Retro Player of the Year Project

[Doctor MJ]

Curious to hear your thoughts on Win Shares and especially D-Win shares in general. I've never been impressed by it as a stat, seems very much for single players trying to individualize something that can't be easily individualized, and for both players and teams quantifying something that can't really be quantified. I dunno, I've looked at some of the formulas and don't really find them terribly impressive, but maybe I'm wrong.

Also, my understanding is that individual Drtg is something akin to raw +/- in that it does not adjust for teammates, correct? If so, that makes it even less close to the mark. Maybe this is another discussion entirely, for the Stats board...

Well, I like WS, though I wouldn't want to put too much stock in it. It's got that classic all-in-one pro/con where it sums up a lot in more number, and thus is a good snapshot, but it also combines very different things, so it's probably better to have all those different numbers handy. Still, I like those snapshots, and there was dire need for really anything that measures total impact (as opposed to per possession played impact).

There's also the matter that on defense, it seems bound to underrate guys who don't rack up box score stats - that said, for guys like that on good defensive teams, they're getting more credit by WS than anywhere else (save +/-).

One of the reasons it's getting used so much in the past, is that it's not so easy to get those numbers separately way back in history. This is unfortunate on a number of levels, one of which being that WS are far less accurate back then due to the stats not recorded. However for that issue, I'm just always making some adjustments in my head. For example, Walt Frazier did really well on WS, despite the fact he almost certainly had steals beyond what WS estimates, so there's reason to believe he was actually underrated by WS - so I've got no majors issues using WS for someone like him.

Re: DRtg. Having trouble thinking this through at the moment. I'd tend to think of DWS as largely about converting DRtg into a format where total accumulated impact is measured instead of per possession impact, and then normalizing it so that it could be put in terms of wins. So neither is really more off the mark than the other, just depends on what you're trying to ascertain. Others thoughts?

[bastillon]

semi,

DRtg is irrelevant because it doesn't give us any idea of their defense without putting it into context. if you wanna use DRtg, you'd have to adjust it to league average.

or you can just use team DWS, it's basically the same thing and translates this impact into wins, too. combined DWS is just a way better metric to rank team defense. it's not more accurate, but it's more useful because you don't have to put it into context. your context is already there.

Wilt

Code: Select all

    year       combined DWS
    60             31.3
    61             23.8
    62             22.1
    63             17.1
    64             37.3
    66             28.5
    67             22.5
    68             33.6
    69             19.5
    71             17.2
    72             31.3
    73             30.0

put it simply: Wilt's peak defensively didn't approach Russell's average year (Celtics consistently over 40 DWS). Wilt NEVER EVER anchored similar defenses as Russell's best teams.

Code: Select all

    Year       Teammates
    ========================================================
    1968-69    S Jones, Havlicek, Sanders
    1967-68    S Jones, KC Jones, Havlicek, Sanders
    1966-67    S Jones, KC Jones, Havlicek, Sanders
    1965-66    S Jones, KC Jones, Havlicek, Sanders
    1964-65    S Jones, KC Jones, Havlicek, Sanders
    1963-64    S Jones, KC Jones, Havlicek, Sanders, Ramsey
    1962-63    S Jones, KC Jones, Havlicek, Sanders, Ramsey
    1961-62    S Jones, KC Jones, Sanders, Ramsey
    1960-61    S Jones, KC Jones, Sanders, Ramsey, Sharman
    1959-60    S Jones, KC Jones, Ramsey, Sharman

well, Sam Jones was a mediocre-to-bad defender so what you really have is three outstanding perimeter defenders playing 25-30 MPG - is that supposed to be some all-time epic support ?

more importantly, we know Russell's impact - it was about 20 DWS. there were multiple occasions when his teammates changed and the difference in DWS was minor, to say the least. what really stood out is the consistency with Russ... and how they fell of a cliff immediately after he retired.

I'm not buying this great supporting cast argument. Russell's teammates weren't a lot above average for his era. the most important position from Russell's perspective was PF and most of the time guys like Heinsohn or Bailey Howell, really bad defenders, played that position. pre-Russ, post-Russ, olympics, college, didn't matter. Russell was on the floor - team defense was all-time epic. Red or not, Hondo or not, KC or not. the only constant was always Russell.

