Who is our best player?

[nickagneta]

So ESPN came out with their initial RPM stat. Yes, it's very wonky and I don't love they they won't publish the formula, though in an article today they did reveal things that make up the formula. Not sure if that is new.

But my point is Jaylen Brown is rated at 12th in the league at 4.23 and is rated as the 2nd best shooting guard. He is also rated as the 9th best player in RPM wins at 2.71.

Jayson Tatum is rated as the 38th best in the league at 2.49 and has only 1.51 RPM wins. His missing games might account for that, but RPM isn't being kind to Tatum thus far. Tatum is currently rated as the 7th best small forward.

Not sure this says anything other than Brown is having a better year but thought it might interest those in the statistical debate going on here.

[footey]

BE curious to see what guys rate higher than Tatum; that seems crazy low.

[PhoSita]

So ESPN came out with their initial RPM stat. Yes, it's very wonky and I don't love they they won't publish the formula, though in an article today they did reveal things that make up the formula. Not sure if that is new.

But my point is Jaylen Brown is rated at 12th in the league at 4.23 and is rated as the 2nd best shooting guard. He is also rated as the 9th best player in RPM wins at 2.71.

Jayson Tatum is rated as the 38th best in the league at 2.49 and has only 1.51 RPM wins. His missing games might account for that, but RPM isn't being kind to Tatum thus far. Tatum is currently rated as the 7th best small forward.

Not sure this says anything other than Brown is having a better year but thought it might interest those in the statistical debate going on here.

The funny thing is that in the past it's been the opposite; Tatum has huge +/- swings associated with his playing time, while Brown is in the negative.

I tend to think these RPM numbers have more to do with the context of which guys spend more time on the floor with bench units as opposed to play with/against starters than anything else.


I continue to think the easy answer here is "Tatum," but the fact that it's a debate at all is really good thing, because it speaks to how well Jaylen is playing.

[colincb]

So ESPN came out with their initial RPM stat. Yes, it's very wonky and I don't love they they won't publish the formula, though in an article today they did reveal things that make up the formula. Not sure if that is new.

But my point is Jaylen Brown is rated at 12th in the league at 4.23 and is rated as the 2nd best shooting guard. He is also rated as the 9th best player in RPM wins at 2.71.

Jayson Tatum is rated as the 38th best in the league at 2.49 and has only 1.51 RPM wins. His missing games might account for that, but RPM isn't being kind to Tatum thus far. Tatum is currently rated as the 7th best small forward.

Not sure this says anything other than Brown is having a better year but thought it might interest those in the statistical debate going on here.

I was going to make the same points about Real +/-, but whatever its flaws, it's better than raw +/- which is worthless for comparing players )though it might be useful for lineup construction). Last year is a good example of raw +/-'s main flaw because when Tatum was off the court Brown was playing with Semi or GW at PF. When Brown was off the floor Tatum was playing with Smart or Haywood at SG or SF instead. All else being equal, Brown was at an extreme disadvantage with raw +/- compared to Tatum.

Also, the OP asked was who is our best player, not who was, or who will be when all is said in done (I believe JT), but it is clear that JB is the best player right now by the numbers.

[Somebody]

So ESPN came out with their initial RPM stat. Yes, it's very wonky and I don't love they they won't publish the formula, though in an article today they did reveal things that make up the formula. Not sure if that is new.

But my point is Jaylen Brown is rated at 12th in the league at 4.23 and is rated as the 2nd best shooting guard. He is also rated as the 9th best player in RPM wins at 2.71.

Jayson Tatum is rated as the 38th best in the league at 2.49 and has only 1.51 RPM wins. His missing games might account for that, but RPM isn't being kind to Tatum thus far. Tatum is currently rated as the 7th best small forward.

Not sure this says anything other than Brown is having a better year but thought it might interest those in the statistical debate going on here.

It's become really wonky recently - their founder left the team a year ago I think, which sucks because tests have OG RPM as one of the "best" +/- metrics out there.

I'd use Estimated Plus-Minus instead of RPM, it has Jaylen ranked 18th in the league while underrating his defensive value in the playoffs.

[Somebody]

This will be debated all season, and beyond.

You really think so?

Tatum missed 5 games, Boston went 2-3 despite getting Walker back for the last 4.  They were 7-3 with Tatum to start the year and without Walker playing at all (and when they close out the Bulls will be 8-3 with Tatum and no Walker).  And despite the Cleveland game yesterday, Boston is still 3.4 points per 100 possessions better when Brown is on the bench.

