RealGM

Yes, ceiling raiser/fpliii has another hot take based on Squared2020's limited data.

So recently, someone posted in the top peaks project updated 1991 data. Bird looks very much like a top 5 player in a limited sample.

Obviously you need a ton more data, but I almost wonder. What if Bird from 90-92 was still elite?

A good comp would be David Robinson after his injury. He took a step back largely because of Duncan's presence. In Bird's case his usage declined with the rise of Reggie Lewis.

These three seasons are usually written off. But if we had data showing his impact was elite, how differently would you view Bird's career?

It's not too out there for Bird to still have top 5 impact across those 3 years on average. He was 12th, 15th and 9th in BPM so he didn't just get those All-Star selection for the name value. Considering some players ahead of him in some of those years have dip years in this period, a top 10 impact across this period is pretty believable, top 5 might be a bit high but not impossible.

Though it doesn't account for these years still being well below his prime level, his defense getting way too much leniency, him playing significantly less games every year and lackluster play-off performances during these 3 seasons.

I think some are too dismissive of these years but even if properly rated they're not the type of seasons to heavily impact the legacy of a top 10 player.

Tough to tell without more sample details (19 games, 17-2 in them, unsure how many Bird played in), but overall Bostons players are likely slightly overrated by the small unbalanced sample.

Of course if it was true it'd be a nice little feather for Bird, putting him a clear half step ahead of Kobe/Dirk for me, probably still just behind Magic/Oscar. Not too much as there'd still be a fair sized *when healthy* asterisk and some not so great playoff performances to explain within those years.

*Edit: I see Justin put out an updated version, so my numbers above aren't currently accurate, but I'd still be concerned about the sample given the base and would need to see more. It's not unbelievable, they were really good with Bird and pretty average without him overall.

If we had full season data, raw on-off data (for comparison), a reliable known source doing a 3 year RAPM with Bird top 5 ...

So that allows a max of say MJ, Magic, Robinson (on/off impact-y stuff never locks but given status, and Robinson's precedents should each be likely) and one other above him.

That would be pretty big. Now I'd caveat with ... the last two years the minutes are significantly hit. We're looking at nearly 6900 minutes for the three years. He also has 2 weak production playoffs with [otoh] uncertain availability (even just early rounds).

Still the impression from what I read is he's a bad defender (poor mobility) at this point, which combined with reduced production has me pretty down on these years. So having him being net in or around that tier (and of course the details matter), especially if one isn't too mean about playoff health/play (I trend lower weight here but ... this is an area of concern, probably moreso as one advances deeper), would be a significant boost to my perception on the longevity of quality. Obviously better years still matter more but accepting the premise (not assessing plausibility) this is adding parts of 3 years at what would seem to be a very high level in absolute terms.

Maybe a bit higher on his peak/more lenient on his scoring issues in the playoffs.

ceiling raiser wrote:Yes, ceiling raiser/fpliii has another hot take based on Squared2020's limited data.

So recently, someone posted in the top peaks project updated 1991 data. Bird looks very much like a top 5 player in a limited sample.

Obviously you need a ton more data, but I almost wonder. What if Bird from 90-92 was still elite?

A good comp would be David Robinson after his injury. He took a step back largely because of Duncan's presence. In Bird's case his usage declined with the rise of Reggie Lewis.

These three seasons are usually written off. But if we had data showing his impact was elite, how differently would you view Bird's career?

I'm glad you emphasized that we need more data because I really don't think we should thinking too much about this from a player impact perspective yet. (On the other hand, if it can tell you something about a particular team's play and trends already, cool.)

But to entertain your hypothetical here:

Magic's extra years of longevity tend to make it real easy for me, and probably everyone else, to just give the edge to Magic over Bird. It's possible we'll find we really shouldn't be doing that.

My own impressions of Bird at the time was that he was still very effective and impactful in a general sense. Its just that his mobility had dropped off even more and he was really getting up there in age. He was still a great player I'd say at least in 90 & 91. I remember watching that final 50 pt game he had and he could still make incredible shots. I don't write these seasons off when I compare him to other atg's.

THe Celts played .689 ball (124-56) in games he played and .530 ball (35-31) the games he didn't.
That's 13 games a year in an 82 game schedule.
I didn't realize this was in question.

DQuinn1575 wrote:THe Celts played .689 ball (124-56) in games he played and .530 ball (35-31) the games he didn't.
That's 13 games a year in an 82 game schedule.
I didn't realize this was in question.

