No worries about picking on me fair or unfair.
One of the nice thing about lawyers is that you can usually bet on them to not take personally something that was clearly not meant personally.
...
It is my belief that while Jrue might have been spotty on offense in G1-G3, his defense was for the most part superb and is a big reason why BOS is holding MIA under 100 points for 3 games plus G2 was more about a team-wide lack of effort, not just Jrue (which is why I decided to break down Exasperated Guy's G2 rant).
Along with my belief about Jrue, I offered up his +/- to show that maybe people who were thinking that Jrue sucked for three games may have overweighted his very visible bad plays and underweighted what he routinely does on the defensive end.
To be clear, I largely agree with you that Jrue has been good to very good, and aside from some isolated plays has been a rock on defense. I genuinely didn't mean to pick on you or the point you were making because I happened to think it was true! But:
But I understand that +/- isn't conclusive. Sure, Jrue might have been super lucky to be on the court when other players went supernova or just took over certain periods, which would inflate his +/-. And frankly, we'll never have anything conclusive about how Jrue played in G1-G3 unless we broke down every play (which I presume the Cs have done).
So I don't mind when people mention +/-, even for one game. I think we're all smart enough to give it the appropriate weight (or lack thereof) depending on what point is trying to be made. I mean PHI in G1-G4 was something like +40 when Embiid was on the court and -38 in the 32 minutes that Embiid was off the court. I think (and other people who have podcasts also think) this says something somewhat meaningful about PHI's roster construction. Maybe you don't. Which is fine. But it's an interesting stat to mention.
And as we always say, YMMV.
Yeah that's the thing - the proposition is likely no more or less true for having cited single-game +/- than if nothing has been cited at all.
There has been a long-term struggle to quantify single-game impact. In baseball, we have runs added from each at-bat, which is helpfully discrete, and can adjust for leverage, on-base and outs situation, and other such things to tell us exactly how much difference a player made on either side of the ball. In basketball, possessions are discrete
enough, usually, but the interplay between players and the fluid impact of defense and decision-making means we'll only ever approach an idea of how much difference a player made on a given possession. But if we don't have that, we can't aggregate up to a single game, so we're left with top-down things that merely approximate.
One of those tools is plus-minus, or adjusted plus-minus, or box plus-minus, or the various derivations of it (which notably include DARKO and LEBRON). Another is game score, which is a nice little heuristic imported from baseball that probably has some value (as long as we note that it just measures what box score stats measure, which avoids a lot of defense). I think it's probably an incremental improvement over +/-, much the same as both are an improvement over "points!", which we thankfully usually avoid around here. +/- does have the benefit of capturing some aspect of a player's defensive contributions
But just as it's very possible to have a player who scores a lot of empty points in a game, it's very possible to have an empty plus-minus, where you were a passenger on some strong passages of play by the rest of your team, or the other team's shooters happened to hit when you were out and miss when you were in. That's why we have adjustments to the raw number, to account for some aspects of shooting variance, garbage time, who you played with, etc (depending on which version and how fancy it is). If someone deeper into the basketball analytics stuff than I am thinks there's a clear winner among the approaches, I'd be interested to learn it.
Hence the need to eliminate some of the noise by accumulating greater sample size of the metric, before we start drawing conclusions with it. And as with other similar metrics where with small sample sizes, the noise can swamp the signal, it is also true that extreme outliers can be notable even in smaller sample sizes. But that means, like, Embiid's 50-point game combined with a few other games. It doesn't mean the difference between guys' individual +/-s within a game can be reasonable compared, or (For small differences) even doing so across a 4-game sample.