APBRmetrics

ideus, my comments were not meant to suggest that the Celtics recently became opposed in principle to making net positive plus-minus trades/personnel moves. They were to note that they had seemingly abandoned their previously apparent and succcessful strategy.

So, it is worth summarizing what the net effect has been (to date) to better anchor the conversation and how this relates to plausible beliefs about how this roster makeover promotes a more promising future.

Kevin Pelton recently wrote that his RPM projection for this years Celtics is 44.8 wins.

http://www.espn.com/nba/story/_/id/2049 ... vin-pelton

By contrast, basketball reference lists Predicted Wins for last year's team at 48.

I infer from this in plus-minus terms that the Celtics "makeover" made them worse in per 100 possession terms by about 1.4 points. And, in so doing incurred a salary bill increase of about $19 million.

So when you write "I don't think the Celtics expect this trade to make them immediately better." that, it seems to me, is a bit of an understatement.

The Celtics are currently about $10 million away from paying a luxury tax, and to get to championship level strength through player acquisition (in plus minus terms) will require significantly more than a top five player. I will admit that I have a limited imagination, but the phrase "Can't get there from here." is all that comes to mind.

So from here they either get on a 50 plus win pace or don't. Make conference final or don't. If they don't for the first, a lot of folks will probably let it slide. "Need time to gel" Will they let the more important goal slide? I dunno. Probably blame it on refs, injury or something. If they blame it on someone, it will probably be Kyrie, probably not Ainge (but maybe), not Stevens, probably not Horford, not Hayward or Tatum. Maybe some on Smart or Morris. Wait n see where deserved. Management and coaching claim the authority & ultimate responsibility. At least in abstract. If things don't go as hoped?

New kinda like the old? Kyrie as super DJ? Hawyward as lesser Bird? Horford as Parish? Tatum as hoped for McHale? Who is Ainge? Rozier? Not close, not yet.

Very little similarity to 2008 Celtics. KG-like? Hell no. Even Perkins, no. Pierce? Not exactly but Hayward is closer than the other comparisons. Ray Allen? Hell no. Rondo? No, no pass first.

Horford pairings with Bradley and Crowder were bad in playoffs. Bradley and Crowder now gone. But Horford in playoffs had negative pairings for all of his top 5 pairings. Horford still there. Team -78 with him on the court, mainly due to very sucky defense. 7 of 9 playoffs he has been negative on the court. 6 of them badly. But they knew that right? Hayward, 2 playoffs, once terrible on court plus minus, one horrendous. Outstanding picks of huge salary leaders! Kyrie, great "on court" playoff plus minus but barely over neutral on / off. Another great pick?


Horford negative in playoffs in all of his 3, 4 and 5 man top 5s except for one 5 man lineup (all 4 other players now gone). 19 negatives, one positive. Cornerstone... Fine in regular season, bad in playoffs (in toto) last season.

Against Cavs, Crowder was good by boxscore stats. Horford sucked. Almost every other Celt sucked besides Green, Jerebko and Zeller.

In final two rounds (12 games), Stevens used one lineup more than 5 minutes per game. Ten more for 1-4 minutes per game. Then 128 dink lineups for less than 1 minute per game each. Those 128 dink lineups cost the Celtics 53 points... in 52 minutes. Celts beat Washington by 7 points. lost to Cavs by 100 points. So I guess dink lineups were only about half the problem overall, 60% of the Cavs series. In Cavs series Stevens won 16% of his lineups used under 1 minute per game. Yuck.

Recognizing that past is past but imagining I had input now...

I'd consider Irving-Hayward-Tatum as my core and actively try to trade almost everybody else. I'd give Ojeleye some time before making him available. I'd be ok if Morris stuck around. Beyond that, move em out for a SG who can really shoot and a defensive Center. My goal would be peak in 2019-20 and 2020-21. I don't see greater things in next two years as is or during the massive overhaul of the massive overhaul.

If it is title or bust, you might have to move on from 1, 2 or all 3 of that big 3. But that is for 2-3 years from now.

Crow wrote:My goal would be peak in 2019-20 and 2020-21

With the Warriors as good as they are, and LeBron still great, that timeframe does indeed seem to make the most sense to peak at. Not many titles available for grabs before that

Thomas & Crowder may have helped attract Horford. At least 2 of 3 probably helped attract Hayward. Hayward & Horford probably helped attract Irving. Thomas & Crowder gone. Horford will be eventually. Will any of the Celtics draft picks become pivotal? Not Smart, Rozier, Brown. Maybe Tatum but I am guessing not enough. Maybe a future first? Better do it fast. They will be shuffling up on Irving salary fairly soon, maybe more than Irving impact. Does Hayward have another gear? I doubt it. Shuffle, shuffle.