Tatum is the most valuable Celtic because he is the best Celtic and frankly I don't think it is all that close.

Genuinely cannot believe you're clinging so desperately to the laughable idea that we're a better team with Brown on the bench by using raw on/off. Funny stuff

And yet historically that stat has been a pretty darn accurate predictor of actual importance to winning.  Look at champions.  The Warriors, for example, Curry, not Durant, is what made them tick.  Durant was a better player than Curry, but he wasn't more critical to the Warriors success.  That was Curry.  Without Curry the Warriors were a good team, with Curry the Warriors were an all time great team and that wasn't the case for Durant.  We saw this quite well during the 17-18 season when the Warriors were 41-10 in Curry's 51 games and 17-14 in the 31 games he missed.  They were 49-19 with Durant so 9-5 without him (there was some overlap in missed games).  Curry's on/off differential was +12.1.  Durant was +1.9.  And that was fairly consistent all 3 of their seasons together.  Curry made the Warriors go even though Durant was the better player.  You see it doesn't necessarily equate to skill, talent, etc., merely how integral the player is to the team. 

Tatum is far more integral to the success of Boston than Brown is.  We've seen this pretty consistently over the last 3 years in that Boston just isn't very good in games Tatum has missed and plays about the same (or better) when Brown is out.  It doesn't mean Brown isn't a good player, he absolutely is, but his value to the scoreboard just isn't very high and the reality is, Boston has just been better with him on the bench over the last 2+ seasons, while Tatum has been by far the leader in that category.  The fact that they both start and have played similar minutes makes that way more striking as you generally find units to have fairly similar numbers, especially when the individual production has been similar enough.  Boston needs Tatum, Boston doesn't need Brown.

What is it in Tatum's game that I'm missing that makes him that great, if we're following your reasoning?

These on/off numbers suggest that Tatum is basically as important as what LeBron James is to the Lakers.
Anthony Davis is a great player, but according (and also to the eye test to be honest) to the numbers the Lakers are only functioning at top level because of LeBron. Same with Harden in Houston in comparison with any other star that played there (Paul, Westbrook).

To me Jayson Tatum looks like 15-20 as a player ranking in the NBA and Jaylen Brown like 25-30. How come this isn't reflected in the numbers?

And will the improvement of Jaylen Brown change this difference between Tatum and Brown?

You're missing the fact that raw on/off numbers are incredibly noisy. Tatum isn't doing that well in adjusted plus-minus metrics.

Except you know you can actually see Boston's record in games when Tatum is out and games when Brown is out.  I've put that in this thread plenty of times but it bears repeating again.

Boston's record without Brown available / with Brown
17-18 season - 8-4 / 47-23
18-19 season - 8-0 / 41-33
19-20 season - 10-5 / 38-19
20-21 season - 0-0 / 10-6

So 26-9 without Brown or 74.3% and 136-81 with Brown or 62.7%


Tatum has been far healthier so more room for variance here and there, but here are his splits
17-18 season - 2-0 / 53-27 (Tatum's rookie year)
18-19 season - 1-2 / 48-31
19-20 season - 2-4 / 46-20
20-21 season - 2-3 / 8-3

So 7-9 without Tatum or 43.8% and 155-81 with Tatum or 65.7%

You can call the stat noisy all you want, but the only place any of this matters is the actual wins and losses and in that, Tatum is way more important than Brown.  It isn't close and it hasn't been close since basically Tatum stepped foot on the floor.  And over the approximate half a season Brown has missed in the last 4 years, Boston actually wins those games at a much higher percentage than in the games Brown actually plays in.  That quite simply is a fact.  Boston wins more when Brown doesn't play.  Period.

Aaaaaaaand adjusted +/- metrics absolutely crush raw on/off when it comes to predicting wins!

You say that and yet you see the actual results when they miss games.  If what you are relying on doesn't match reality, then it isn't reality.

It's like basketball isn't a complex game that can confuse raw +/- data with lineups, rotations and schemes

The notion that raw on/off sucks at predicting reality compared to adjusted +/- = adjusted +/- doesn't match reality and isn't reality is hilarious even by your standards though

[Moranis]

So ESPN came out with their initial RPM stat. Yes, it's very wonky and I don't love they they won't publish the formula, though in an article today they did reveal things that make up the formula. Not sure if that is new.

But my point is Jaylen Brown is rated at 12th in the league at 4.23 and is rated as the 2nd best shooting guard. He is also rated as the 9th best player in RPM wins at 2.71.

