How good is Derrick Favors? - An Email Conversation

Andy says: Hello there, Mr. Yucca! I am here to argue with you on the internet because I believe that you are wrong about something. As a result, I must do something about it. Never has this XKCD comic been more appropriate.

Nevertheless, we have such differing opinions at times that it is as if we watch different basketball games, and I wanted to talk with you about your vision. What is it that you see with regards to the Utah Jazz? In particular, you claimed Favors was "the best player on the team right now" and "It's absurd to pretend that Favors isn't already the greatest two-way post player the team has ever had." Those statements are crazy.

Moving past the possibility that he's been our worst big man in the preseason, I think I see how this conclusion could come about: Derrick Favors is the closest thing that we have to a post player who is excellent offensively and defensively. I agree with that. I agree that he could become our best big, he certainly has potential to be great. And yet he's just not there. In every statistic that takes offense and defense into account, he performs worse than Millsap and Jefferson. Consider:

• Favors' offensive PER of 17.5 compared to Millsap's 22.3 and Jefferson's 23.1 means that his defensive advantage (allowing a 15.0 PER compared to Millsap and Jefferson's 17.4 and 16.7) is significantly counterbalanced.
• His 2 year adjusted +/- (from basketballvalue) is only +1.43 points per 100 possessions, compared to Millsap's 10.60. Last year, Favors' +/- was excellent (10.47), but still didn't surpass Millsap's 14.42(!).
• On offense, Favors ranks 202nd in the NBA with a 0.9 PPP from Synergy, whereas Millsap ranks 93rd with a 0.96 PPP. Jefferson ranks 68th with 0.98 PPP.
  
• On defense, Favors ranks 341st in the NBA, with a 0.91 PPP allowed. Millsap ranked 147th, with 0.82 PPP allowed. Jefferson ranks 199th with a 0.84 PPP. x
• Favors ranked fifth in Win Shares, behind Millsap, Jefferson, Hayward, and Harris, and 3rd in Wins Produced, behind Millsap and Hayward.
  

If this were just one or two statistics, then that would be fine, and we could have an interesting debate about the value of one statistic compared to another. But it's all of them. Indeed, in every single database I encountered, I could not find a single statistic in which Favors was leading the team. That's worrisome for your argument.

But even those who don't consider statistics don't look at Favors as the best player on the Jazz, indeed, they don't even consider the possibility. NBARank showed this, but so do all of the season previews of the Utah Jazz: while they all talk about the "logjam", as we've come to refer to it, making a trade is universally considered a likely sacrifice of the "win-now" in order to promote "win-later". Millsap and Jefferson's superiority in the "win-now" perspective makes it clear that Favors has some catching up to do in order to reach those players performance heights.

.

Yucca says: Hello back, Andy.

First, let's get the first thing out of the way:

First of all my claim that Favors is the best "two-way player the team has ever had" suffers from poor word choice. A better description of what I mean is that if you're looking for a post player who is a defensive anchor, who single-handedly transforms the defense of the entire game, and has a strong offensive game . . .; if you're talking about that type of player, then Favors is the best we've ever had.

That out of my way, I stand by my claim that Favors is the best player on the team right now. The key words in that phrase are right now. I did not say Favors had the best overall season last year. I said the best player right now.

When a parent asks me how her son is doing at spelling right now, I don't pull out spelling tests from ten months ago. I pull out recent work. This is especially important if there has been a change in performance over the past 10 months. The old work is useful as a means for comparison. But it shows nothing about how a student is performing right now.

Basically, with all of your stats, you are pulling out spelling tests from 10 months ago. We know he made a significant improvement after the All-Star break. We know this kind of improvement is typical of young players (and thus less likely to be a random hot streak). So it makes no sense to bring up how he played last January when trying to evaluate how good he is right now.

Here's a snapshot on how much Derrick Favors improved over last season. I'm using Wins Produced per 48 minutes. I know it's not your favorite stat, but just accept it as a snapshot and an indicator of how different Favors was at the end of the year compared to the beginning.

