Push the Pace

One of the odder NBA dynamics around the end of last season was that lots of people were simultaneously talking about the unique awesomeness of Steve Nash and advocating that other teams adopt a Phoenix-style high-paced offense. Seemingly, though, those can't both be right. Either Nash is a brilliant talent, capable of quarterbacking an unorthodox style of play effectively, or else the unorthodox style is just better and anyone would put up Suns-esque numbers by pushing the pace.

Well . . . score one for Nash, even though the Suns aren't doing very well. John Hollinger points out that the Denver Nuggets have managed to actually exceed Phoenix's pace this year and for their trouble they rank 22nd in offensive efficiency. Meanwhile "this is still better than Toronto and Charlotte, which rank third and fourth in pace but 23rd and 28th in offensive efficiency." Boston, I might add, is fifth in pace and 25th in efficiency. Phoenix itself, however, is managing a respectable sixth in efficiency while maintaining the league's second-fastest offense. It's not good enough, in other words, to just run; you need players who can actually make it work. Charlotte, in particular, is just turning the ball over like crazy.

Comments

Either Nash is a brilliant talent, capable of quarterbacking an unorthodox style of play effectively, or else the unorthodox style is just better and anyone would put up Suns-esque numbers by pushing the pace.

It would be funny if every NFL team learned from the Atlants Falcons that the thing you want to do is run a super-simple run/pass option offense that lets your team take advantage of Mike Vick's athleticism (even though he plays for the Atlanta Falcons).

Posted by dj moonbat | November 14, 2006 2:12 PM

Well, I for one am glad not all of your basketblogging has gone over to DCist. I would point out that about every couple of years or so since 1990, the Nuggets announce they are going to return to their run-and-gun roots -- to "take advantage of the elevation" even though highly fit NBA athletes are least likely to feel the adverse effects, and even though the Utah Jazz have done very well at only a slightly lower elevation with a slow style. I expect the Nuggets' current flirtation will last about a month, at which time George Karl will go back to an offensive style he is more used to coaching.

Posted by Colorado Luis | November 14, 2006 2:39 PM

It would be funny if every NFL team learned from the Atlants Falcons that the thing you want to do is run a super-simple run/pass option offense that lets your team take advantage of Mike Vick's athleticism (even though he plays for the Atlanta Falcons).

Put a sock in it, Easterbrook.

Posted by Pooh | November 14, 2006 2:56 PM

Charlotte, in particular, is just turning the ball over like crazy.

I think you've hit upon the problem of the pace metric. Teams that turn the ball over a lot tend to give up lots of transition points, which naturally increases the pace of the game. Plus, turn overs tend to end possessions early, also increasing the apparent pace of the game. The numbers only tell you so much about how "fast" a team actually plays.

FWIW, my guess is that, other things being equal, a good defensive team will appear to play slower than a bad one, because the opposition has to work harder to get shots. Long possessions = 'shorter' games.

Posted by Pooh | November 14, 2006 3:01 PM

FWIW, my guess is that, other things being equal, a good defensive team will appear to play slower than a bad one, because the opposition has to work harder to get shots.

Well, for some sense of "good defensive team" that's probably true. But in efficiency terms there's more (a lot more!) to good defense than making it hard for your opponent to get good looks. Things like not fouling the other team, forcing turnovers, and cleaning the defensive glass all make for good defense and all serve to increase the pace of the game.

Posted by Matthew Yglesias | November 14, 2006 3:13 PM

Good points, MY. I'm not familiar enough with how the pace metric is calculated to speculate on how well it takes those factors into account. Without empirical research either way, I'm still going to say that, in general good defense tends to slow the pace of the game. There are obviously exceptions - championship Bulls Mark I, Russell's Celtics, Showtime Lakers. But I'm going to guess that modern Spurs/Mavs, Stockton-Malone Jazz, Bad Boy Pistons are more representative of good defensive teams.

