A reader, express34texas, had some good questions regarding PCS with a lot of valid points. They were in the comments section of the “Full Explanation” post, but I feel like they deserve their own post in case anyone else had similar questions. Here’s his/her questions in bold, followed by my answers:
Thank you for explaining this even more, and seems to make more sense now evaluating players’ careers overall and not necessarily who’s better, but I think they go hand-in-hand. However, I see additional problems with it, regardless if the conclusions are fairly close to reality, though it’s quite interesting.
1. I’m not sure how to incorporate height as a measure, but this is a great example how it’s impossible to rate players like this and think it’s reality or even close to reality. There’s so many other variables. Size(height) matters a lot in most sports, especially basketball. I don’t think Iverson would ever be as good as Kobe or Jordan if he was 6 inches taller, but if he was, he’d be a lot better of player. His quickness would decrease some, but his improvement most other areas would improve a lot more. Just think if KD was only 6-6 instead 7-0, how much worse he would be. The examples are endless. Sure, tiny players can be good players, but the taller you are generally the better you are.
Size matters a lot in sports, specifically height for basketball (though Shaq’s weight mattered quite a bit too.) BBR is almost certainly right that it correlates strongly with how likely a player is to get into the Hall of Fame. Taller players are more likely to have better careers and are therefore more likely to get inducted. I just didn’t want to use it as a specific part of the formula. It feels built in to awards/win shares. Until very recently, almost all offenses were built around big men, so historically, big men have better stats which leads to more MVP votes and win shares.
2. I thought 3rd team all-nba selections were counted, as you mentioned in another post. (Oh, I see it mentioned later, but not initially). I don’t agree 2nd team all-nba should count the same as 1st team though.
Third team selections are not counted as they were only introduced in 1989. It’s the same reason I don’t use All-Defensive selections, Defensive Player of the Year, Most Improved Player, etc. I just wanted to use awards that go all the way back to at least the mid-50s. That works for All-NBA First and Second Teams, and All-Star Teams. There is a problem in that 1956 is the first year the MVP was awarded, but that was just too important an award to leave out. I thought about trying to work out who the MVP should’ve been those years (I still might, haven’t ruled it out yet, but I would need a really strong way to determine it down to say, fifth place.) But as this only affects 5-10 players – George Mikan, Paul Arizin, Dolph Schayes, and Neil Johnston, maybe Ed Macauley, Vern Mikkelsen, and Harry Gallatin – I thought it fair enough to start at 1956 rather than retroactively awarding the MVPs (a future project if ever there was one.)
(All-NBA First and Second Teams are also not the same. 15 points for a First Team, 10 points for a Second Team. It’s just that either one still counts as a “full credit” triggering award.)
3. Are defensive team selections not counted, doesn’t seem to be? I understand that you need to be consistent with all the eras, and this presents a big challenge, but this still needs to count. More on this in #4.
See answer #2. Not something I’m entirely opposed to doing – I would just want a solid way of retroactively designating those award winners prior to 1969. They probably should count, and probably will at some point one way or another, I just haven’t figured out the best way to go about it yet. I may give in and just add them at something like 5/2.5 points a piece, but I still don’t like that this will ultimately skew away from guys like Russell and West who are really well-represented as it is now.
4. While there were only 2 rounds in the playoffs early on, we just can’t eliminate the 1st/2nd rounds that have been going on for a long time now. Take RW for example, who’s the best player in the nba this season. He’ll be lucky to make the 2nd round, and he won’t make the WCF. Which presents other issues with this, as we have great players on weak teams relative to the top teams anyway, and this happens every year. RW will be missing out a lot.
Playoff performance prior to the Conference Finals is tricky. I’ve thought about approaching this two (well, three) different ways. First, I could continue halving the coefficients, i.e.: 1.25 points per win share for a second round exit, 0.625 points for a first round exit. Second, giving just one point (or half a point) for each win share through the first two rounds, so it wouldn’t skew toward modern players as much. Or three, leaving it as is with no points awarded for those two rounds. I’m leaning towards #3 because it already skews towards modern players: the modern playoffs are considerable longer – extra rounds means extra games, and extra games means more opportunities to snag win shares. It’s a tricky issue, and I need to play with it more to see how it affects things, but that’s how it stands right now.