[semi-sentient]

penbeast stated that Sam Jones was average, so I included him. Removing him doesn't exactly make that much of a difference.

I never stated that Russell had all-time epic support either, only that he had more help than Wilt. If you can't admit that then there's no point in continuing. Yes, he was the only constant and he's a better defender than Wilt, but when looking at just how dominate the Celtics were you have to look at all the other factors.

RE: falling off a cliff -- that's not necessarily true. They were still one of the better defensive teams in the league (above average), and a big drop is to be expected when replace a guy like Russell with Hank Finkel and Jim Barnes.

That's basically like replacing Dwight Howard with Travis Knight and Mark Madsen. Tom Sanders also missed 25 games that year.

Now in the 1970-71 season, they added Cowens to the mix and saw a nice yearly improvement until got back to playing high level defense (1972-73) similar to how they were playing in the mid/late 60's so...

[TrueLAfan]

I’ve largely stayed out of the discussion of statistical analysis of pre 1969 teams, especially on defense, for a couple of reasons. First, I don’t think the tools are necessarily accurate when we have thorough numbers, as we do today. I like WS and DWS; I think they do what they do about as well as any other metric. But that’s not saying a huge amount. I think the idea that Lamar Odom was one of the top 7 defenders in the league in 2009-10 is absurd; that Chris Bosh was 142nd in league in 2010 in DWS is equally off. I think these types of metrics can give us general information…but that’s all it is.

When we’re going back into periods where we don’t have complete statistical breakdowns, it makes an already semi-questionable tool far more problematic. I can see team breakdowns being somewhat effectively determined; I see much bigger problems for individuals.

And this leaving out that defense was conducted in a different manner in the past. Players were allowed to do certain they aren’t allowed to do now. Stylistic shifts in offense dictated different approaches to defense. Even if we had complete numbers, they wouldn’t necessarily show us anything. We’d be looking at an analytic tool designed for 21st century basketball to judge a game that is played in an often dissimilar way. (A good comparison would be the stolen base in baseball … it was used extensively in some periods, and not as extensively in others. If you choose a metric that puts more—or less—value on the stolen base, you’re going to be rating a team based on a system that “rates” a team, or player, by the value you want…not the value at the time.) So using that tool as anything other than a (very) minor supporting argument for eyewitness accounts is not useful. It is easy to say that first hand accounts are biased and unclear. Sometimes they are. But they also are contemporaneous and, when generally agreed upon, should be given more weight than any statistical tool. Much more, in my opinion.

I appreciate the statistical work done by others on this project, and I find it interesting and useful. But I absolutely, positively don’t think it should be used as anything other than support for what was commonly viewed and recognized and noted. These analyses lack the integrity to stand on their own as measurement tools.

[bastillon]

20 DWS less IS the very definition of falling of a cliff. it doesn't matter what place they were. it matters how much they won with their defense. 69 Celtics won over 40 games with their defense. that means they were on the level of 2008 Celtics defensively. 70 Celtics were average team defensively. now whatever the difference between 2008 Celtics and average is for you, it was the difference between Russell-led team and Russell-less team. huge meltdown.

but you have a great point about the 70s Celtics. in 73 they had 45.3 DWS - that's better than 90s Knicks, KG's Celtics and any version of 70s Knicks. one of the best defensive teams of all-time... Havlicek, Cowens, Chaney, JoJo White, Silas... no wonder why.

[bastillon]

I’ve largely stayed out of the discussion of statistical analysis of pre 1969 teams, especially on defense, for a couple of reasons. First, I don’t think the tools are necessarily accurate when we have thorough numbers, as we do today. I like WS and DWS; I think they do what they do about as well as any other metric. But that’s not saying a huge amount. I think the idea that Lamar Odom was one of the top 7 defenders in the league in 2009-10 is absurd; that Chris Bosh was 142nd in league in 2010 in DWS is equally off. I think these types of metrics can give us general information…but that’s all it is.

When we’re going back into periods where we don’t have complete statistical breakdowns, it makes an already semi-questionable tool far more problematic. I can see team breakdowns being somewhat effectively determined; I see much bigger problems for individuals.

