Brad Stevens assessment at first season halfway mark

[Crow]

This thread is intended to gather information about his performance.

I have just started a review.

Among the first things I see most are negative :

Celtics 25th in 3PTA rate. 24th on FTrate. Above average on attempts from all mid-range distances. 25th on assists. 25th on TO%.

8 of top 9 most used lineups have negative +/-.

24th on offensive rating. Bottom 10 on 6 of the Four Factors *2. Only above 15th on one (opponent eFG%).

5th most opponent shots within 5 feet.

4th highest foul rate.

9 straight losses and 14 of 17.

4-18 against best 16 teams.


Any signs of advanced analytic points of success?

Slightly above average own shots within 5 feet.

6th least 3 pt attempts by opponents.


I know the real test is in the future. And probably not next year. But there are fewer positive signs than I expected and few signs of positive impact from advanced analytics either. Most disappointing is that the Celtic team shot distribution remains as dreadful (to my way of thinking) as it has been for years.

[AcrossTheCourt]

A lot of this is personnel. They have one of the sorriest looking rotations in the league.

Not high in three-pointer attempted rate? Again, personnel. Brad Stevens urged Jared Sullinger to become a three-point shooter, and he has -- from a 0.022 3PAr to 0.214. He turned notorious shot-selection Jordan Crawford into a point guard. His two highest scorers, Jeff Green and Avery Bradley, are not good outside shooters. Brandon Bass is a midrange shooter; that's just who he is. There's no way to make this a good, competitive team, and I'm sure they'd be even less of a three-point shooting team with a more conventional coach.

[Mike G]

Our 11 prognosticators (Predictions thread) gave the Celts an avg of 27 wins for the year. They're on pace for 32.
ESPN crew figured them for 26 wins.
Only 7 teams are faring as much better than forecast.

[Crow]

With a .333 overall win percentage, they would be appear to be on pace for 25 wins or less than the consensus. Or are you using something else to predict 32?

After going 4-14 recently and trading away some talent, I'd guess that is more likely that they will do worse than what their overall win percentage suggests than equal or better.

They have underperformed their Pythagorean expected wins by 1 game so far.

[italia13calcio]

I believe he's getting the 32 wins from the b-r projections http://www.basketball-reference.com/fri ... f_prob.cgi

[Crow]

I see, thanks. I should have checked that but was in a rush.

I still think they will probably finish with under or well under 30 wins. But I guess they have one of the easiest schedules remaining.

[Crow]

I forgot about Rondo til now. That would account for much of the expectations gap. But I am still less optimistic than the B-R projection.

[Mike G]

I think the b-r.com forecast is based on performance to date, future schedule, etc. Not on lineup changes or any other specific.
Celts now expected to go 19-23 the rest of the way; 32.8 wins projected; 28% chance of making the playoffs.
They've had the toughest schedule in the East, so far. So that does indeed suggest an easy road the rest of the way.
And that assumes Eastern teams will continue to generally suck.

Their SRS is -3.52, which is 21st in the league. In a balanced league, they should have won 39% of their games; but they've won 35%.

[Crow]

... He turned notorious shot-selection Jordan Crawford into a point guard. ...

On the other hand, one could also say that Rondo's injury and Avery Bradley's failure as starting PG the first four losing games of the season drove the move to turn Crawford into a PG. His shots per 36 minutes actually nudged up slightly this season, instead of any decline. His eFG% barely budged and though more 3 pointers were taken in more minutes his rate of taking them did not increase at all. His assist rate did go up; it basically had to on this team playing that role. Not that this is sufficient / reliable information but his raw on court +/- fell by 4 points from the previous season and his on/off fell by 2 points. The team won 30% of his time on the court. They won more when he was off. With the trade, apparently the team did not strongly want to continue with that experiment.


"There's no way to make this a good, competitive team"
I don't know about that. They probably were going to be sub .500 but maybe they could have been somewhat better than .350 win% with different lineups and coaching strategy. Right now the 'best college coach in the country' (according to Ainge) has the Celtics in 25th place. A good number of NBA assistants (with previous NBA head coaching experience or not) could probably have taken them to this level (or maybe better), so far. Time will of course give more information to evaluate. This is just a first look.