Jayson Tatum is rated as the 38th best in the league at 2.49 and has only 1.51 RPM wins. His missing games might account for that, but RPM isn't being kind to Tatum thus far. Tatum is currently rated as the 7th best small forward.

Not sure this says anything other than Brown is having a better year but thought it might interest those in the statistical debate going on here.

I was going to make the same points about Real +/-, but whatever its flaws, it's better than raw +/- which is worthless for comparing players )though it might be useful for lineup construction). Last year is a good example of raw +/-'s main flaw because when Tatum was off the court Brown was playing with Semi or GW at PF. When Brown was off the floor Tatum was playing with Smart or Haywood at SG or SF instead. All else being equal, Brown was at an extreme disadvantage with raw +/- compared to Tatum.

Also, the OP asked was who is our best player, not who was, or who will be when all is said in done (I believe JT), but it is clear that JB is the best player right now by the numbers.

Brown's top 5 man unit last year without Tatum was Walker, Smart, Hayward, Theis (those same 4 were also Tatum's top 5 man unit without Brown).  Those 4 with Brown were -17 in 85.4 minutes, however with Tatum were +52 in 170.8 minutes.  Overall and of the top 20 5 man units by minutes played, Brown was only in 2 of them without Tatum that one and one with Smart, Hayward, Ojeleye, Theis.  Tatum on the other hand was in 5 of the top 20 units without Brown.  The Walker, Smart, Hayward, and Theis was actually the 3rd most common unit last year.  After that though, the 8th most common unit was Tatum with Wanamaker, Smart, Ojeleye, and Kanter.  Tatum also appeared with Edwards, Wanamaker, Ojeleye, Kanter and Walker, Smart, Ojeleye, Theis and finally Wanamaker, Smart, G. Williams, Theis.

It isn't surprising that Tatum appeared on more units as Brown missed 15 games so Tatum just played more.  The key point is however in those 15 games Boston went 10-5 winning games at the exact same rate as when Brown played.  That is the point I've been making this entire thread.  Boston just keeps right on winning when Brown doesn't play at all and that most certainly is not the case when Tatum doesn't play.

[colincb]

So ESPN came out with their initial RPM stat. Yes, it's very wonky and I don't love they they won't publish the formula, though in an article today they did reveal things that make up the formula. Not sure if that is new.

But my point is Jaylen Brown is rated at 12th in the league at 4.23 and is rated as the 2nd best shooting guard. He is also rated as the 9th best player in RPM wins at 2.71.

Jayson Tatum is rated as the 38th best in the league at 2.49 and has only 1.51 RPM wins. His missing games might account for that, but RPM isn't being kind to Tatum thus far. Tatum is currently rated as the 7th best small forward.

Not sure this says anything other than Brown is having a better year but thought it might interest those in the statistical debate going on here.

I was going to make the same points about Real +/-, but whatever its flaws, it's better than raw +/- which is worthless for comparing players )though it might be useful for lineup construction). Last year is a good example of raw +/-'s main flaw because when Tatum was off the court Brown was playing with Semi or GW at PF. When Brown was off the floor Tatum was playing with Smart or Haywood at SG or SF instead. All else being equal, Brown was at an extreme disadvantage with raw +/- compared to Tatum.

Also, the OP asked was who is our best player, not who was, or who will be when all is said in done (I believe JT), but it is clear that JB is the best player right now by the numbers.

Brown's top 5 man unit last year without Tatum was Walker, Smart, Hayward, Theis (those same 4 were also Tatum's top 5 man unit without Brown).  Those 4 with Brown were -17 in 85.4 minutes, however with Tatum were +52 in 170.8 minutes.  Overall and of the top 20 5 man units by minutes played, Brown was only in 2 of them without Tatum that one and one with Smart, Hayward, Ojeleye, Theis.  Tatum on the other hand was in 5 of the top 20 units without Brown.  The Walker, Smart, Hayward, and Theis was actually the 3rd most common unit last year.  After that though, the 8th most common unit was Tatum with Wanamaker, Smart, Ojeleye, and Kanter.  Tatum also appeared with Edwards, Wanamaker, Ojeleye, Kanter and Walker, Smart, Ojeleye, Theis and finally Wanamaker, Smart, G. Williams, Theis.

It isn't surprising that Tatum appeared on more units as Brown missed 15 games so Tatum just played more.  The key point is however in those 15 games Boston went 10-5 winning games at the exact same rate as when Brown played.  That is the point I've been making this entire thread.  Boston just keeps right on winning when Brown doesn't play at all and that most certainly is not the case when Tatum doesn't play.