Posted by Pooh | November 14, 2006 3:18 PM

Things like not fouling the other team, forcing turnovers, and cleaning the defensive glass all make for good defense and all serve to increase the pace of the game.

Not all of those things are equally easy to affect, though. Making it harder to get good looks is probably the easiest thing to do. Also, transition defense is hard. Teams that run just are going to give up a lot of points.

Really, someone should test the offense/defense measures out against data from the 80s, and in particular, for the Showtime Lakers.

Posted by SomeCallMeTim | November 14, 2006 3:21 PM

SCMT: After a quick check in basketball-reference.com the results are interesting. The 80s Lakers had a relatively high pace (usually ranked between 6 and 10), a pretty good defense (almost always top 10 at least) and a scorching offense (ranked first or second most of the time).

Posted by Carlos | November 14, 2006 3:51 PM

Thinking this over, here's another wrinkle. There are really two distinctive elements to the Phoenix offense. One is the high pace and proclivity for running. The other is that they shoot three pointers like crazy. The point of the running isn't to get the traditional fast-break layup, but rather to generate transition treys. A lot of the other neo-speedy teams don't play like that at all.

Super-fast Denver, in particular, features almost comically poor three point shooting. It's possible, however, that the love of the three pointer is actually Mike D'Antoni's strategic breakthrough.

Watching games, it often seems to me that players don't understand that three points is a lot more than two points in the relevant sense; i.e. that the expected value of three point shot you have a 36 percent chance of making is bigger than that of a conventional bucket you have a 49 percent chance of making.

Posted by Matthew Yglesias | November 14, 2006 4:13 PM

as red would have told you, a fast-break offense requires a quarterback, appropriate athletes, and familiarity with the style. teams just implementing a fast-break style will almost always look sloppy about it at first (at least, that was true back in the days when most teams ran, although that was back before the three).

i of course, being an ancient and all, grew up with red holzman's knicks, who were pretty singular in being a non-running team in those days, and some of the battles between the heinsohn-coached fast-breaking cowens-hondo-jojo celts and the holzman-coached set-it-up frazier-debusschre-bradley knicks were quite fascinating in terms of which team would manage to assert its style of play. i'm probably fogging my memories, but i recall most of the regular-season games between the two were blowouts in one direction or another: either the celts would run it up or the knicks would slow it down....

Posted by howard | November 14, 2006 6:18 PM

your example shows an expected value of 1 either way (ok, just below 1 for a normal bucket). i think you'd need a bigger difference for this to matter (it may be that the suns have better shooters of 3s, which is why it works for them).

Posted by dj superflat | November 14, 2006 7:01 PM

.1 pt/shot is not an insignificant difference.

(3 * .36 = 1.08)
(2 * .49 = .98)

a 5 point swing will win you a fair few extra games. I think MY's comparison is a little superficial (there is some value to not having "empty" possessions, beyond the points, I think. Plus, if you shoot a ton of threes you either aren't rewarding your big guys, or you don't have big guys, etc...)

Posted by Pooh | November 14, 2006 7:26 PM

(...and just above one for a three.)

Carlos, I looked at basketball-reference, but they don't give their formula for defense there. They do say that it measures points per 100 possessions, but (a) I can imagine that the Lakers' offense became their defense--teams had to shoot faster because they needed to catch up, and (b) I remember that the Lakers' transition defense was excellent, while their half-court offense was average, and in a running game with a lot of possessions, small differences in defensive ability (broadly) will be increased. I'm wondering what the Lakers' defense looked like if you either (a) removed steals from the equation, or (b) limited your analysis to relatively low scoring games.

Posted by SomeCallMeTim | November 14, 2006 7:28 PM

it may be that the suns have better shooters of 3s, which is why it works for them

But it is also true that the style leads to better looks at the trey thus higher averages. A transition trey is usually an uncontested open look taken in rythm off a pass. If not, the shooter will soon be on the bench, because as a coach I would pull anyone that took a contested three on a break. Aslo defense tend to sag on Nash to prevent his penetration, leading to similar types of shots.