[xkonk]

Celtics 25th in 3PTA rate. 24th on FTrate. Above average on attempts from all mid-range distances. 25th on assists. 25th on TO%.

8 of top 9 most used lineups have negative +/-.

24th on offensive rating. Bottom 10 on 6 of the Four Factors *2. Only above 15th on one (opponent eFG%).

5th most opponent shots within 5 feet.

4th highest foul rate.

9 straight losses and 14 of 17.

4-18 against best 16 teams.


Any signs of advanced analytic points of success?

Slightly above average own shots within 5 feet.

6th least 3 pt attempts by opponents.

You obviously expect the team to perform differently with the trades they made, but it still seems important to compare to where they were last year. Some of the measures you mentioned:
24th in 3PA rate, 16th in FTA rate, 12th on eFG%, 23rd on TOV%, last in ORB%, 14th in FTA/FGA, 6th, 11th, 20th, and 23rd on the opponent equivalents. 24th on ORtg and 7th on DRtg. SRS of -.62.

So after losing a number of players, and arguably some of their best ones, their ORtg is in the same place (albeit a bad place), their DRtg has slipped from good to average, and their SRS has plummeted. Would there have been more of a decline without Stevens? Less of a decline? How would we know?

[Mike G]

A few days old now, but my most recent 'improved' list shows Celtics mostly better than last year:

Code: Select all

eW+    per36 rates    Eff%   Sco   Reb   Ast   e484  2013 Eff%   Sco   Reb   Ast  e484
.82   Bradley,Avery  .507   17.0   5.2   1.5    .73      .459   12.0   3.0   2.5   .34
.68   Bass,Brandon   .525   14.6   8.2   1.6   1.04      .528   12.1   7.2   1.3   .70
.53  Sullinger,Jared .505   18.8  10.2   2.3   1.31      .526   11.2  10.9   1.3  1.00
.42  Crawford,Jordan .528   17.5   4.0   6.2   1.12      .506   17.9   4.4   4.4   .92
.30   Humphries,Kris .569   15.0  10.7   1.8   1.31  Brk .505   11.6  11.6   1.0   .98

.23   Lee,Courtney   .570   16.6   3.4   2.1    .74      .541   12.1   3.7   2.5   .52
.03   Brooks,Marshon .537   17.4   8.0   1.6    .92  Brk .516   15.5   4.1   3.0   .69
-.13  Green,Jeff     .537   18.7   5.3   1.6    .88      .552   17.8   5.3   1.9   .94
-.66  Wallace,Gerald .527    7.1   5.3   3.6    .25  Brk .480    9.8   6.1   3.4   .69

Lee is out, and Bayless is in, but Lee did add more than expected wins.
This group in total created 2.23 'excess' wins in the first 33 games.

[Crow]

xkonk: "Would there have been more of a decline without Stevens? Less of a decline? How would we know?"

Can't say with confidence, especially right now. One man in that chair at a time. Eventually we might be able to get a coaching APM estimate of his impact. 3, 5 or more years from now. The rating will be more accurate if / when someone else takes the chair with a good number of the same guys.



Comparing player ratings before and with Stevens may be suggestive.

Mike, the left columns are 2013-14 and the right ones are 2012-13? Almost all got better on scoring, most got worse on rebounding and it is was close to half and half on assists.

I find Green and Bradley's assist rates unacceptable, especially for this team's situation. Bradley has shown some month to month improvement byt Green has regressed. If you are not tanking, I find Gerald Wallace's FGA rate way too low given his team best eFG% (if you ignore Bogans). Humphries probably deserves more than 19 minutes per game. With the worst raw on court team +/-s and on / offs one or more of Bass, Olynyk and Faverani should probably be played less or just used differently. Encouraging Sullinger to take 3 point shots might or might not pay off in future but it hasn't yet. The payoff is similar to a mid-range shot, the worst in the game. His overall jump shot frequency has jumped from 44 to 56% of his shots. Not a trend I would encourage. Bradley may be better than his worst season in 2012-13 but he is still well below 2011-12 and still well above league average on winshares per 48 minutes. It is still early but right now Bayless is about as bad as he has ever been.