You don't think that him being top 5 in impact is in question?

Okay well
1) The absences skew towards older McHale, older Parish teams.
2) I would imagine based on what I've heard of Bird's absences that they skewed towards Boston back-to-backs.
3) I wouldn't think .159 is a massive, elite level difference in WoWY win % because
a) from a similar era and span (92 and 93) STATS Inc had
Jordan .611
Olajuwon .458
Barkley .392
Rodman .383
Mutombo .295
Daugherty .258
Robinson .250
Mullin .178
among noteworthy with at least 5 games missed (they have Bird and Wilkins tied at .148, Kemp .0366, Coleman -0.18).
Sampling of others with from the source data with +.2 seasons (excluding those with a negating, significantly worse, clearly into negative seasons e.g. '93 Bol hurt by 92 Bol, '92 Dreiling by '93 Dreiling). Obviously single season means smaller samples.
Michael Adams '93 .217
Barros '93 .249
Brickowski '92 ? ('93 barely above neutral) .329
Tom Copa .309
Daugherty '93 .341
Chris Dudley '93 .501
Sean Elliott '93 .212
Winston Garland '93 .212
Tate George '92? .279 (circa neutral with more balanced sample ... larger off ... in '93)
...
b) this type of impact where it's injury and bench guy becomes starter, non-rotation guy[s] added to rotation, should - I think - skew optimistic.

I feel like I must be misreading the "this" (perhaps actually intending "Bird was useful") but given the subject (and discussion) of the thread that would seem a non-sequitur.

ceiling raiser wrote:Yes, ceiling raiser/fpliii has another hot take based on Squared2020's limited data.

So recently, someone posted in the top peaks project updated 1991 data. Bird looks very much like a top 5 player in a limited sample.

Obviously you need a ton more data, but I almost wonder. What if Bird from 90-92 was still elite?

A good comp would be David Robinson after his injury. He took a step back largely because of Duncan's presence. In Bird's case his usage declined with the rise of Reggie Lewis.

These three seasons are usually written off. But if we had data showing his impact was elite, how differently would you view Bird's career?

Wait did u change ur name from fpliii lol

Owly wrote:If we had full season data, raw on-off data (for comparison), a reliable known source doing a 3 year RAPM with Bird top 5 ...

So that allows a max of say MJ, Magic, Robinson (on/off impact-y stuff never locks but given status, and Robinson's precedents should each be likely) and one other above him.

That would be pretty big. Now I'd caveat with ... the last two years the minutes are significantly hit. We're looking at nearly 6900 minutes for the three years. He also has 2 weak production playoffs with [otoh] uncertain availability (even just early rounds).

Still the impression from what I read is he's a bad defender (poor mobility) at this point, which combined with reduced production has me pretty down on these years. So having him being net in or around that tier (and of course the details matter), especially if one isn't too mean about playoff health/play (I trend lower weight here but ... this is an area of concern, probably moreso as one advances deeper), would be a significant boost to my perception on the longevity of quality. Obviously better years still matter more but accepting the premise (not assessing plausibility) this is adding parts of 3 years at what would seem to be a very high level in absolute terms.

Yes. The current dataset the person doing the research and data collection has 1985 Bird with an unimpressive number relative to The expectations of prime/peak Bird. Also, not sure how to take the data when it’s a single person doing it regarding veracity if number.

MyUniBroDavis wrote:

ceiling raiser wrote:Yes, ceiling raiser/fpliii has another hot take based on Squared2020's limited data.

So recently, someone posted in the top peaks project updated 1991 data. Bird looks very much like a top 5 player in a limited sample.

Obviously you need a ton more data, but I almost wonder. What if Bird from 90-92 was still elite?

A good comp would be David Robinson after his injury. He took a step back largely because of Duncan's presence. In Bird's case his usage declined with the rise of Reggie Lewis.

These three seasons are usually written off. But if we had data showing his impact was elite, how differently would you view Bird's career?

Wait did u change ur name from fpliii lol

Yeah, in December or so.

Owly wrote:2) I would imagine based on what I've heard of Bird's absences that they skewed towards Boston back-to-backs.

You heard wrong, Bird played 15 of team's 21 back-to-backs in 91 and 10 of 20 in 92. That's right in line with 60 and 45 games.

