Can anyone shed light on why the sim is stingy with D ratings for earlier seasons? There are only 15 player seasons with a 90 D rating throughout the 1960s, and none in the 1950s.

Okay, the league was smaller so there is a smaller pool of players, but that can't account fully for the discrepancy. By contrast, there are 100 players seasons at 90+ D in the '70s, and 123 in the '80s.

Is this a matter of a subjective determination that defense was less of a focus, or defenders were simply less skilled? Or is there a stat-based algorithm at work here?

1st team defense equals an automatic 90 rating.
2nd team defense equals an automatic 70 rating.


All defense wasn’t a thing until 1969.

hence the Bill Russell travesty

(also those stats that do help to determine some part of the D rating (steals and blocks) were not kept until the mid 70s so when you look at player stat tables from seasons prior to that what you're seeing is made up nonsense)

Well, I know this is a pretty humorous topic around here, but in reality the early NBA (post-shot clock) was all about shooting as quickly as possible and that's why all the FG% from that era look so bad while rebound averages look so great. The flip side of this is that legitimately good defenders kind of get understated by an algorithm looking for actual data. Mostly why I don't mind the Bill Russell low defense numbers. As far as the sim is concerned, his defensive prowess is anecdotal. Everyone from that era agrees that he's great, but we aren't totally sure how destructive his defense was in his era compared to the Ben Wallace, Tim Duncan, or even Al Horfords of the past 15 years.

interesting post from a user at bball forums

I think both Robinson and Wilt have a case here. Wilt might have peaked higher defensively (largely due to era differences IMO, much easier to impact the game defensively when there's no 3 point line and thus less offensive spacing), but he was a little less consistent than Robinson.

As you guys might have noticed my approach in these threads has been entirely impact-based, i.e. how much a player's presence actually improves his team. Best way to look at this IMO is to simply look at the results - examine team DRtg (points allowed per 100 possessions) relative to league average with and without the player in question over their careers and analyze trends.

Negative relative DRtgs are good, positives are bad. e.g. If the league average in a particular year is 100 points allowed per 100 possessions, and a team allows 95 points per 100 possessions, their relative DRtg that year is -5

Warriors' Defensive Rating relative to league average:
1959: -1.3
1960: -4.5 *Wilt joins
1961: -1.5
1962: -0.4
1963: +1.7
1964: -5.8

Wilt has an immediate -3 impact his rookie year. From 61 to 63 we see the defense suffer, coinciding with the huge offensive load that Wilt has to carry. Then in 64, by most contemporary accounts this is the first year Wilt puts tremendous effort on the defensive end (emphasized by new coach Hannum) and the numbers reflect that - one of his best defensive years at arguably around -5 impact.

1965: -0.7 *Wilt's last season w/ Warriors, plays only 38 games
1966: -0.2

An expected drop in 65 given he plays only 38 games. Without him in 66 they drop a little further.

Sixers' DRtgs:
1964: +2.5
1965: +0.4 *Wilt joins, plays only 35 games
1966: -3.4 *Wilt plays full season

In 65 Wilt improves the defense by -2 playing only half the season. When he plays the full season in 66 he improves them by -6 compared to 64. Further evidence that Wilt can be at least a -5 defensive impact player (when he's not focusing as much on offense, as his scoring is starting to dip here).

1967: -2.4
1968: -5.6 *Wilt's last season w/ Sixers
1969: -1.6

The defense stays fairly consistently great. I think we can say he's still having around -4 to -5 impact in 67 and 68. Indeed, in 1969 when he's gone we see the defense get worse by 4 points.

Lakers DRtgs:
1968: +0.5
1969: -0.1 *Wilt joins
1970: -1.5 *Wilt plays only 12 games
1971: -1.4
1972: -4.4
1973: -3.7 *Wilt's last season
1974: -1.1

Less impressive changes on defense here from 69 to 71. Perhaps harder to impact the defense of the much more offensively-oriented Lakers here (they led the league in Offensive Rating). Not to mention this is Wilt past his prime.

Notice the -3 change from 1971 to 1972, however. This coincides with Wilt's offense dropping from 15 FGA / 21 PPG in 1971, to 9 FGA / 15 PPG in 1972. When he's focusing less on offense, the defense improves greatly. We saw that on his previous teams and we see that here. Coincidence?

Then in 1974 without Wilt the defense falls by 2.6 points.

