Clutch Celtics: The Big 4 in Crunch Time

by drza44 on Mar 9, 2011 12:12 AM EST

There has been a big emphasis in the basketball public lately upon what happens in crunch time. There was a buzz last month about Kobe Bryant's crunch time myth, started by Henry Abbot at TrueHoop. Essentially, he pointed out the numbers in the clutch in recent years, showing that Kobe's "late game assassin" reputation may be overstated. Now, this week, everyone is all over LeBron James and the Miami Heat for how they have failed to produce in crunch time of late. The interesting thing is that LeBron's clutch numbers actually look relatively good this year, but he's failed very visibly in some very marquis games and the 1-for-19 (or whatever the exact number is) that the Heat are shooting in the last 10-seconds of one possession games looks pretty terrible.

Anyway, those are Lakers and Heat issues. In the words of Rajon Rondo, I can't relate.

But all of this talk has got me thinking again about how the Celtics perform in the crunch. With Paul Pierce and Ray Allen, the Celtics have two players historically known as great closers. Kevin Garnett's clutch reputation is a bit shakier, though those that have followed his career closely know that he's actually been one of the better crunch time bigs in the league, very comparable to rival Tim Duncan late in games. And Rajon Rondo is a lightening rod, even among Celtics fans. He has his supporters and his detractors, but one of the main criticisms he faces is that his poor free throw and jump shooting make him a crunch time liability.

So, I decided to take a look at how the Pierce, Allen, Garnett and Rondo have performed in crunch time during the almost 4 years that this Celtic squad has been together. For today, I'll focus on the clutch numbers from 82games.com (defined as the last 5 minutes of a game within 5 points). Here are the totals since October of 2007, per 48 minutes played:

	pts	FGM	FGA	3s	FG%	EFG%	FTM	FTA	FT%	Oreb	Dreb	Reb	asts	TOs	blks
Garnett	21.9	8.1	17.3	0.0	46.6	46.6	5.7	7.3	78.3	2.5	9.9	12.4	3.3	2.3	1.8
Pierce	28.9	7.9	20.2	1.5	39.3	43.0	11.6	13.7	84.3	1.0	5.0	6.0	5.7	2.6	0.6
Allen	25.9	7.4	17.2	4.3	43.1	55.5	6.8	7.7	89.2	1.2	4.5	5.7	1.4	1.8	0.2
Rondo	15.2	6.4	13.8	0.0	46.1	46.1	2.5	3.9	65.1	2.5	5.2	7.7	8.1	3.1	0.1

Now, here are the per-48 minute clutch stats for the 2010-11 season thus far:

	pts	FGM	FGA	3s	FG%	EFG%	FTM	FTA	FT%	Oreb	Dreb	Reb	asts	TOs	blks
Garnett	21.0	8.0	17.2	0.0	46.3	46.3	5.0	5.9	85.7	2.9	10.1	13.0	2.5	2.9	0.8
Pierce	28.8	8.8	19.3	1.5	45.3	49.1	9.8	10.9	90.0	0.4	5.8	6.2	6.2	1.5	1.1
Allen	31.0	8.7	19.7	6.2	44.4	60.2	7.3	8.4	87.0	1.5	4.7	6.2	1.1	1.5	0.4
Rondo	10.7	4.9	13.4	0.0	36.7	36.7	0.9	1.3	66.7	2.7	3.1	5.8	14.3	3.6	0.0

Here are some of the things I take away from these numbers:

First, just like over the rest of the game, in crunch time the Celtics still preach Ubuntu. Unlike the Heat or the Lakers, the Celtics' crunch time offense doesn't consist of giving one player the ball and getting out of the way. Instead, our main players maintain roughly the same roles as they always have. The scoring is relatively equal opportunity, with Pierce and Allen scoring a bit more than Garnett and even Rondo getting his shots as well. Pierce draws the fouls best, Ray is nails from downtown. Rondo is the main distributor, though Pierce and even Garnett get their fair share of crunch time assists. Garnett is the rebounder and help defender, though Rondo, Pierce and Allen all do a good job crashing the board for their positions. All in all, in crunch time like the rest of the game, this Celtics team knows their roles and performs them at a high level.

Looking at this year, it's remarkable how similar most of the numbers are when compared to the previous four years. This team is essentially locked in...in the clutch, they know what to do. Pierce and Allen are shooting a bit better this year, KG maybe crashing the boards harder, but really...they are who they are at this point. I was most interested in Rondo, though, as my impression would be that he's taken a larger role on the team this year than in years past. Interestingly, while he's getting a similar number of shots he seems to be really struggling from the field in the clutch this year as opposed to in years past. On the other hand, he's taken a much larger role as a clutch distributor this year, with over 14 assists/48 min (almost twice his average over the past four years).