Favors' WP/48 during the first two months (Jan-Feb): 0.038. Favors' WP/48 during the final two months (March-April): 0.191.

Whoa.

Now, because I'm interested in right now, and because those last two months are the closest thing we can evaluate to understand what kind of player he is right now, let's look at what Favors was doing those final two months. And let's look at the full range of his production: scoring efficiency (TS%), rebounding (RBD/36), passing (AST/36), deflections (blocks + steals/36), turnovers (TO%), and fouling (PF - offensive fouls / 36 — I'm subtracting offensive fouls because 1) they were already counted in turnovers and 2) they don't lead to FT's, which is the primary reason fouls are bad). And for the sake of time, right now let's just compare Favors these final two months to his competition for best post player: Millsap and Jefferson (since these are also his main competition for "Best Player" title):

	TS%	Rbd/36	Ast/36	Stocks/36	TO%	PF/36
Favors	54.60%	12.4	1.4	2.9	16.20%	3.1
Millsap	54.50%	9.7	2.6	2.9	10.30%	3.4
Jefferson	52.00%	10.1	2.3	2.6	5.20%	2.6

I call that pretty much a wash.

But do you know what isn't a wash? The one aspect of the game not accounted for here: defense. It's not accounted for because there are no stats that correctly show a player's defensive ability. Hollinger admits this about PER. Berri admits this about WP. Locke wrote an entire post explaining why Synergy's stats are useless when trying to see a player's real defensive contributions. PER-against has the exact same flaws that Synergy's stats do. Adjusted +/- struggles to give definitive results with 5 years of data on a player … it's useless when looking at a two month period.

There's nothing but the eyeball test. But a quick look at video tells us something very quickly: Favors is terrific. Defensive Player of the Year kind of terrific. Millsap and Jefferson are bad.

With everything else a wash, but Favors performing that much better at essentially half the game …

My conclusion is simple. Favors is very likely the best post player on the team right now.

.

Andy says: I won't deny that you brought it (the "it" here referring to your A game) with that email. There's no question that you've convinced that Favors is a good player. My problem with your argument, at this point, is one of timing. You play around with minutes, seasons, months, etc. and pull out much of the meaning from the data in order to get a small enough sample size, on which Favors then equals Millsap and Jefferson.

1. 2 months of the season recently passed is not enough to say very much about a player's future, or even current ability. Consider C.J. Miles, who had an excellent 2 months to end the 2009-2010 campaign. Consider Andrei Kirilenko's tendency to have an excellent six week stretch in any given season. Consider even Paul Millsap's first two months of last season, in which he was an All-Star level player. When evaluating a player, we have to look at a larger body of work than two months at a time.
2. Per 36 minute stats can tell some funny stories. Consider Jeremy Evans' rookie season, and the same stats that you presented:

	TS%	Rbd/36	Ast/36	Stocks*/36	TO%	PF/36
Favors	54.60%	12.4	1.4	2.9	16.20%	3.1
Millsap	54.50%	9.7	2.6	2.9	10.30%	3.4
Jefferson	52.00%	10.1	2.3	2.6	5.20%	2.6

Evans 67.8%(!) -- 7.5 -- 1.9 -- 2.6 -- 12.6% -- 3.9

Look at how much better Evans is than everyone! He should get all of the minutes! This comparison seems ridiculous at first, because Evans played such limited playing time. But he actually played about as many minutes in his rookie season as Favors did in the last two months of last year: in short, the sample sizes are actually pretty close.

*Steals are much more valuable than blocks because you're guaranteed to gain possession, whereas with blocks you aren't. So you're being slightly unfair to Millsap.

3. Preseason. This is a little bit of a debate low-blow as you wrote your proclamation before the preseason. But to whatever extent you can take something from preseason, Derrick Favors' performance has not been "best player on the team" worthy. He's ranked 7th on the team in points, 4th in rebounds, <1 APG, and has the 3rd highest TO/G rate. He's shot only 45.8%. Kanter, on the other hand, has led the team in points, rebounds, FG%, and even has the fewest TO/G, in almost the same mount of time Favors has played (21.3 vs. 22.1 MPG). In short, if you take preseason into account at all, Favors isn't even close to "best-on-team" production.