And this leaving out that defense was conducted in a different manner in the past. Players were allowed to do certain they aren’t allowed to do now. Stylistic shifts in offense dictated different approaches to defense. Even if we had complete numbers, they wouldn’t necessarily show us anything. We’d be looking at an analytic tool designed for 21st century basketball to judge a game that is played in an often dissimilar way. (A good comparison would be the stolen base in baseball … it was used extensively in some periods, and not as extensively in others. If you choose a metric that puts more—or less—value on the stolen base, you’re going to be rating a team based on a system that “rates” a team, or player, by the value you want…not the value at the time.) So using that tool as anything other than a (very) minor supporting argument for eyewitness accounts is not useful. It is easy to say that first hand accounts are biased and unclear. Sometimes they are. But they also are contemporaneous and, when generally agreed upon, should be given more weight than any statistical tool. Much more, in my opinion.

I appreciate the statistical work done by others on this project, and I find it interesting and useful. But I absolutely, positively don’t think it should be used as anything other than support for what was commonly viewed and recognized and noted. These analyses lack the integrity to stand on their own as measurement tools.

that's the problem with your line of thought: we're not even touching individual DWS, because it's useless. team DWS is basically DRtg relative to league average translated into wins.

I also think individual DWS is nonsense. it's trying to sort of share the team DWS between the players using mins, rebs, stls and blks. so IIRC Ray Allen was negative DWS player in 2007 but was one of the best in the league in 2008 because he played with Garnett.

DWS as any boxscore-to-wins based metric does a great job of asigning wins to team, but is poor in sharing the credit among the players.

it's great on team level, performs poorly on individual level.

[lorak]

Elgee,
could you post your ORtg and DRtg numbers for 1971, 1972 and 1973 seasons? Because these numbers who are in POY threads are incomplete (many teams are missing). Thanks in advance.

[Doctor MJ]

Guys, I'm planning now to go through '54-55. There seems to be a decent contingent either for this or '55-56, and those threads would overlap in our schedule anyway. This would mean we'd have 2 weeks left in the project, and we could talk about the results simply as shot-clock era basketball, which sounds good to me.

As I think about it, because of the overlaps, '56-57 is the other real option, which would make this the last week. I'll still listen to sharp protests for those who can't take 2 more weeks of this though.

[Sedale Threatt]

I'm down to just voting now. I have virtually nothing to say or add at this point.

[Dr Positivity]

I'm down to just voting now. I have virtually nothing to say or add at this point.

Me too - picking between Hagan, Schayes, Yardley etc. is just throwing darts at the wall. Luckily I don't really care whether Hagan finishes 75th or 63rd... the only ones that matter to me right now are Russ and Pettit and I know enough about them to vote confidently, so anyways

[drza]

I'm down to just voting now. I have virtually nothing to say or add at this point.

Me too - picking between Hagan, Schayes, Yardley etc. is just throwing darts at the wall. Luckily I don't really care whether Hagan finishes 75th or 63rd... the only ones that matter to me right now are Russ and Pettit and I know enough about them to vote confidently, so anyways

Ditto on the feeling of throwing darts, but surprisingly to me, I have the same feeling thus far for Pettit. Coming into this project I knew he was thought to be great, perhaps the GOAT pre-Malone PF, and I was looking forward to really learning about him. So far, though, I haven't really seen what the fuss is about. I mean, he's clearly a great player, but so far he hasn't even differentiated himself from Hagan on his own team, let alone shown himself to be the juggernaut that I had hoped to find. And I still have no feel for his game...several years into his run, I still see him only as a collection of numbers.

[bastillon]

drza, Pettit was a very good outside shooting PF/C who also happened to be a great rebounder and could draw fouls A LOT. poor defender though: white, somewhat unathletic, not long enough to contest shots and thus no shotblocking presence.

I think, as in the case of all poor defensive bigs, his numbers overrate his impact. he's not as awful as, say, Lucas but that's enough to disqualify him from any GOAT PF discussions for me.

[Doctor MJ]

drza, Pettit was a very good outside shooting PF/C who also happened to be a great rebounder and could draw fouls A LOT. poor defender though: white, somewhat unathletic, not long enough to contest shots and thus no shotblocking presence.