Owly wrote:3) I wouldn't think .159 is a massive, elite level difference in WoWY win % because
a) from a similar era and span (92 and 93) STATS Inc had
Jordan .611
Olajuwon .458
Barkley .392
Rodman .383
Mutombo .295
Daugherty .258
Robinson .250
Mullin .178

Jordan hardly missed any games, 2-10 games is just too small a sample.
So I used a 15 game cutoff for these guys 90-93.
I looked at top group and get the data below
total team wins,losses, played wins played losses, dnp wins and losses and difference
Bird is second and fourth of the group; behind Rodman and Barkley splitting him,
It's also harder for Bird as he and Hakeem are the only ones whose teams are above .500 without them, and it is harder to improve a plus .500 team than a negative one.


w l on w on l off w off l tot % on % off % diff %



93 Rodman 40 42 36 26 4 16 0.488 0.581 0.200 0.381
91 Bird 56 26 46 14 10 12 0.683 0.767 0.455 0.312
91 Barkley 44 38 39 28 5 10 0.537 0.582 0.333 0.249
92 Bird 51 31 31 14 20 17 0.622 0.689 0.541 0.148
90 Daugherty 42 40 24 17 18 23 0.512 0.585 0.439 0.146
93 Mullin 34 48 20 26 14 22 0.415 0.435 0.389 0.046
91 Hakeem 52 30 36 20 16 10 0.634 0.643 0.615 0.027

DQuinn1575 wrote:

Owly wrote:2) I would imagine based on what I've heard of Bird's absences that they skewed towards Boston back-to-backs.

You heard wrong, Bird played 15 of team's 21 back-to-backs in 91 and 10 of 20 in 92. That's right in line with 60 and 45 games.

Owly wrote:3) I wouldn't think .159 is a massive, elite level difference in WoWY win % because
a) from a similar era and span (92 and 93) STATS Inc had
Jordan .611
Olajuwon .458
Barkley .392
Rodman .383
Mutombo .295
Daugherty .258
Robinson .250
Mullin .178

Jordan hardly missed any games, 2-10 games is just too small a sample.
So I used a 15 game cutoff for these guys 90-93.
I looked at top group and get the data below
total team wins,losses, played wins played losses, dnp wins and losses and difference
Bird is second and fourth of the group; behind Rodman and Barkley splitting him,
It's also harder for Bird as he and Hakeem are the only ones whose teams are above .500 without them, and it is harder to improve a plus .500 team than a negative one.


w l on w on l off w off l tot % on % off % diff %



93 Rodman 40 42 36 26 4 16 0.488 0.581 0.200 0.381
91 Bird 56 26 46 14 10 12 0.683 0.767 0.455 0.312
91 Barkley 44 38 39 28 5 10 0.537 0.582 0.333 0.249
92 Bird 51 31 31 14 20 17 0.622 0.689 0.541 0.148
90 Daugherty 42 40 24 17 18 23 0.512 0.585 0.439 0.146
93 Mullin 34 48 20 26 14 22 0.415 0.435 0.389 0.046
91 Hakeem 52 30 36 20 16 10 0.634 0.643 0.615 0.027

I'll trust that you're accurate regarding back to backs and in each of the season numbers above.

I don't love raw Win% WoWY, I just happened to have the data relatively to hand but fwiw, I'd still debate some stuff here

In terms of mathematics, if measuring the team in wins, I believe it's easier to improve team around a pivot point of .500. Going from being outscored by 50 (9500 to 10000 into 10050 to 10000) is (varying slightly with different versions of pythagorean wins) is worth around 3 wins. A move of 100 away from the norm at the extremes (8500 to 10000 going down to 8400 to 10000 or 11500 to 10000 changing to 11600 to 10000) only moves pythag wins by around 1.

It might on average require a higher standard of player to improve a good team, but this would depend very much on the specifics of whose minutes you are replacing.

Don't love the use of games missed as a marker here there would be instances where a player could make a high mark on the leaderboard but didn't miss enough, but if they missed a few more wins ... then they'd make it.

This chops out for instance .340588988 '93 Daugherty from the top named stars samples and would chop .501 Dudley '93
(each 11 games) from the general group. Olajuwon '92 (.405 12 games).
Jordan applies less here because it was a two season thing but the point still applies, substantial sign of impact is missed in such instances that would be if the sample included more "out games" and the player were not missed at all (indeed the team substantially better) in that additional span. I would think a loss requirement would be better, personally.

Then too if the board is just that offered minus those with less than 15 games missed in a season that's not a lot of players in this comparison.