Raw +/- is crap, so why keep using it?

[gouki88]

So ESPN came out with their initial RPM stat. Yes, it's very wonky and I don't love they they won't publish the formula, though in an article today they did reveal things that make up the formula. Not sure if that is new.

But my point is Jaylen Brown is rated at 12th in the league at 4.23 and is rated as the 2nd best shooting guard. He is also rated as the 9th best player in RPM wins at 2.71.

Jayson Tatum is rated as the 38th best in the league at 2.49 and has only 1.51 RPM wins. His missing games might account for that, but RPM isn't being kind to Tatum thus far. Tatum is currently rated as the 7th best small forward.

Not sure this says anything other than Brown is having a better year but thought it might interest those in the statistical debate going on here.

I was going to make the same points about Real +/-, but whatever its flaws, it's better than raw +/- which is worthless for comparing players )though it might be useful for lineup construction). Last year is a good example of raw +/-'s main flaw because when Tatum was off the court Brown was playing with Semi or GW at PF. When Brown was off the floor Tatum was playing with Smart or Haywood at SG or SF instead. All else being equal, Brown was at an extreme disadvantage with raw +/- compared to Tatum.

Also, the OP asked was who is our best player, not who was, or who will be when all is said in done (I believe JT), but it is clear that JB is the best player right now by the numbers.

Brown's top 5 man unit last year without Tatum was Walker, Smart, Hayward, Theis (those same 4 were also Tatum's top 5 man unit without Brown).  Those 4 with Brown were -17 in 85.4 minutes, however with Tatum were +52 in 170.8 minutes.  Overall and of the top 20 5 man units by minutes played, Brown was only in 2 of them without Tatum that one and one with Smart, Hayward, Ojeleye, Theis.  Tatum on the other hand was in 5 of the top 20 units without Brown.  The Walker, Smart, Hayward, and Theis was actually the 3rd most common unit last year.  After that though, the 8th most common unit was Tatum with Wanamaker, Smart, Ojeleye, and Kanter.  Tatum also appeared with Edwards, Wanamaker, Ojeleye, Kanter and Walker, Smart, Ojeleye, Theis and finally Wanamaker, Smart, G. Williams, Theis.

It isn't surprising that Tatum appeared on more units as Brown missed 15 games so Tatum just played more.  The key point is however in those 15 games Boston went 10-5 winning games at the exact same rate as when Brown played.  That is the point I've been making this entire thread.  Boston just keeps right on winning when Brown doesn't play at all and that most certainly is not the case when Tatum doesn't play.

Raw +/- is crap, so why keep using it?

You won’t get a reasonable answer - don’t bother

[Moranis]

So ESPN came out with their initial RPM stat. Yes, it's very wonky and I don't love they they won't publish the formula, though in an article today they did reveal things that make up the formula. Not sure if that is new.

But my point is Jaylen Brown is rated at 12th in the league at 4.23 and is rated as the 2nd best shooting guard. He is also rated as the 9th best player in RPM wins at 2.71.

Jayson Tatum is rated as the 38th best in the league at 2.49 and has only 1.51 RPM wins. His missing games might account for that, but RPM isn't being kind to Tatum thus far. Tatum is currently rated as the 7th best small forward.

Not sure this says anything other than Brown is having a better year but thought it might interest those in the statistical debate going on here.

I was going to make the same points about Real +/-, but whatever its flaws, it's better than raw +/- which is worthless for comparing players )though it might be useful for lineup construction). Last year is a good example of raw +/-'s main flaw because when Tatum was off the court Brown was playing with Semi or GW at PF. When Brown was off the floor Tatum was playing with Smart or Haywood at SG or SF instead. All else being equal, Brown was at an extreme disadvantage with raw +/- compared to Tatum.

Also, the OP asked was who is our best player, not who was, or who will be when all is said in done (I believe JT), but it is clear that JB is the best player right now by the numbers.

Brown's top 5 man unit last year without Tatum was Walker, Smart, Hayward, Theis (those same 4 were also Tatum's top 5 man unit without Brown).  Those 4 with Brown were -17 in 85.4 minutes, however with Tatum were +52 in 170.8 minutes.  Overall and of the top 20 5 man units by minutes played, Brown was only in 2 of them without Tatum that one and one with Smart, Hayward, Ojeleye, Theis.  Tatum on the other hand was in 5 of the top 20 units without Brown.  The Walker, Smart, Hayward, and Theis was actually the 3rd most common unit last year.  After that though, the 8th most common unit was Tatum with Wanamaker, Smart, Ojeleye, and Kanter.  Tatum also appeared with Edwards, Wanamaker, Ojeleye, Kanter and Walker, Smart, Ojeleye, Theis and finally Wanamaker, Smart, G. Williams, Theis.