Perhaps a final note on the dreariness that is the Kyrie Irving trade, and it relates not to plus-minus considerations in a direct sense.

Perhaps it is indeed too much to expect that executives, the best basketball minds the US has to offer, paid millions of dollars a year to manage/coach their rosters, keep current of cutting-edge analytic developments of a decade previous. But can the same pass be given for lacking a basic understanding of the actual NBA aging curve?

If you look at the commentary on the KI trade, the text and subtext emphasizes that he is "young", but this is simply not so. Two examples:

Here's Doc Rivers comment, uncritically related by Chris Forsberg: "The Celtics gave up a lot to get him, but he's 25 years old, and that's the other thing people forget about: He's so young." http://www.espn.com/nba/story/_/id/2061 ... on-celtics

And here's Danny Ainge accepting this exact premise from the social media interviewer about two minutes from the end: https://twitter.com/celtics/status/9051 ... 70890246.7

As "everyone" knows, the average age of a player in the NBA is 26+ (basically an eternal fact, with basketball reference showing 26.7 for last year) what represents the approximate peak of the plateau of the aging curve. And regardless of the precise peak, the average expected improvement at an around that point, even if positive, is trivial.

But there is, of course, better reason to believe that that Kyrie Irving is actually on the downslope of the aging curve because of the well-above average miles he has already travelled. Aging curves, of course, represent the combined effects of game experience (a curve with always positive but diminishing slope) and calendar age (what rises then falls with the vicissitudes of age). And Kyrie has had vastly more experience that the typical 25 year old, having all six years of his career playing starter's minutes.

And this directional effect (and its size) can be estimated by referring to the history of similar NBA players.

Looking at guards who have played as many or more minutes (13,024) within their first 6 seasons by the age 25 (technically in B-R terms he was 24, but almost 25, and the sample size is very small with 24) you can calculate the average aging effect in terms of BPM in subsequent years.

Looking at the careers who have played three years thereafter (what is the length of KI's contract and what represents pretty much the length of the aging curve "plateau") you can fiddle around with this or that but basically the net effect, on average, is slightly negative, with average BPM in that period being -0.6 to -0.9 lower.

So, that's the average, but what about the variability? What player career paths in history can Danny Ainge hope that Kyrie Irving lucks into, so as to redeem this trade. Well, it's pretty much just one: Russell Westbrook or bust, at least in plus-minus terms. (All the other "similar" players who improved, didn't improve enough to overcome the aforementioned 1.4 point hole that this off-season's maneuvers have yielded.)

And a note to careful, regular readers, Russell Westbrook is a known and curious BPM anomaly, with his skillset matching "perfectly" with the underlying regression result, yielding "curious" results, e.g. a BPM last year of 15.6. Accordingly, for him and him alone, I used RPM for the years in question, where he still represents the only historical straw for Danny Ainge to grasp at.

Finally, it should be noted that minor, average improvement can be expected when one eliminates from the sample guards smaller than Kyrie (i.e. 6'2" and below). But in that case, the sample size shrinks to only seven players, and there is an average gain in BPM in the three year window of all of 0.2 (what is only made positive by the inclusion of good ol' Russell Westbrook).

So, that's that. Bottom line is that a 25 year old player who has played 13,000 minutes isn't young; he's middle-aged, and it's simply ridiculous that folks in authority would represent this truly basic fact otherwise (and that no one in the media would offer the short and sharp correction on the point).

Yep. Read it here.

To get an alternative future (to not good enough for a title), Celtics need a big hit in draft (Tatum or later or both) or another star that wants to play with Irving and the others. No immediate guesses as to who that could be but will ponder some from now til eternity or collapse of their dream. "This" isn't enough or probably even close (with a 4-1 finals loss not counting as close, usually and a conference finals loss probably not either unless decided by a few points total).

Dink lineups averaging less than 1 minute per game in playoffs were a colossal failure for every Cavs' opponent but the Celtics and Stevens were the worst. Even the Warriors only won 21% of their dink lineups against them. Might not be a fluke. Maybe Lue and / or LeBron need to get some recognition for this. Haven't checked prior years. I assume with his lower # of lineups used Lue had a majority of less-dink lineup matchup advantages. Will coaches in general learn to be less dinky? Not if everyone in organization continues to defer to them about this habit of game coaching.

In 2015-16 most playoff opponents did better with their dink lineups against Cavs but still not good. The average for all opponents was better but still terrible. Raptors and D Casey only won 4% of their dink lineups, pulling down the average.

There could be plenty more study of this. Minutes of dink lineups, dink vs dink, bigger vs. dink, yr to yr trends, etc. I wonder how much teams have studied this. Practice still seems sub-optimal Probably can't completely eliminate dink lineups but logic (less familarity and less continuity contributes to possibility of less effective play) and results say reduce them. But the multi-millionaire coaches appear largely to just wing it.