This is pretty clear evidence that Wilt, when focused on defense rather than offense, is a -5 DRtg impact player, if not higher, during his best defensive seasons. Outside of his best defensive years, he's around -3 to -4 impact which is still excellent. And in years in which he focuses more on offense (e.g. 61-63, 69-71), the defense suffers. I don't think these are coincidences.

A similar analysis of Robinson shows he was a -4 to -4.5 DRtg impact player for almost his entire career. I won't do another lengthy breakdown but you can see right from the start of his career that he improves the Spurs from +0.1 to -3.9 his rookie year (-4 impact), and he stays around that kind of impact for pretty much his entire career, even alongside Duncan. Defensive metrics support Robinson having similar impact to Duncan when he's on the court from 98 to 03.

An analysis of Pippen, at his absolute peak, shows he approaches -3 impact (arguably -3.5 in 1995, the best defensive peak of any perimeter player ever), which is absolutely HUGE for a perimeter player but that's about as high a perimeter can impact the game. If you're doing an honest assessment of these players' defensive impacts, I think that's pretty clear. Elite defensive bigs simply have more impact than the best perimeter defenders due to the nature of their positions.

If more people took a more fact-based approach rather than a reputation or highlight video approach, I don't think we'd see Pippen getting voted this high IMO.

so you could actually do something along these lines for Russell as well - as a matter of fact the Cs Drtg goes from 89 (1st in the league) in 68-69 (Russell's last season) to 99 and 8th in the league (of only 14 teams at the time) the next season (and they go from winning a title to not making the playoffs) - there are similar behaviors at the beginning of his career

That's a pretty fantastic analysis.

He is using rel-drtg from bbref. A look at Russell's Celts puts this thing to bed.

From 1.4 before Russell to -4.9 as a rookie.

Then
-5, -6, -6, -8, -8.5, -8.5, -11, -9, ,-7, -5, -4, -7

Back to 0 69-70.

These may be estimated numbers, but they align with anecdotal evidence that #6 would rank as the best defender of his era and probably of all time.

Could he match Ben, Drob and Duncan? Again, anecdotaly he could defend more positions than they did. The numbers seen to align with him having as big an impact as anyone ever has.

To have his wis d rating below 96-100 makes no sense.

Posted by superrobb420 on 5/12/2018 11:23:00 AM (view original):
Well, I know this is a pretty humorous topic around here, but in reality the early NBA (post-shot clock) was all about shooting as quickly as possible and that's why all the FG% from that era look so bad while rebound averages look so great. The flip side of this is that legitimately good defenders kind of get understated by an algorithm looking for actual data. Mostly why I don't mind the Bill Russell low defense numbers. As far as the sim is concerned, his defensive prowess is anecdotal. Everyone from that era agrees that he's great, but we aren't totally sure how destructive his defense was in his era compared to the Ben Wallace, Tim Duncan, or even Al Horfords of the past 15 years.

Sorry robb, but just watch clips of the guy play and you'll never again compare him to Al Horford (or anyone but the most elite defenders).
His SIM D ratings #s are a travesty.

oh look here's footage of Russell taking down a rebound, going coast to coast in 7 steps and jumping OVER a dude, you know, just like Al Horford....

https://www.youtube.com/watch?time_continue=4&v=SF22xiRxHv8

PS I always find it funny to watch old footage of guys dribbling the ball because they keep their hands on top of the ball and dont try any of the palming/carrying mess guys get away with now so it looks odd

Posted by copernicus on 5/13/2018 12:26:00 PM (view original):
PS I always find it funny to watch old footage of guys dribbling the ball because they keep their hands on top of the ball and dont try any of the palming/carrying mess guys get away with now so it looks odd

60's and 70's basketball will also look "strange" to younger people because players really were only allowed 2 steps instead of 3 or 4, teams actually worked the ball around to get shots closer to the basket instead of jacking-up half their shots from 23 feet or more, and guys weren't allowed crab dribbles, Euro-steps and other crap.

the other thing is that the dude's analysis was done a few years ago and his observations about the role of big vs perimeter might be a little dated - a roving 4 like Draymond (or in his post Pippen) may be more impactful now given the change in game play

very good discussion this turned out to be

also Russell, who, if you look at the footage, was an absolute gazelle, could totally defend all over the floor - he'd be a freaking terror

Posted by copernicus on 5/14/2018 12:38:00 PM (view original):
also Russell, who, if you look at the footage, was an absolute gazelle, could totally defend all over the floor - he'd be a freaking terror

Indeed. And in today's game it's reasonable to suppose he'd be stronger (due to modern condition regimens) and be a more efficient scorer.