It isn't surprising that Tatum appeared on more units as Brown missed 15 games so Tatum just played more.  The key point is however in those 15 games Boston went 10-5 winning games at the exact same rate as when Brown played.  That is the point I've been making this entire thread.  Boston just keeps right on winning when Brown doesn't play at all and that most certainly is not the case when Tatum doesn't play.

Raw +/- is crap, so why keep using it?

because it isn't.  Everyone keeps saying that, but yet historically if you compare players, it almost universally shows how important a player is to the only thing that matters, which is winning.  Tatum and Brown are both starters playing similar minutes on the Celtics, yet the team falls apart without Tatum and continues right along without Brown.  This is best illustrated when they miss full games, where the team is well below .500 without Tatum, yet wins at a higher rate without Brown.  It doesn't mean Brown is a bad player, he isn't, Brown is very good player, he just doesn't impact winning all that much.  And that is all that really matters to me when thinking about a player's value. 

[td450]

So ESPN came out with their initial RPM stat. Yes, it's very wonky and I don't love they they won't publish the formula, though in an article today they did reveal things that make up the formula. Not sure if that is new.

But my point is Jaylen Brown is rated at 12th in the league at 4.23 and is rated as the 2nd best shooting guard. He is also rated as the 9th best player in RPM wins at 2.71.

Jayson Tatum is rated as the 38th best in the league at 2.49 and has only 1.51 RPM wins. His missing games might account for that, but RPM isn't being kind to Tatum thus far. Tatum is currently rated as the 7th best small forward.

Not sure this says anything other than Brown is having a better year but thought it might interest those in the statistical debate going on here.

I was going to make the same points about Real +/-, but whatever its flaws, it's better than raw +/- which is worthless for comparing players )though it might be useful for lineup construction). Last year is a good example of raw +/-'s main flaw because when Tatum was off the court Brown was playing with Semi or GW at PF. When Brown was off the floor Tatum was playing with Smart or Haywood at SG or SF instead. All else being equal, Brown was at an extreme disadvantage with raw +/- compared to Tatum.

Also, the OP asked was who is our best player, not who was, or who will be when all is said in done (I believe JT), but it is clear that JB is the best player right now by the numbers.

Brown's top 5 man unit last year without Tatum was Walker, Smart, Hayward, Theis (those same 4 were also Tatum's top 5 man unit without Brown).  Those 4 with Brown were -17 in 85.4 minutes, however with Tatum were +52 in 170.8 minutes.  Overall and of the top 20 5 man units by minutes played, Brown was only in 2 of them without Tatum that one and one with Smart, Hayward, Ojeleye, Theis.  Tatum on the other hand was in 5 of the top 20 units without Brown.  The Walker, Smart, Hayward, and Theis was actually the 3rd most common unit last year.  After that though, the 8th most common unit was Tatum with Wanamaker, Smart, Ojeleye, and Kanter.  Tatum also appeared with Edwards, Wanamaker, Ojeleye, Kanter and Walker, Smart, Ojeleye, Theis and finally Wanamaker, Smart, G. Williams, Theis.

It isn't surprising that Tatum appeared on more units as Brown missed 15 games so Tatum just played more.  The key point is however in those 15 games Boston went 10-5 winning games at the exact same rate as when Brown played.  That is the point I've been making this entire thread.  Boston just keeps right on winning when Brown doesn't play at all and that most certainly is not the case when Tatum doesn't play.

Raw +/- is crap, so why keep using it?

because it isn't.  Everyone keeps saying that, but yet historically if you compare players, it almost universally shows how important a player is to the only thing that matters, which is winning.  Tatum and Brown are both starters playing similar minutes on the Celtics, yet the team falls apart without Tatum and continues right along without Brown.  This is best illustrated when they miss full games, where the team is well below .500 without Tatum, yet wins at a higher rate without Brown.  It doesn't mean Brown is a bad player, he isn't, Brown is very good player, he just doesn't impact winning all that much.  And that is all that really matters to me when thinking about a player's value.

To me, this is an analysis of just how bad the Walker-Smart ball handling duo was and still is. Without Tatum, it can get ugly.

I do think that Jaylen has become more independent of this as the team has learned to just give him the ball more. The dynamics have changed dramatically since last year.