Celtics' Isaiah Thomas is a fourth-quarter killer, but which team is he killing?

[Denis998]

.

[Vermont Green]

I am shocked that the Celtic's defensive rating in the 4th Qtr. with IT on the court is so bad.  IT has surprised me at how good of a primary scorer he has become, maybe he will surprise me and figure out a way to improve on defense.  There is no reason to sugar coat this.  It is bad.  We won't win in the playoffs without better defense (including rebounding).

[Denis998]

BUMP...because I added a bunch of words..

This whole problem is overblown. No one seems to wonder why the Defensive rating is so low with it on the court. Let's start with out NUMBER ONE issue which is rebounding. Our bigs are horrendous at rebounding the ball, and guess who has to pick up the slack for it? The guards. So to start off games our lineups included IT AB Jae Amir and Horford, really weak defensive rebounding team thanks to our bigs. Now what does that mean? It means that for every extra rebound the other team gets, be it offensive or defensive, it directly increases opposing teams offensive rating and decreases our defensive rating. That's really a given, more shot opertunities means more opportunities to score thus a better rating.

Now that we have established that our bigs are horrible at rebounding, let's see how that in turn effects IT. Due to this issue, what has been said time and time again on what the team needs to do in order to negate this issue? Well, what BS always says it has to be a team effort, everyone on the floor needs to go for the defensive board. AB is a clear example of this and his team high 6.9 boards a game. But guess what? IT is only 5'9, the only rebounds hes going to get are the ones that just happen to fall in his hands. As I said before, the team relies on 5 people to rebound the ball on any given play, but having IT out there just means that they have one less player out there to rebound, and thus increasing the rebounding disparity and boosting the other teams offensive efforts.

Now on to the next issue, the whole on/off the court statistics. Lets take a look as to what the line up turns into when he goes to the bench for the first time in the game. Usually (up until BS changed the starting lineup), IT and Amir would be the first guys to get a rest, and guess what? Those two guys are our worst rebounders in the starting lineup. And who do they get replaced with you ask? Usually it two of Smart, Jaylen, and KO, all of which are good rebounders. You might say KO isn't really that great of a rebounder, but he is second on the team in defensive rebound percentage (first being AB). So basically our defensive improves because the rebounding improves quite a bit compared to the starting lineup, which ultimately gives the other team less offensive opportunities, thus lowering their offensive efficiency and increasing our defensive efficiency.

So basically, its not that IT is a horrid defender, it is more so that our BIGS are horrendous at rebounding.

Im including a stat which shows that when IT is guarding his opponent they are shooting only 1.9% better, which at the end out the day the 1.9% isnt going to lead to the horrid D.EFF stats that everyone is talking about today.
http://stats.nba.com/players/defense-dash-overall/#!?Season=2016-17&SeasonType=Regular%20Season&TeamID=1610612738&sort=PCT_PLUSMINUS&dir=-1

[mctyson]

I think having Jerebko and Horford down low will help defensively.

[mmmmm]

These sorts of analysis have to be very careful.   The sample sizes involved cutting down a half season into just the 4th period means you are cutting a very small sample -- one that is normally considered way too small for statistical significance when doing plus/minus analysis -- into an even tinier one.

I would also caution that with plus-minus data individual and two-man analysis is often extremely misleading.  Collinearity effects can be dramatic, especially when some players play a ton of minutes together while others hardly ever play together.   A good example of this was last year when some 'analysts' decided to look at 2-man data that seemed to show that everybody played so much better without David Lee on the floor.  What their analysis didn't show was that the minutes people played without David Lee included large shares of time WITH Isaiah on the floor last year whereas David Lee himself played most of his minutes WITHOUT Isaiah on the floor.   

That presented a false result that it was Lee who was dragging the plus/minus of others down when in truth it was absence of time with IT.

Many fans here were surprised when Lee joined Dallas for the rest of last year and San Antonio this year and has posted very, very good numbers in each stint -- including excellent plus/minus numbers!   But folks who looked past the simplistic 2-man analysis were not surprised.

When using plus/minus data, it is much better to look at actual 5-man units.   These let you get a better picture of the actual configurations that are working and are not working.

One picture that materializes is that the three-guard problem is a real one - but especially when it involves using only one true big.

For example, the Isaiah+Avery+Marcus+Jae+Al 5-man unit has a truly horrid 127 defensive rating.

Every one of those players also has plenty of minutes on _other_ 5-man units that have fantastic defensive ratings.   For example, just swap Marcus for Jonas and you get a unit with a fantastic 97 DRtg!!!

Alternatively, however, if you instead swap out Crowder and put in Olynyk, (so you have a lineup of IT+AB+MS+AH+KO), that _also_ drops the DRtg significantly, down to 103.8, which is still excellent.

Neither of those swaps mean that Marcus or Jae are bad defenders.  Or that Jonas or Kelly are elite defenders capable of shaving dozens of points off the DRtg.

The p----poor rating of the first 3-guard unit up above doesn't mean _any_ of those 5 players are bad defensive players.   It means that it was a horrible defensive _unit_.

The alternative units with either Jonas or Kelly are far superior defensive units because they are better utilizations of the players involved.

Marcus at the 3 and/or Jae at the 4 are things that you can get away with situationally, but overall, they just aren't good defensive alignments.

We have 8 different 5-man units with at least 50 minutes on the floor that include Thomas.  Only two of those units -- the one noted above plus another made up of IT+MS+JC+AH+KO -- have 'bad' DRtgs (both are above 120).   

The other 5 most-used lineups with Thomas all are at or better than the team's overall 109.8 rating and go as low as the stifling 97 rating of the lineup with JJ mentioned above.