One note re: defense. Yes, defensive stats have flaws. But rather than dismissing them all off-hand because of their flaws, let's use them to illuminate points in the video that we may have otherwise missed. For example: a large deal of Favors' struggles in Synergy's defensive stats are his quite-poor numbers against spot-up shots. Looking at the video, it's clear that this is the case; for the most part, Favors likes to be in mid-to-post area, possibly leaving his man if he is defending a perimeter 4. As a result, PFs with shooting ability have done extremely well against Favors, memorably Matt Bonner. He's not a complete defensive player, and the stats and eyeball test show that. The stats, then, can elucidate the relative advantages and disadvantages of Millsap, Favors, and Jefferson on defense. Of the stats, I find the PER against and defensive +/- most useful as tell a reasonable tale: Favors as one of the best defensive players on the team, and Millsap/Jefferson as average-to-below-average. But then looking at the offensive side of each of those stats, Millsap and Jefferson are enough better than Favors to more than make up the difference.

.

Yucca says: You are correct that two months is not enough to definitively say anything. I can agree with that. There are no guarantees. And I think many fans, the coaches, and the front office are hesitant to call Favors the Man until they have guarantees. Unfortunately there will be no guarantees until he is actually given a starting spot and 30+ minutes per game. Only then will anyone really know.

But to address your main counterpoints:

1. You're correct that 2 months of the season isn't always enough to say anything. But that kind of leap for a new NBA player is often (not always, but often) representative of what will come in future seasons. I did a quick look at Basketball-Reference.com, and here's a list of ended an early season (Roookie to 3rd year) with a two-month surge that fairly accurately predicted what would be coming in future years: James Harden, Russell Westbrook, Ty Lawson, Joakim Noah, Manu Ginobili, LaMarcus Aldridge, Brandon Roy, Rudy Gay, Andrew Bogut, and Al Horford.

There are no guarantees, but it happens enough for me to feel somewhat confident that what we saw after the All-Star break is pretty representative of what Derrick Favors is capable of right now.

2. You have brought up Jeremy Evans twice now (once here, and once in your DB). So I take it you don't think he can be an effective player with more PT? Maybe that should be our next argument, because I think he'll be very effective.

At the same time, there are some major differences between Jeremy Evans and Derrick Favors … and that is minutes per game AND what kind of game situations they played in. I expect that both of us can agree that playing in the last 5 minutes in a game that is already decided is quite different than playing 20 minutes against competition still doing everything they can to win. Besides, during the period I'm talking about Favors played 721 minutes. Evans played 463 in his rookie year. That's not really very close to the same.

3. Yes, Favors has underwhelmed in the pre-season. I don't deny this at all. It's been pretty disappointing for a guy solidly in the pro-Favors camp. Though he at least came alive the final couple games. In the end, we'll see whether the pre-season portends anything about what our guys will produce this year. But if it does, we shouldn't be too excited about Al Jefferson's upcoming season either.

4. Defense. The stats can be useful if you look at them for what they are actually telling you. Your example regarding Favors and spot-up shooters is a perfect example of this done right. However, for a player like Favors, his real strength on defense is how he affects the entire team—not just how he guards his own guy. PER-against doesn't give us this. Defensive +/- has WAY too much noise to tell anything definitively without at least two years of data (and even then there's still a lot of noise). No matter how you frame it, defensive stats only give us a very fuzzy, hazy, and nebulous reading of a guy's defensive abilities. This is in strong contrast to offensive stats. While not perfect, they do give us a fairly complete view of a player's contributions on offense.

Regardless, Favors' production during the final two months suggests to me that the difference between his offensive abilities and that of Millsap and Al is much smaller than you suggest. Quite a bit smaller than the difference in their defensive abilities.

In the end, it really depends on how much you believe that two-month surge is representative of what Favors can do right now.