Anyway, this was just a quick snapshot. Make of these numbers what you will, but they suggest to me that this Celtics squad isn't about any one late-game assassin. There's no one "closer" on this team. Instead, we've got 4 high-caliber options that specialize in what they do best. In closing time we don't have to put the burden on any one person's shoulders...instead, we've got a clutch TEAM. On the whole, seems like that's a much better position to be in.

Be respectful and keep it clean. Thanks.

Tweet Comment 15 comments | 0 recs |

Do you like this story?

Comments

Another great piece, drza44.

The results here are not surprising.

It would be useful to also include a block with the ‘non-clutch’ (i.e. normal) per-48 numbers so one could see how the team’s play changes from the regular course of the game to when they are in clutch time.

The opponent who scares me the most for closing a game is Dirk Nowitski. His closing numbers are pretty impressive and I wasn’t surprised when he was key in stealing both of Dallas’ victories from the Celtics.

"We will take it easy, and walk on down our road at our pace with our focus ahead, and your focus on us." - Warrior Spirit, on what Celtics' Ubuntu is.

by mmmmm on Mar 9, 2011 10:13 AM EST reply actions

Good point on the normal numbers...

I started to do that. ElGee does it in one of the articles that I linked. But I decided that my point wasn’t so much about “who steps up more?”, more an indication of the style that the Celtics play in the clutch. When I come from that direction, it doesn’t really matter so much (for the sake of my point) what the Cs do in normal situations…all that matters is their distribution of labor in crunch time.

That said, at some point I still may make the kind of normal vs crunch comparison that you mention. Another direction I could take is to compare the Cs crunch-time numbers to other teams, looking at both overall quality of play as well as distribution of labor.

by drza44 on Mar 9, 2011 10:42 AM EST reply actions

My suspicion is that

the Celtics’ play doesn’t really change much in clutch time.

Without checking, those numbers don’t look dramatically different from their overall numbers to me.

If anything, that’s probably a good thing.

"We will take it easy, and walk on down our road at our pace with our focus ahead, and your focus on us." - Warrior Spirit, on what Celtics' Ubuntu is.

by mmmmm on Mar 9, 2011 10:55 AM EST up reply actions

Link for the stats

Where did you get all these? 82games.com?

Reporter: About a year ago I asked about the Celtics-Lakers thing and you said you'd really only seen it on TV. Now you've been through it, can you talk about playing the Boston Celtics for the NBA championship?
Kobe: It Sucks.

by Tom Bellinger on Mar 9, 2011 10:56 AM EST reply actions

nevermind

I must have missed it the first time, I see it now.

by Tom Bellinger on Mar 9, 2011 10:58 AM EST up reply actions

I used 82games.com stats...

…but I did have to do some calculations of my own to get the composite numbers for the last 4 years. 82games has a ‘clutch’ link for each player, but in that link they don’t have all of the raw numbers. But they do have raw shots taken, free throws taken/made, per-48 min scoring average, FG%, EFG%, and percent of crunch time minutes played. From that, I could do some algebra to come up with raw points scored and minutes played, and from there I could convert all of the other per-48 minute stats they keep into raw numbers. From there it got a lot simpler.

Because of all of this there may be some rounding errors in what I posted (i.e. I posted KG’s per-48 clutch scoring average as 21.0 ppg for this year, but 82games actually has him at 21.1 ppg. Stuff like that). On the whole, though, these numbers should be pretty accurate.

by drza44 on Mar 9, 2011 11:19 AM EST up reply actions

Even with PP and KG getting those assists

Rondo leads the league in clutch tim assists/48, and he’s also tied for 3rd in steals.

by BballTim on Mar 9, 2011 11:41 AM EST reply actions

For some reason the player cards...

…don’t include the steals stat, which is why I don’t have it listed here. But it’s cool to see that Rondo (and Garnett) are both among the league leaders there this year. Again, this would fit with my intuition…the Cs essentially continue to play their game through most clutch situations. Rondo is leading the league in assists and steals in normal times, and he continues to do so in the clutch.

by drza44 on Mar 9, 2011 11:54 AM EST up reply actions

Awesome post

Trying to further find a story in the numbers, it looks like much less reliance on the Pierce iso and more of the usual offensive sets. 16.7 assists for the three years and 24 this season per the 48 minutes. 8.9 turnovers for the 3 years compared to 12.5 turnovers this year(I may have screwed those up, haha). Of course, this could just be a case of small sample sizes, but all of the Celtics’ big 4 are over 100 minutes in the crunch time for this season.