5. Any use of win shares isn’t a good idea really, which I know we disagree on. But, you’re relying on this way too much. The problem is how to distinguish each player then if you don’t use a stat like WS. You really can’t. But if you look at each season, you can probably see huge problems with WS. Even if it’s fairly accurate, you’re incorporating another advanced stat not just for one of your criteria but several of your criteria into your own advanced stat.
I don’t think I’ll be able to convince anyone who doesn’t like win shares that they should put any stock in them. They usually correlate with who you’d imagine is the best player on the team, but they don’t always do that. I think as long as the method is being applied consistently to all players, there is room for some surprises, a la Pau Gasol getting more win shares in 2009-2010 than Kobe Bryant. (Although I watched every game of those playoffs – I’ll admit to being a Lakers fan here – and I thought Gasol had his moments of being the best guy on the court. I definitely think he should have won the Finals MVP in 2010.) And I could have come up with a points/rebounds/assists benchmark instead of win shares for the baseline of measuring longevity and regular season success. But I didn’t like how much that rewarded big stat guys on bad teams and didn’t reward the third or fourth best players on really good teams. Win shares just seemed to more accurately fit the bill for what I was looking for.
6. I don’t understand the credits fully. Why would you give 1/2 credit for not hitting a benchmark? However, if you need 3 AS to earn a full credit, 1 to earn 2/3(4/6) credit, then why is 2 AS only 3/4 credit? Shouldn’t it be 5/6 credit to stay accurate?
I admit this is weird, but the “full credit” caveat is there to prevent everyone who was on the 50s and 60s Celtics from having a higher score than they probably should. (Or any other dynasty team.) I wanted to give team leaders more credit than role players, so I set three All-Star games as the cutoff because 50.1% of players who make one All-Star team end up making three or more, so it seemed like a fair point to differentiate true star players from players who had one or two really good seasons. Initially, I was going to make both one and two selections the same at 75% credit, obviously halfway between the half credit of no All-Star selections and three of them. But I decided I wanted to give slightly more credit to guys who made two teams. I thought about doing the following: none All-Star selections = half credit, one = 66.7%, two = 83.3%, three or more = 100%. I could still adjust it to that, which you’re right, might make more sense. But I still wanted that midpoint between 50 and 100 to be incorporated. It isn’t a math/statistical reason, I’m sad to say, just what felt reasonable to me. It’s probably worth looking at again. (This is again subjective, I know, but bear with me.)
The highest rated player this affects is Terry Porter at at 101.8. He’d go to 113.1 if we set the two All-Star coefficient to 83.3%. It isn’t enough of a difference that it’s likely to affect someone’s chances to get into “my” Hall, so I haven’t worried about it too much yet.
7. Putting the ABA at 40% is way too low. This might affect only Dr. J as far as elite players go, but it’ll affect fringe HOFers a lot. And Dr. J was at his best in the ABA, too. Yes, his stats were inflated some, but you could be saying the same thing about his whole era compared today or different eras. The league changes. Offense is picking up quite a bit lately. Defense was huge in the early 2000s.
It’s interesting you use Swen Nater as your example. After looking over his career, he seemed better in the nba. He made AS team in 74 and 75 and all-aba 2nd teams both years. He goes from averaging 15 and 16 on .542 shooting in 1975 to averaging 10 and 10 on . 492 shooting in 1976(his last season in the ABA). He then averages 13 and 12 on .528 shooting in 1977 in the nba. 3 seasons after that he leads in the nba in rebounding, which he did once in the ABA, too. If anything, he was better in the nba. Sure, more players to compete against so harder to get awards, but no way knowing how close he was to 2nd team though. I bet he was very close at least once. And while Erving’s stats decreased some in the nba, Gervin’s increased a lot. Sure, Gervin matured more as a player, but that much? He made 0 1st-team all-aba, but 5 1st team all-nba.