Some other things that come out of the analysis of 5-man data is that the vast majority of our 'better' 5-man units (both defensively and overall net rating) include at least 2 of our three best bigs (Al, Kelly & Amir) and the few good ones that don't include two from that set tend to include Al & Jonas.   

We have very few 5-man units with only one big that aren't in the negative, whereas almost all of our 5-man units with two true bigs are in the positive.

To me, all this suggests strongly this isn't an individual player issue.   This is a configuration & utilization issue.

[Celtics4ever]

What is our record in these games? There goes the analysis and we would be crap without IT4.

[TrueFan]

What is our record in these games? There goes the analysis and we would be crap without IT4.

Exactly.

So much IT4 hate today.

[jdz101]

I believe the way to solve our defensive woes is to get a center that can help on dribble penetration, guard the pick and roll with length, and rebound the basketball. We give away far too many easy basic drives and shots when Thomas and Bradley inevitably get hung up on that first screen. Boston has been missing that 7 foot defensive general ever since KG was traded and it's been the very obvious achilles heel since then. I'd love for rebounding centers to stop drooling when they see us on the schedule.

Thomas being so small gets so easily knocked off by screens on and off the ball. After that point we're helping and scrambling for the rest of the possession. If we had a big guy with length and mobility who can help on that sort of action or intimidate the drive, a lot of those defensive issues would be mitigated and we would become a much more difficult team to play.

[chilidawg]

I believe the way to solve our defensive woes is to get a center that can help on dribble penetration, guard the pick and roll with length, and rebound the basketball. We give away far too many easy basic drives and shots when Thomas and Bradley inevitably get hung up on that first screen. Boston has been missing that 7 foot defensive general ever since KG was traded and it's been the very obvious achilles heel since then. I'd love for rebounding centers to stop drooling when they see us on the schedule.

Thomas being so small gets so easily knocked off by screens on and off the ball. After that point we're helping and scrambling for the rest of the possession. If we had a big guy with length and mobility who can help on that sort of action or intimidate the drive, a lot of those defensive issues would be mitigated and we would become a much more difficult team to play.

So Noel? Or did you have somebody else in mind?  That doesn't seem to describe Bogut, Monroe, Vucevic, Chandler or anyone else I can think of who has been rumored to be available.

[rickyfan3.0...]

We have ALL been seeing this, most of us probably didn't realize how bad it actually is. Eye opening.

[celticsclay]

I think we just need a better defensive center to help anchor the defense. This would also raise the question of if IT is so bad for us why do we have the 5th best record in the conference and why were we abysmal when he was injured?

[chilidawg]

In reading the article again, it seems to say that going small in the fourth quarter doesn't work.
Which I used to say when Doc Rivers did it all the time.

I seem to remember our 4th Quarter lineup during the championship season being Garnett, Posey, Pierce, Allen and Rondo.  That seemed to work out pretty well.

[chilidawg]

What is our record in these games? There goes the analysis and we would be crap without IT4.

Exactly.

So much IT4 hate today.

Doesn't seem like hate so much as a reasonable attempt at understanding what's going on.  Unless by hate you mean anything that isn't "IT is the MVP of the league"

[CelticPride2016]

What is our record in these games? There goes the analysis and we would be crap without IT4.

That is true, and there's no reason to hate on him. I'm not sure anyone is. Someone below Celtics4ever is defining some posts as hatred. I don't see it.

There are a lot of big what ifs that haven't played out yet.

What if we draft the phenom pg?

I mean, this is Isaiah's team for now and perhaps his whole career.

But at some point, you can't keep them all.

We lost Evan Turner because of the structure of the league. How much will Bradley want, $25 million?

I'd rather have a lot of quality players than be stuck with two or three Brinks Truck guys making that impossible.

There is only so much money. Isaiah might not be worth $30 million. The truth hurts. It takes a team effort to win and management with vision to avoid trajectory sinkholes. Golden State looks invincible, but give it a year or two when they end up whining like LeBron.

I don't think NBA teams can just spend money like George Steinbrenner without severe penalties.  I'm not sure, but there doesn't seem to be a hard cap, only steep luxury taxes. So that's what's going on in Cleveland. And we saw what happened to Miami, and to Golden State, I'll just say that nothing lasts forever.

[greece66]

I think we just need a better defensive center to help anchor the defense. This would also raise the question of if IT is so bad for us why do we have the 5th best record in the conference and why were we abysmal when he was injured?

I agree with these points celticsclay.

But to keep it fair to the article, Moore had anticipated them. First, Moore is not saying IT is bad for us in the absolute sense.

First, it's never one thing when the metrics are this messy. It's not like every single bucket scored is because Thomas is unable or unwilling to defend, and that's important in determining where to go from here. Here's what this data doesn't tell us:

It doesn't tell us that Isaiah Thomas is a bad NBA player who should be benched.

It doesn't tell us that the Celtics should bench Isaiah Thomas in the fourth.

It doesn't tell us that Thomas is the reason they are bad defensively.

And second, he makes it clear this is about the postseason

The Celtics are undefeated this season when leading going into the 4th quarter, and are 30-18 overall. They're winning anyway. There's a philosophy that as long as the Celtics are winning games, it's fine. They have the 2nd best record in the East, and a win over the Raptors Wednesday would give them a full two-game lead for 2nd. Does any of this matter?

Yes. It does. This isn't about whether the Celtics can beat the Pistons or Kings or Dallas in January. It's whether their defense will be good enough to hold off teams in the playoffs. Boston's 8-12 record vs. teams .500 or better is telling, as is their 1-8 record vs. the Warriors, Spurs, Rockets, Cavs, and Raptors. Boston can hold off teams with their offense in the regular season. They have to find a better formula for the postseason.