Posted by Tiparillo | November 14, 2006 7:34 PM

somecallmetim, by the metric (so to speak, since it's entirely subjective!) that matters to me - ability to shut down the other guys at critical times in the game - the early '80s Lakers played only fair defense, but the mid-late '80s Lakers were an excellent defensive team....

Posted by howard | November 14, 2006 7:54 PM

empty possessions

Missed threes also tend to lead to long rebounds which in turn lead to transition buckets. Thus, does the slightly higher expected value overcome the probability of giving up more easy buckets?

I tend to see threes as fine as long as they come in the flow of the offense or on the break. But teams can fall in love with the shot, to their detriment. Believe me, having watched the Warriors for many fruitless years, I have seen it first hand.

Posted by Tiparillo | November 14, 2006 7:55 PM

oh yes, i meant to say something about the lakers half-court offense: obviously matters changed as kareem aged, but in the early '80s, the half-court offense was well above average: you had one of the great point guards in history, one of the great low-post scorers in history, and worthy too. they knew what to do if they had to do it....

Posted by howard | November 14, 2006 7:56 PM

Missed threes also tend to lead to long rebounds which in turn lead to transition buckets. Thus, does the slightly higher expected value overcome the probability of giving up more easy buckets?

I meant to say this. However, the "long rebounds" thing could also lead to more O-boards. Without seeing numbers on this, I'd hesitate to advance that theory. However, I will say that generally more missed shots = more break opportunities...

And while I'm at it, it's hard to get fouled shooting a 3, and getting the other teams better players on the bench is always a Good Thing...

(Of course, I still love the three. Without it, I would never have made any team past the 8th grade...)

Posted by Pooh | November 14, 2006 8:00 PM

Chucker.

Posted by SomeCallMeTim | November 14, 2006 8:14 PM

You say that like it's a bad thing. Former teammates have said that I don't so much have shot selection as shot timing. It's not if I shoot, but when...

Posted by Pooh | November 14, 2006 8:18 PM

Nash plays the point guard position like a central midfielder in soccer. I don't know whether you can replicate that without having the personnel with Nash's background. That doesn't mean you can't emulate it, but it's not easy to coach the vision thing.

Posted by pseudonymous in nc | November 14, 2006 8:46 PM

SCMT: I'm a bit lost about what exactly are you looking for in offensive/defensive and pace metrics. The basketball-reference.com formula for efficiency is simply points per 100 possesions and the formula for possesions is (basically) FGA+TO+(0.4*FTA)-Off.Reb. Now, this formula measures the average number of times the ball changes possesion in a game and as such has some limitations. A high pace may mean a team who gets a lot of fastbreaks or it may mean a team who simply shots early on the clock. A medium pace may hide a team who gets a lot of fastbreaks points but milks the clock on its other possesions. If you want to analyze efficiency further, you may try a four factors analysis (Efg%, Turnovers, Off.Reb. and FTAs). Basketball-reference has it.

Posted by Carlos | November 14, 2006 10:33 PM

I'm a bit lost about what exactly are you looking for in offensive/defensive and pace metrics

Me too. I think I was wondering how the various parts of defense relate to offensive style. Something like: does an increase in offensive pace mean good defense comes in the form of steals rather than lower fg%. But I'm not sure. Thanks for the info, though.

Posted by SomeCallMeTim | November 14, 2006 11:26 PM

Where is the date for pace and efficiency?

Posted by Dan | November 15, 2006 11:11 AM

Err.. I mean where is the data for the pace and efficiency rankings? Is it behind the sub-only access? nba.com?

Posted by Dan | November 15, 2006 1:52 PM

It's at basketball reference.com, and it's not AFAIK, broken out. You can see it as part of each team-year entry.

Posted by SomeCallMeTim | November 15, 2006 2:04 PM