The ABA is tough to quantify. For that percentage, I looked at the number of players and award winners in the ABA and compared them to their equivalents in the NBA. For example, there were seven players to win the ABA MVP, one went on to win the NBA MVP, or 14% of them (Julius Erving) – to me this means 14% guys who were the very best in the ABA were capable of being the very best in the NBA. The same goes for the other equivalent awards as you work your way down. For the All-League Teams, 42 players made an All-ABA team, and eight of those guys eventually made an NBA team – 19% (Rick Barry, Billy Cunningham, Erving, George Gervin, Connie Hawkins, Spencer Haywood, George McGinnis, and David Thompson). With All-Stars, 16 of 92 ABA All-Stars made an NBA All-Star team – 17% (Barry, Zelmo Beaty, Cunningham, Erving, Gervin, Artis Gilmore, Cliff Hagan, Hawkins, Haywood, Dan Issel, Bobby Jones, Maurice Lucas, Moses Malone, McGinnis, Charlie Scott, and Thompson; this was tougher to track down, let me know if I missed any). Comparing the All-Time teams (and the NBA’s from 1996 is badly outdated by now) five of the 30 members of the All-Time ABA team (17%) made the NBA 50 greatest players list – not a perfect analogue, but close. Finally, there were 518 different players in ABA history; 207 of them played in at least one NBA game (40%).
I think Nater and Gervin are outliers, and I probably shouldn’t have used Nater as the example. If you look at the majority of the ABA, just as a few examples: Barry, Cunningham, Gilmore, Hawkins, McGinnis, Scott, James Silas, and especially Erving, all of their per game and advanced statistics for their ABA years are higher than their NBA years, and usually significantly so. I realize it’s hard to separate out which guys were in their primes in the ABA vs. their NBA years – that might be another project down the road since the ABA is truly fascinating. I kept the value at 40% since that’s the number of ABA guys who made it in the NBA and to give the benefit of the doubt, but the translation at the highest level is much lower, around 17-18%. The 40% honestly feels too high to me, and I’m thinking about bumping it down to something like 25%.
The ABA is worth considering on its own merits when debating Hall of Fame worthiness – I’d consider guys like Mel Daniels, Dan Issel, and McGinnis as Hall-worthy based on their ABA careers alone. (It’s a shame McGinnis isn’t in already.) But giving the awards a fraction of their counterparts seems like the most reasonable way to incorporate them into the score.
8. PCS is an objective stat only in the sense that every player is rated with the same formula. However, all the components that make it up are subjective(either someone voting or the person who made up WS), just like any other advanced stat that specifically rates players. All of your criteria are subjective, though not by you. However, the way you weight each criteria you deem important is subjective according to you. I understand your disclaimer what you’re saying, but I think you’re getting too caught up in it. This is very subjective. Just take the 40% ABA part for example. Even if we say you’re right about that for argument’s sake, there’s nothing objective about that.
I’m not sure the best way to explain this. I realize subjectivity isn’t something I’ll completely get away from. You’re right, the weighting system is subjective – but it is also objectively applied to every player equally, which is what I meant. (I should have clarified.) On subjectivity as a whole, I don’t think it will ever be done away with entirely when judging players, nor would I really want that. It’s fun and interesting to argue whether you’d rather have Bird or Magic, or Wilt or Russell, or anything like that. There’s just far too much context to boil everything down to a single number. I mean, you can’t quantify style no matter how hard you try. And there’s still an in-their-prime argument to be made whether you use numbers there or not.
I guess I’d say I’m not trying to reduce a player to a single number, I’m trying to stack careers against one another as a guideline for looking at the Hall of Fame. It might not seem like much of a distinction to others, but it’s an important one to me. If I wanted a Hall of Fame based solely on numbers, I’d just set a cutoff (like 200) and say everyone with more is in, no exceptions, everyone with less is out, no exceptions. But I want people to vote and get discussions and debates going. Do I think Bernard King should be in despite being lower than 200? Absolutely. Do I think Chris Bosh deserves to be in even though he’s slightly over 200? I’m not real sure about that one; it might be recency bias, but right now I’d say no. PCS shouldn’t be the sole determiner of Hall of Fame-worthiness, just one facet of it for looking at careers as a whole.
If anyone has any other questions, comments, or feedback, please let me know and I’ll get back to you and answer them as best I can.