed players in the league? Does this correlate in any way with the amount of dunks by left handed players?

Kevin Pelton: You know, I was thinking about this the other day. Is there any place with comprehensive data on lefties? I'd love to see that at Basketball-Reference.com. We probably don't talk enough about handedness, given the role it plays in our mental picture of a player. Earlier this week I did player comparisons for Hall of Fame candidates at True Hoop ( http://myespn.go.com/blogs/truehoop/0-39-31/If-Hall-of-Fame-Candidates-Played-Today.html ) and it occurred to me that one of the reasons Michael Redd seems like a good comp for Chris Mullin is simply the fact that they're both lefties.

Wendy (Madrid): Is Salmnons the Bulls' MVP, or is he just a really good sidekick to Rose? They've been so much better with him.

Kevin Pelton: The numbers at 82games.com don't show a huge difference. I haven't seen them play enough since the deadline to say for certain.

strupp (madison): Is this the last gasp for the Spurs in this incarnation or do they have one more run in them? Do they make an effort for a big gun to pair with Parker and move duncan to a later David Robinson type role?

Kevin Pelton: Well, I think that's what they considered doing with the Vince Carter rumors. If - and this is a big if - Duncan and Ginobili both are healthy next year, it's still too early to write San Antonio off. But years of lengthy playoff runs (and international basketball) have taken their toll. Ginobili's style has always been conducive to injury. The Spurs need to find more guys who can create their own shot -- maybe George Hill evolves into that for them.

I can see San Antonio winning the division next year, but I have a hard time seeing them breaking back into the league's top three or four teams.

uoduckfan33 (Portland, OR): In your mind, what is the statistical value of an assist in the NBA? I believe it was Simmons who came out with an article about the expanding definition of the assist, recently, and it has got me into thinking about the real value of an assist. In your mind is the assist/TO ratio better? What about an assist/possession statistic?

Kevin Pelton: The Wall Street Journal had an interesting story on the subjective nature of assists recently: http://online.wsj.com/article/SB123855027541776617.html

Adding the context of possessions helps somewhat, but ultimately the big problems are that the standard for assists is very different depending on the team/arena (an issue we don't really have with other NBA stats) and that simply looking at assists doesn't account for their difficulty or quality. I'm not sure how we help address the second factor without introducing even more subjectivity.

It's also hard to quantify the value of assists because at the team level they aren't really that important (due to the value of having players who can create their own shot).

Cole (Portland): Do you think the Blazers could win a 1st round match-up with homecourt?

Kevin Pelton: Blazer's Edge lighting round before we wrap things up here ...

Yes, definitely. It's going to be a challenge for a team to steal a win at the Rose Garden in the postseason.

Reid (Portland oregon): If the Blazers face the Spurs in the playoffs, do you think that the Blazers have a good chance to win that series with or without homecourt?

Kevin Pelton: I think that would be the one series where I'd pick Portland without the benefit of home-court advantage. Even with Ginobili, I thought San Antonio was fairly vulnerable because their point differential has been middling all season long.

Sam (UT): The Blazers have really seemed to come together down the stretch, and I know you don't think their inexperience is going to play a major factor in the playoffs. I used to think their ceiling this year was the second round, but now I think they could push it even further than that. What do you think?

Kevin Pelton: I think you have to give them a chance of reaching the Western Conference Finals. Their point differential has been impressive throughout the season, and that tends to be a better predictor of postseason success than win-loss record, at least when you account for seed.

Joel (Seattle): Which team do you think Portland matches up better with and gives them the best chance of a first-round win in the playoffs: San Antonio or Houston?

Kevin Pelton: Definitely San Antonio at this point.

lennyd (Portland): Hi Kevin, Do you see any team that can seriously challenge the Lakers in the West?

Kevin Pelton: Coincidentally, Henry Abbott has a good post about that topic this morning:
http://myespn.go.com/nba/truehoop

By the numbers, the Lakers aren't quite as overwhelming a favorite as it seems. I do think using L.A.'s regular-season numbers slightly understates their chances because the Lakers tend to play to their opposition, as I've shown statistically.

I think the team with the best chance of knocking the Lakers off might be Houston. Ron Artest and Shane Battier could make for a very frustrating seven-game series for Kobe Bryant.

dshugert (ohio): Next year's most improved team will be the ___________

Kevin Pelton: Oklahoma City Thunder or the Minnesota Timberwolves.

Kevin Pelton: Alright, thanks to everyone (including the BE Army) for the great questions. We'll try to do this again in the not-so-distant future. In the meantime, keep reading Basketball Prospectus.