Are they tanking? That would change the story. They have moved out or away 4 of the top 7 players on raw on/off +/= and are using the very best rated (Humphries) fairly lightly. I find that somewhat suspicious.

[Mike G]

updated Celts' improvement table

Code: Select all

eW+   2014   per36    Eff%   Sco   Reb   Ast   e484  2013 Eff%   Sco   Reb   Ast  e484
1.15  Bradley,Avery  .500   17.3   5.1   1.6    .77      .459   12.0   3.0   2.5   .34
.69  Sullinger,Jared .500   18.3  10.9   2.1   1.33      .526   11.2  10.9   1.3  1.00
.62   Bass,Brandon   .519   14.4   8.1   1.5    .96      .528   12.1   7.2   1.3   .70
.40   Humphries,Kris .561   14.1  10.7   1.8   1.29  Brk .505   11.6  11.6   1.0   .98
.40  Crawford,Jordan .518   17.2   3.7   6.2   1.08      .506   17.9   4.4   4.4   .92

.14   Lee,Courtney   .570   15.7   3.2   2.0    .65      .541   12.1   3.7   2.5   .52
-.30  Green,Jeff     .522   17.7   5.7   1.5    .83      .552   17.8   5.3   1.9   .94
-.88  Wallace,Gerald .535    7.0   5.2   3.3    .22  Brk .480    9.8   6.1   3.4   .69

Bradley gives up 1 assist (per 36), gains 7 pts and 2 reb
Sullinger gains an assist and 7 pts
Bass picks up 2 pts and a reb
Humphries loses one reb, gains 2.5 pts and an ast
Crawford lost half a pt and reb, gained 2 ast
Lee gained 3.6 pts vs losing half a reb and ast

Not all gains and losses are equal.

[Crow]

Things to watch:

Rondo-Bradley on court together last season was bad from a team +/- perspective (around -5 per 48 minutes). First game together this season the +/- was -15 in 15 minutes. Will this combination work? Will Stevens make it in particular better? What kind of players fit his system and strategy? Is it consist with the way Ainge has stocked the team in the past or more importantly will it be the same going forward?

Rondo-Sullinger was slightly negative on raw +/- last season and was moderately negative last night. Last season Sullinger shot 55% without Rondo and only 44% with. Last night with Rondo, Sullinger shot 40%. Without 44%.

Rondo-Green was slightly negative last season and terrible in first game last night.

Rondo-Bass was moderately weak last season. It wasn't used last night.

Which of these pairs are even "promising to be good eventually under Stevens" building blocks for team success?

[Crow]

I now see that Humphries minutes are up over the last 10 games. Faverani's are down with a DNP last night.

Starting lineup had a good game but lost 8 points when it played a few minutes without Humphries.

Both Bradley and Rondo shot below 25% on the court together. Bradley hit 50% when Rondo was off.

Lost another game by 3 pts or less. 24th in league on that stat. 28th on road overall.

Jeff Green's FGAs per game are up 20% compared to December. His FG% are down considerably. How much freedom do you let him have to take extra shots? Same story with Bradley. Sullinger's overall FG% has been sinking month to month to under 39% in January. He trimmed one shot per game. Might be good to trim 1-3 more until he can get back close to or above 45%.

Pressey is shooting 19% for the month but has gotten 20+ minutes in last 3 games. Negative winshares per 48.

Tanking or poor management remains a key question. I'd think that by the end of the season there will likely be a pretty strong case that it is one or the other but maybe not clear which.

Really early but Rondo's first 2 games give him a ws/48 that is 3rd worst in league for guys playing 20+ minutes per game. Long climb back to average (what he barely was last season) and then All-Star level (only one season above .160 in his career, back in 2008-9). Do they build around him? RAPM last season was barely above average and well below his past level. What will it be this season and next without the stars that used to surround him and bring out the best in him?