I think, as in the case of all poor defensive bigs, his numbers overrate his impact. he's not as awful as, say, Lucas but that's enough to disqualify him from any GOAT PF discussions for me.

Curious on your thoughts of Baylor?

I think people are too quick to dismiss Pettit based around his "unathleticism". He seems to have clearly been at least the equal of Baylor as a forward, and Baylor was incredibly athletic.

[bastillon]

after this project I'm much more pro-West than Baylor. he didn't seem to have great effect on offenses when he was their anchor. above average defender for his times, but would struggle in modern game with his tweener body (as did Barkley). I think of him as Barkley with -10 TS%. I don't regard Barkley as highly as others so you can guess I'm not a huge Baylor fan.

[penbeast0]

Bastillon, I've never read anything that said Pettit was a poor defender, unlike, say, Tommy Heinsohn. Thus I've always assumed average defender with great defensive rebounding skills; maybe better than that since Heinsohn said he was the most relentless player Heinsohn had ever faced. Nor have I heard contemporaries calling Pettit unathletic (skinny, yes, particularly coming into the league when they were playing him at center where he felt he was getting beat up).

I have heard Hagan was a good defender; also that despite his off the court demeanor that Hagan was an incredible thug on the court (unlike Pettit who was supposedly a gentleman on the court -- which may be a negative defensively though I've heard it about good defenders).

[Doctor MJ]

after this project I'm much more pro-West than Baylor. he didn't seem to have great effect on offenses when he was their anchor. above average defender for his times, but would struggle in modern game with his tweener body (as did Barkley). I think of him as Barkley with -10 TS%. I don't regard Barkley as highly as others so you can guess I'm not a huge Baylor fan.

Well, I think everyone's a bigger fan of West than Baylor - but that simply is part of recognizing that Baylor was only the 5th best guy in the 60s. I guess my point is that you're tearing Pettit up for the things he lacked, but to the extent that those things existed, they were evident in his day, and he was still considered to have tremendous overall impact in the conversation with the 60s superstars of greater athleticism.

[ThaRegul8r]

Ditto on the feeling of throwing darts, but surprisingly to me, I have the same feeling thus far for Pettit. Coming into this project I knew he was thought to be great, perhaps the GOAT pre-Malone PF, and I was looking forward to really learning about him. So far, though, I haven't really seen what the fuss is about. I mean, he's clearly a great player, but so far he hasn't even differentiated himself from Hagan on his own team, let alone shown himself to be the juggernaut that I had hoped to find. And I still have no feel for his game...several years into his run, I still see him only as a collection of numbers.

I wish I had more time, but there were so many people I was doing research on, but not enough time, particularly with this project. In the future, I might perhaps make posts providing more information on players people are more interested in learning about, so that it might be of some use even after we've concluded this project. For edification, or even the next time RealGM does an All-Time League, so that an owner of a team isn't penalized based on a general lack of knowledge about a certain player(s) they may have drafted, which might cost them a matchup.

[bastillon]

Bastillon, I've never read anything that said Pettit was a poor defender, unlike, say, Tommy Heinsohn. Thus I've always assumed average defender with great defensive rebounding skills; maybe better than that since Heinsohn said he was the most relentless player Heinsohn had ever faced. Nor have I heard contemporaries calling Pettit unathletic (skinny, yes, particularly coming into the league when they were playing him at center where he felt he was getting beat up).

I have heard Hagan was a good defender; also that despite his off the court demeanor that Hagan was an incredible thug on the court (unlike Pettit who was supposedly a gentleman on the court -- which may be a negative defensively though I've heard it about good defenders).

I don't see how white 6'9 big can be a good defender. maybe I'm generalizing too much, but you know...

[Doctor MJ]

I don't see how white 6'9 big can be a good defender. maybe I'm generalizing too much, but you know...

So quite literally, you're downgrading this guy in your mind just because he's white.

I'll point out two things just cause:

1) Andrei Kirilenko

2) Pettit was still racking up the rebounds in his 30s as it became par for the course for teams to have guys bigger than him. Now rebounding alone doesn't mean great defense, but this was hardly a guy who couldn't mix it up on the inside.

[semi-sentient]

I heard a couple of other honkies were pretty good defenders back in the 60's as well. West and Hondo anyone?