Fantasy Blogs  |  The Court Jester  |  

 

Don Nelson, why do you torture us? You’re probably headed to the Hall of Fame. You make millions of dollars. You’re obviously not going hungry. Yet you sit up late into the night cackling and rubbing your bony hands together as you plot the ruin of fantasy owners whose only crime is to believe in one of your players.

As any owner you haven’t already sent to the lunatic asylum knows, you have a bounty of guys who would be major fantasy studs if given consistent minutes and touches. The names Morrow, Randolph, Watson, and even Turiaf leap to mind. Yet the only sure bet on any given night is that Monta Ellis will get too much run and three of those other four dudes will not get enough.

Playing like a monster is not enough to guarantee playing time on Nelson’s squad. Oh no. Consider Anthony Morrow, who was number one – that’s right, number one – in the league in preseason. From February 8^th to 17^th Morrow scored 33, 26, 23 and 25 points. As a reward his PT was cut over the next three games to 29, 26 and 27 minutes. On December 5^th Randolph got 35 minutes and notched 28 points, 13 boards, five assists and a block. His reward was to be cut the next game to 25 minutes and then to 20, 19, and 14. On the 23^rd Randolph once again got a decent, 36 minute run and amassed 12 point, 8 boards and 8 blocks. Good old Nellie responded by cutting his subsequent minutes to 26, 24, 19, 20, 22, 24 and 12.

Nellie, you have a bench full of budding fantasy superstars, and your team is horrendous and riddled with injuries. Why is it that you seem to believe none of these guys deserves consistent time on the court?   

In fairness to Nelson, there may be a good reason. I suspect that the terrific statistical performance of Golden State’s fringe players has to do with team pace.

When it comes to speed, not all teams are created equal. This is shown by the number of possessions teams generate over the course of the season. I derive this by the simple formula FGA+FTA/2+TO, which pretty neatly captures the number of a team’s offensive series outside of the odd three point play or technical free throw.

There’s a noticeable difference between teams here. The league’s slowest-paced team, the glacial Portland Trailblazers, slog to 102 possessions a game. The fastest team is none other than the Warriors, who sizzle to 114.  

This difference might not seem like a lot, but try multiplying Portland’s players’ stats by 1.12 and see what happens—the cumulative effect is large. (For example, for the season it would raise Lamarcus Aldridge from 57^th to 34^th.)

I have always assumed this is what happens. After all, more possessions must equal more shots, more rebounds, more blocks, more steals, and more threes, mustn’t they? For this reason I always have given a player a considerable bump on my preseason spreadsheet when he changes addresses and goes to a squad that plays a more run-and-gun type of game. And this is still a good general rule.

However, my recent fiddling with the numbers tells a very interesting story. Here are the teams ranked by number of possessions, along with the overall team fantasy score (as you would get if you added up the contributions of all their players.) See if you can discern the relationship.

 

Rnk	team	pos/g	score
1	GSWarriors	114	1.65
2	MinnesotaT-wolves	112	-6.53
3	DenverNuggets	111	4.61
4	MemphisGrizzlies	111	-1.46
5	SacramentoKings	111	-3.19
6	HoustonRockets	110	-0.59
7	IndianaPacers	110	-0.55
8	PhoenixSuns	110	6.00
9	ChicagoBulls	110	-0.32
10	LALakers	110	4.33
11	MilwaukeeBucks	109	0.41
12	NYKnicks	108	0.73
13	OklahomaCityThunder	108	2.27
14	WashingtonWizards	108	-3.73
15	UtahJazz	107	3.73
16	TorontoRaptors	107	0.60
17	Philadelphia76ers	107	0.82
18	AtlantaHawks	107	4.19
19	LAClippers	107	-3.06
20	NOrleansHornets	106	0.83
21	DallasMavericks	106	6.88
22	DetroitPistons	106	-9.26
23	SanAntonioSpurs	106	0.98
24	NJNets	105	-8.82
25	OrlandoMagic	105	1.25
26	CharlotteBobcats	105	-4.27
27	ClevelandCavaliers	104	3.47
28	MiamiHeat	104	-0.77
29	BostonCeltics	103	0.85
30	PortlandTrailBlazers	102	-1.06

 

Did you see the pattern? I hope not—because there is no pattern. I ran the correlation between these numbers and came up with a coefficient of .075. This means there is absolutely no association between team tempo and player stats, at least based on this year’s numbers. Lots of possessions or few, it seems to make no difference in team fantasy production.

How, you might ask, could this possibly be? Don’t more possessions simply have to mean better overall numbers?

The missing link in the explanation here is team records. Here’s the same table, this time sorted by team winning percentages. This time there is a pattern:

 

Rnk	team	PCT	pos/g	score
1	ClevelandCavaliers	.766	104	3.47
2	LALakers	.730	110	4.33
3	OrlandoMagic	.683	105	1.25
4	DallasMavericks	.672	106	6.88
5	DenverNuggets	.661	111	4.61
6	BostonCeltics	.650	103	0.85
7	AtlantaHawks	.645	107	4.19
8	UtahJazz	.645	107	3.73
9	PhoenixSuns	.615	110	6.00
10	OklahomaCityThunder	.607	108	2.27
11	SanAntonioSpurs	.600	106	0.98
12	PortlandTrailBlazers	.578	102	-1.06
13	TorontoRaptors	.533	107	0.60
14	MilwaukeeBucks	.532	109	0.41
15	HoustonRockets	.508	110	-0.59
16	MiamiHeat	.508	104	-0.77
17	MemphisGrizzlies	.508	111	-1.46
18	ChicagoBulls	.500	110	-0.32
19	NOrleansHornets	.492	106	0.83
20	CharlotteBobcats	.492	105	-4.27
21	LAClippers	.397	107	-3.06
22	Philadelphia76ers	.361	107	0.82
23	WashingtonWizards	.356	108	-3.73
24	NYKnicks	.339	108	0.73
25	SacramentoKings	.339	111	-3.19
26	DetroitPistons	.339	106	-9.26
27	IndianaPacers	.317	110	-0.55
28	GSWarriors	.274	114	1.65
29	MinnesotaT-wolves	.222	112	-6.53
30	NJNets	.113	105	-8.82

 

You might have spotted it with the naked eye, but if not, the correlation coefficient between tempo and wins is -.302. What this means is that, based on this year’s data, the loser teams tend to run and gun while the winners are more likely to slow it down and grind it out. The reason for the lack of a correlation between fantasy performance and team speed is that the weaker teams generate more possessions but their inept players cannot take advantage of the situation.

Any statistician would tell you that correlation does not mean causation, and there are a couple of possible explanations for this story.

First, run-and-gun basketball might simply be bad basketball. A deliberate style means milking the most out of every possession, while casting off at the first opportunity means squandering precious possessions on low-percentage shots. A high-octane game might kill team performance.

Another possibility is that teams that are destined to be bad decide to go uptempo. Perhaps coaches and GMs of sorry squads figure that if they’re doomed to playing losing ball they might as well play entertaining ball.

In any event, this helps explain the Nelson phenomenon. Yes, under Nelson’s system minutes mean numbers. But it isn’t because the players are necessarily any good. This helps explain why Nelson is quick to bench dudes who are putting up big numbers; they’re basically statistical illusions.

So in the future beware of adding big-number flash in the pan players from high-speed, losing teams. If given playing time the Randolphs of the world may put up the stats, but their fundamental weakness as actual players means there’s no guarantee playing time will continue to flow. Another statistical star may well be lurking even farther down the bench. Place a greater value on guys who are putting up numbers on more deliberate, winning teams and let Nellie work his Jedi mind tricks on your opponents, not you.