rico381 wrote:Russell: My big concern is his offensive impact. I've seen the numbers about how, despite the reputations of some of Russell's teammates as offensive stars, the Celtics won as an all-time great defensive team with an average-to-poor offense most years. What concerns me is that the discussion then becomes solely about how much value Russell provided on the defensive side of the ball, without any concern for the offensive side. Looking at the Rel_ORTG and Rel_DRTG provided on basketball-reference, we see the Celtics go from a +0.7 SRS team in '56 (+1.9 O/ +1.4 D) to a +4.8 SRS team in '57 (-0.4 O/ -4.9 D). If you just look at the defensive side of the ball, there's a 6.3-point shift there. That's certainly the way it appears in posts like this:
http://elgee35.wordpress.com/2010/12/31 ... ve-impact/ (which, despite my minor criticism here, is truly excellent work). Looking at the total package, though, it's a 4-point shift that we're looking at, because the offense dropped off by two points the year Russell came in (after being at least 1.9 points above average each of the six years before), and stays below average every year of Russell's career. There are a few possibilities here that I can see, but all of them are concerning in some way:
(snip)
A few times in this thread, I've seen Russell supporters argue that Boston was an average-to-poor offensive team throughout Russell's prime, and then use this as a point to Russell's credit. While yes, it does make their defense more impressive, we can't just ignore their struggles, either. Right now, the evidence I'm seeing indicates that Russell was a below-average individual offensive performer on a below-average team offense.
I saw at least one poster estimate that Russell was a +10 on defense, while the best offensive players were +9/9.5. The problem is that we're then left comparing Russell's offense, which looks like it was at best average in his own time (and possibly well below) to Jordan's defense, where he's on the shortlist to be the best perimeter defender ever. That makes me think Jordan's total impact would be higher than Russell's, unless you'd take issue with the numbers used and would place an even larger gap there. If you think Russell's defense is several points more valuable than the offensive GOAT's offense, then sure, but that sounds like it's stretching the case a bit far. I am definitely looking forward to hearing input on this issue from Russell supporters, though.
This was another aspect of your post that I wanted to respond to, the concept of whether Russell could have had as much impact on (primarily) defense to have been worth more than Jordan's more two-way impact.
1)
Can't add those numbers like that. I'd say that the math that you mention in the underlined wouldn't be linear. If I'm not mistaken, those numbers were first posted as more of a "scale of 1 - 10" than any type of impact estimate. It might make sense, for example, to try to come up with an individual SRS-like value for offense vs defense and then sum it to approximate total impact. But if you're doing it on a scale of 1 - 10, that addition falls apart in large part because there's nothing higher than 10 and as you approach perfection, the improvements are non-linear. OK, that got/is getting way nerdier than I intended. My point is that in this type of comparison, when using this type of approach, you'd be better served to estimate their entire impacts and compare them than to do the arbitrary values for offense and defense.
2)
Portability. As someone (Owly? Westside?) pointed out up-thread, impact isn't universal. We can only quantitatively compare based on the situation that was to the extent of the numbers we have. However, each of us can, in forming our opinions, decide what we believe would have happened in different contexts. The idea of portability has played a big part in our last few projects like this. And while I've noticed that in general the offensive player is given the benefit of the doubt over an equivalent defensive player, one area in which I think the defender is better served is in the portability of their maximum impact. And ironically, the best way I can illustrate that is using a similar net points approach to the one I just discouraged you from using in (1).
Let's say that Jordan's and Russell's overall impact in their actual situations was a 10 each. Both of those 10s had elements of offense and defense in them, but in Jordan's case a large part of that value was offensive whereas the majority of Russell's impact was on defense. What would have happened, then, if we put these players on other teams with other situations. Obviously this is a thought exercise, but it's one that I've seen ElGee perform several times and it's useful to me.
While there are many different types of team situations that one could get into, ElGee has argued that it is less valuable what one could do in a bad situation than what one could add to a good situation. If we use that approach, then let's look at what we think would happen if Jordan and Russell were transplanted to other good teams that weren't built like their own. Specifically, what if they went into situations where there was redundancy of talent?
To me, there are a larger range of teams where Russell could have maximum impact than there are teams where Jordan could. Because defense is, in general, more additive than offense. If you add an elite offensive player to an already strong offensive team, the star (in this case Jordan) can only improve them by so much. And if the improvement comes by making himself the focal point, there is a good chance that this gain comes at the expense of any other player that relied on on-ball usage. If, on the other hand, Russell is added to a team that is already strong on defense, he would still be able to come closer to maximizing his own impact without it detracting from his fellows.
I'll illustrate with an example (note: this is not intended to be a rigorous proof. Just a real-life situation that is close enough to my point that it helps make that point clearer. If you understand my point but don't agree, nitpicking the examples isn't really something I'd find all that useful).
a) LeBron and Wade. We've seen over the last couple of seasons that Wade has worn down physically, and that can't be blamed on LeBron. In fact, you could even argue that LeBron has helped Wade sustain his level for longer. However, when both were at their best, they just could not be maximized at the same time. There was too much overlap in their offensive styles, both could perform well off the ball but could have their max impact with the ball in their hands.
b) Duncan and Robinson. Duncan and Robinson had a similar overlapping skill set to LeBron and Wade. Similarly, Robinson's health deteriorated over their tenure together. However, because Robinson was able to focus his impact on defense, he continued to maintain elite per-minute impact all the way up until he retired even while playing next to such a similar player.
Again, my feeling is that in general, elite defense is more additive and portable than elite offense. So even if I stipulate that Jordan and Russell may have both been 10 impacts in their actual situation, I believe that Russell could have maintained that 10 impact on his team's bottom line in more situations than Jordan could. In short, I think that Russell was more portable, which helps make him more valuable.