I’ve been wondering if the league insistence on isos in crunch time is one of those things where coaches keep doing it because “that’s how we’ve always done it” even if it statistically performs poorer. It does obviously have it’s advantages since you can get the clock to exactly where you need to take the shot and help defenders can’t poke the ball away. I love what this team has evolved to under Doc.

by jdpapa3 on Mar 9, 2011 12:17 PM EST reply actions

ball security and clock management

those indeed are two very, very valid reasons for running an ‘iso’ in the closing possessions.

The ‘clutch time’ definition that drza44 is using is the broader ‘last 5 minutes, with the score within 5 points’ definition that 82games.com uses.

Thus it includes a lot more possessions where full-clock usage is not key – and in fact if you are trailing not even desired. Possession scoring efficiciency will be the desired factor, independent of time.

In the ‘last possession’ scenario, the Celtics still go predominately to a Piece ISO and for very good reasons. The ball control and clock management are two. Statistically also, is that Pierce is the most likely Celtic to generate points – whether through making a shot or drawing a foul – with the ball in his hands.

Even on an ‘iso’ play, the Celtics still run actual plays. Everyone on the floor has a role to help create the shot for Pierce or provide an alternative depending on what the defense does. Watch when Pierce sets up how everyone moves to specific locations (and sometime keep moving) and then when Pierce makes his first move and the defense reacts, everyone moves accordingly to either clear space for Pierce, draw a defender out of the way or get clear for a catch & shoot.

What is different this year over prior years is that Pierce is far more likely to pass out of the iso when the defense collapses on him, even in the last second. We’ve clearly seen that many times this year with last-second heroics by KG & Ray off assists from Paul.

"We will take it easy, and walk on down our road at our pace with our focus ahead, and your focus on us." - Warrior Spirit, on what Celtics' Ubuntu is.

by mmmmm on Mar 9, 2011 1:40 PM EST up reply actions

Not sure where you got the turnover numbers

It’s 9.8 for the big 4 over the last 4 years. 9.5 for this year. Not really any difference.

by phi7 on Mar 9, 2011 3:06 PM EST up reply actions

2008 and comparing to other teams

Cool post drza. I believe the Celtics struggled in the clutch in 2008 though — at least the big 3. That might skew the numbers a bit: I wonder if the overall snapshot from 09-11 is even better/more consistent.

Cheating a little, we can infer team performance from high-minute +/- numbers: http://elgee35.files.wordpress.com/2011/03/clutch-pm-by-team.jpg

Of course, it might be better just to use ORtg since we are primarily looking at offensive performance.

http://elgee35.wordpress.com/

by ElGee35 on Mar 9, 2011 4:40 PM EST reply actions

Thanks ElGee. And about 2008...

A glance at my spreadsheet suggests that Pierce had a poor year shooting in the clutch in ’08 (his worst of the last 4 years), and KG was slightly below his average. Ray and Rondo, though, had good clutch shooting years in ’08. Then, in ’09 Pierce and KG had by far their best clutch year. Then, in ’10 they struggled again, and KG also struggled in other areas like rebounding (not surprising). Then, in this year they bounced back again.

All told, it doesn’t look like there is any kind of consistent pattern in their clutch numbers outside of what one might infer based on circumstances. In ’08 they were learning each other, so they struggled a bit (relatively speaking) late. In ’09 they were locked and confident, and everyone was able to play great in the clutch. Last year injuries really derailed them, and they gave up a bunch of leads late, so not surprising that the crunch time numbers suffered. Then, this year, there has been more health so the numbers in the crunch are creeping up again.

by drza44 on Mar 9, 2011 4:54 PM EST up reply actions

The Problem

Rondo too many turnovers and Pierce has the same problem trying to take over with the three pointers when he should be driving in and drawing a foul ….. Also way to many substitutions to early in the game …… Murphy should be allowed to play with the first team awhile alonger and Green the same thing not enough time to get going don’t forget
both were start before………. Rondo way too many dumb passes and not enough brains ..

by juanpan on Mar 10, 2011 1:41 PM EST reply actions

With rising assist numbers, you can't help but get higher TO numbers.

The more you pas the ball, the more likely a TO will happen. No getting around that.

Rondo’s “hands” rating, which factors in bad passes as part of a general ball-handling rating, is 28.7. Among various top PGs this puts him:

Nash 32.9
Rondo 28.7
Paul 28.7
DWill 27.4
Harris 24.1
Rose 23.6

Source: 82games.com

In that score, while Rondo does have a higher number of bad passes than Paul, he has fewer dribbling, offensive foul and other TO marks. Thus the two end up tied in overall rating.

"We will take it easy, and walk on down our road at our pace with our focus ahead, and your focus on us." - Warrior Spirit, on what Celtics' Ubuntu is.

by mmmmm on Mar 10, 2011 7:10 PM EST reply actions