[Paint Touches] Plus/minus leaderboard

[MU82]

So ... Markus scored 89 points in consecutive games against Providence and Nova ... and he was only a combined +3.

He plays mediocre hoops against SH and DePaul but is a combined +9.

Interesting stat.

[Benny B]

+/- is a bit deceiving in basketball, because unlike the sports where it’s a meaningful metric (e.g. hockey), you don’t have specific lines you rotate through a game.  In other words, Marcus could be on the floor in one of 7 or 8 combinations of players, whereas in hockey, if you’re the center, most of your time on the ice is with your line mates on the wings.

So I’m not going to read too much into stand-alone +/-, but what would be more enlightening is a 2D matrix that shows +/- for pairs of players, i.e. how does Marcus compare when he and AR are on the floor together vs. when he’s on the floor with Greg or Sacar. 

[PaintTouches]

So I’m not going to read too much into stand-alone +/-, but what would be more enlightening is a 2D matrix that shows +/- for pairs of players, i.e. how does Marcus compare when he and AR are on the floor together vs. when he’s on the floor with Greg or Sacar.

I've been trying to make my spreadsheets filter-able, but Wordpress is awful and wants me to pay up to get that feature. Still thinking about it.

For now, just let me know if you have any requests:
Rowsey, no Howard: +19 in 56 minutes
Howard, no Rowsey: +11 in 51 minutes

Pretty similar overall. But this is my focus:
Rowsey/Howard/Anim: -15 in 73 mins
Rowsey/Howard/Cain: -6 in 35 mins
Rowsey/Howard/Elliott: +21 in 58 mins 

Moar Elliott, please. 

[PaintTouches]

Another update bump:

https://painttouches.com/2018/01/22/big-east-tiers-1-22-18-edition/

[MU82]

I've been trying to make my spreadsheets filter-able, but Wordpress is awful and wants me to pay up to get that feature. Still thinking about it.

For now, just let me know if you have any requests:
Rowsey, no Howard: +19 in 56 minutes
Howard, no Rowsey: +11 in 51 minutes

Pretty similar overall. But this is my focus:
Rowsey/Howard/Anim: -15 in 73 mins
Rowsey/Howard/Cain: -6 in 35 mins
Rowsey/Howard/Elliott: +21 in 58 mins 

Moar Elliott, please.

Doesn't it matter who the opponents were? For example, did AR, MH and GE happen to play together more against crappy teams or against excellent teams, and wouldn't that have made a difference?

Not ripping on the research, just genuinely curious.

[PaintTouches]

If we were only talking about a one or two game sample, sure, but we're 7 games in, so all the lineups have had opportunities against "good" and "bad" teams alike.

What does matter much more than opponent is time on court, that's why I include it, or else it has no real perspective. For example, all but 2 of the Rowsey/Howard/Elliott minutes (56) have come in the last 5 games, and the team is +18 in that span. Rowsey/Howard/Anim has fewer minutes the last 5 games (50 min) and is a -25. 

This stat does not explain the why and shouldn't be used as an end all, be all, but I do think tracking trends like that one does have a lot of value.

[MU82]

If we were only talking about a one or two game sample, sure, but we're 7 games in, so all the lineups have had opportunities against "good" and "bad" teams alike.

What does matter much more than opponent is time on court, that's why I include it, or else it has no real perspective. For example, all but 2 of the Rowsey/Howard/Elliott minutes (56) have come in the last 5 games, and the team is +18 in that span. Rowsey/Howard/Anim has fewer minutes the last 5 games (50 min) and is a -25. 

This stat does not explain the why and shouldn't be used as an end all, be all, but I do think tracking trends like that one does have a lot of value.

Thanks for running the numbers.

I don't know whether it has value or not, but I do think it's interesting to see and talk about.

[manny31]

Thanks for running the numbers.

I don't know whether it has value or not, but I do think it's interesting to see and talk about.

I agree that this is interesting. I like the tangible metrics but I also love the intangible. I know 7 games is a reasonable sample size, I am curious what the numbers look like after another 4 or 6 games? Does G. Elliott improve in those games to make these numbers even more convincing? I really like the Freshman I think they all are going to be  huge contributors and one or more might wind up being really, really good. Warriors will beat X!

[Benny B]

Drew Cannon has made himself a comfortable living being Brad Stevens' right-hand quant man, whose job essentially translates into generating lineup combinations.  It's amazing that there aren't more younger coaches who are catching onto this.

In fact, isn't this all something that the fabled SportVu system covers, i.e. analyzing favorable lineups, positioning, etc.?  As I recall, Duke and Marquette were two of the first college programs to sign up, so I've always assumed that Wojo is/was a believer in quant and SportVu may have been part (read: not major, but some iota of consideration) of attracting him to Milwaukee.

Can anyone confirm whether MU is still using SportVu or did that go the way of the Hokie?

[PaintTouches]

Update bump.

https://painttouches.com/2018/01/02/marquette-plus-minus-leaderboard/

As to your question Benny, the BMO Bradley Center does have SportVu and MU has access to that, but I am not positive about Kasten. I had an interview lined up with the staff 2 years ago, but it fell through and I got lazy AF, so I never rescheduled. While SportVu doesn't cover +/- per se, it does provide a lot of spacing data, which can be translated to lineup info.

For example, MU's death lineup (Rowsey/Howard/Elliott/Cain/Hauser) was a -7 in 3:54, giving up 17 points and scoring 10 in 7 possessions. Watching those 4 minutes, X feasted with no shot blocker down low. Hauser/Cain did not provide any help on drivers, so it was either a layup, dunk or kickout for an open 3. SportVU would provide more tangible evidence on distance to nearest help defender, distance traveled by driver, speed on rotations. All that jazz.