'90 Cavs data may be pretty junky with Nance missing 20 games, Price 9 (and many different players in the 55 to 35 game range), '92 Celtics too has a substantial McHale absence.

It's nice to see some more numbers but I haven't seen anything to move me off my points (only that Bird apparently did not avoid the second game of b2bs).

Fwiw, looking at ElGee's Bird WoWYs from that span they paint a mixed picture (individual ones very positive on '91 far less so on the others, overall 90-92 pretty good but I think not noteworthy [Bird in, Parish in, 51 games out, 3.3 SRS change, 2.9 WoWY score]).

Owly wrote:

DQuinn1575 wrote:

Owly wrote:2) I would imagine based on what I've heard of Bird's absences that they skewed towards Boston back-to-backs.

You heard wrong, Bird played 15 of team's 21 back-to-backs in 91 and 10 of 20 in 92. That's right in line with 60 and 45 games.

Owly wrote:3) I wouldn't think .159 is a massive, elite level difference in WoWY win % because
a) from a similar era and span (92 and 93) STATS Inc had
Jordan .611
Olajuwon .458
Barkley .392
Rodman .383
Mutombo .295
Daugherty .258
Robinson .250
Mullin .178

Jordan hardly missed any games, 2-10 games is just too small a sample.
So I used a 15 game cutoff for these guys 90-93.
I looked at top group and get the data below
total team wins,losses, played wins played losses, dnp wins and losses and difference
Bird is second and fourth of the group; behind Rodman and Barkley splitting him,
It's also harder for Bird as he and Hakeem are the only ones whose teams are above .500 without them, and it is harder to improve a plus .500 team than a negative one.


w l on w on l off w off l tot % on % off % diff %



93 Rodman 40 42 36 26 4 16 0.488 0.581 0.200 0.381
91 Bird 56 26 46 14 10 12 0.683 0.767 0.455 0.312
91 Barkley 44 38 39 28 5 10 0.537 0.582 0.333 0.249
92 Bird 51 31 31 14 20 17 0.622 0.689 0.541 0.148
90 Daugherty 42 40 24 17 18 23 0.512 0.585 0.439 0.146
93 Mullin 34 48 20 26 14 22 0.415 0.435 0.389 0.046
91 Hakeem 52 30 36 20 16 10 0.634 0.643 0.615 0.027

I'll trust that you're accurate regarding back to backs and in each of the season numbers above.

I don't love raw Win% WoWY, I just happened to have the data relatively to hand but fwiw, I'd still debate some stuff here

In terms of mathematics, if measuring the team in wins, I believe it's easier to improve team around a pivot point of .500. Going from being outscored by 50 (9500 to 10000 into 10050 to 10000) is (varying slightly with different versions of pythagorean wins) is worth around 3 wins. A move of 100 away from the norm at the extremes (8500 to 10000 going down to 8400 to 10000 or 11500 to 10000 changing to 11600 to 10000) only moves pythag wins by around 1.

It might on average require a higher standard of player to improve a good team, but this would depend very much on the specifics of whose minutes you are replacing.

Don't love the use of games missed as a marker here there would be instances where a player could make a high mark on the leaderboard but didn't miss enough, but if they missed a few more wins ... then they'd make it.

This chops out for instance .340588988 '93 Daugherty from the top named stars samples and would chop .501 Dudley '93
(each 11 games) from the general group. Olajuwon '92 (.405 12 games).
Jordan applies less here because it was a two season thing but the point still applies, substantial sign of impact is missed in such instances that would be if the sample included more "out games" and the player were not missed at all (indeed the team substantially better) in that additional span. I would think a loss requirement would be better, personally.

Then too if the board is just that offered minus those with less than 15 games missed in a season that's not a lot of players in this comparison.

'90 Cavs data may be pretty junky with Nance missing 20 games, Price 9 (and many different players in the 55 to 35 game range), '92 Celtics too has a substantial McHale absence.

It's nice to see some more numbers but I haven't seen anything to move me off my points (only that Bird apparently did not avoid the second game of b2bs).

Fwiw, looking at ElGee's Bird WoWYs from that span they paint a mixed picture (individual ones very positive on '91 far less so on the others, overall 90-92 pretty good but I think not noteworthy [Bird in, Parish in, 51 games out, 3.3 SRS change, 2.9 WoWY score]).

back to back games and game logs are all from Basketball-Reference, which I believe is a reliable source.
Like you said a lot of noise in all the plus and minus, just think that 91 will show real positive numbers with Bird, and 92 fairly decent