OWS distribution reveals championship style pg play

Moderator: Doctor MJ

trainwreckog
Banned User
Posts: 752
And1: 110
Joined: Jul 13, 2012

Re: OWS distribution reveals championship style pg play 

Post#21 » by trainwreckog » Sat Nov 3, 2012 5:20 pm

the data does support what i'm saying.

what i'm saying is that regular-sized pg's that have ows higher than 7.5 almost never make the nba finals (2 times in the last 30 years). this is a historical, statistical fact that cannot be argued. (keep in mind when considering the rest of this post, that almost all championship teams have forwards, centers or big guards that had ows higher than 7.5 as a standard, but never regular-sized pgs).

what does this mean? not exactly sure. but i interpret this to mean that having the smallest guy on the floor making most of the offensive decisions is not an optimal way to play the game. guys like nash and chris paul that dominate the ball are employing an inferior style of basketball that cannot succeed at the highest level.

if you look at the championship teams, teams with small pg's running the show are pretty rare. whereas teams that use more of a "position-less" approach with no defined pg (80's lakers, 90's bulls, 2000's lakers, 2011 mavs, 2012 heat) seem to be the ones winning all the rings.

think about it - is having a guy 6'1" make most of the offensive decisions going to be more optimal/harder to stop than having all 5 guys involved in the offensive decision-making? i'm pretty certain it isn't more optimal.

but that's just my interpretation. you can interpret the ows trend however you want.
trainwreckog
Banned User
Posts: 752
And1: 110
Joined: Jul 13, 2012

Re: OWS distribution reveals championship style pg play 

Post#22 » by trainwreckog » Sat Nov 3, 2012 5:26 pm

i guess one conclusion that i would make from the ows trend, is that the lakers should NOT hand over the reins to nash and let him run his pick-n-roll show - this is not an optimal style - the lakers should stay patient with their read-and-react offense and learn it.
User avatar
Hendrix
RealGM
Posts: 17,030
And1: 3,662
Joined: May 30, 2007
Location: London, Ontario

Re: OWS distribution reveals championship style pg play 

Post#23 » by Hendrix » Sat Nov 3, 2012 6:19 pm

You really should read what people are saying to you, and try and understand it instead of just repeating the same thing over and over.

The game of basketball is just about ORTG>DRTG. There's really no reason at all a team can't have a great differential between ORTG and DRTG when their PG is a very good player.

Like others have said, you are finding a cause becuase of some correlation over a small data sample. For example, I've played a ton of poker, and have a program that tracks every hand. It's quite possible to be a net loser with pocket kings over the small sample size of 30 hands. That doesn't mean you determine KK is a loser hand, and stop playing it. When you look at my database of over a million hands played KK is a massive winner in the long run.
oak2455 wrote:Do understand English???
trainwreckog
Banned User
Posts: 752
And1: 110
Joined: Jul 13, 2012

Re: OWS distribution reveals championship style pg play 

Post#24 » by trainwreckog » Sun Nov 4, 2012 9:45 pm

Hendrix wrote:You really should read what people are saying to you, and try and understand it instead of just repeating the same thing over and over.

The game of basketball is just about ORTG>DRTG. There's really no reason at all a team can't have a great differential between ORTG and DRTG when their PG is a very good player.

Like others have said, you are finding a cause becuase of some correlation over a small data sample. For example, I've played a ton of poker, and have a program that tracks every hand. It's quite possible to be a net loser with pocket kings over the small sample size of 30 hands. That doesn't mean you determine KK is a loser hand, and stop playing it. When you look at my database of over a million hands played KK is a massive winner in the long run.


the reason nash's style is not optimal is because good teams figure out the two-man iso game (pnr) in a 7 game series. that's why it has never won when it has been the staple of a team's offense. playoff games are usually won on just one or two possessions - the offensive rating of nash's offense in the clutch times of close playoff games is lower than it is at all other times.

the ows data over a sample size of 30 is pretty decent in my opinion. the game of basketball was invented in the early 1900's and at that time it was decided that the smallest guy is a pg and he runs the whole show. given the advancements in the game, this approach is obviously not the optimal way to play the game. and the fact that almost all the championship teams like the heat, mavs, lakers, bulls, rockets, 80's lakers, don't even use this antiquanted idea of a traditional pg only further supports what i'm saying. it's not an optimal way to play the game.

and given all the great, short ballhandlers there are nowadays and it is only increasing, they will one day be a dime a dozen and will be a minimal part of the game.
trainwreckog
Banned User
Posts: 752
And1: 110
Joined: Jul 13, 2012

Re: OWS distribution reveals championship style pg play 

Post#25 » by trainwreckog » Sun Nov 4, 2012 9:49 pm

so when kobe calls everyone idiots saying they don't know what a championship team is, and to be patient with them learning a more position-less, read-and-react offense, which is the more optimal way to play, i think he is being serious.
User avatar
Hendrix
RealGM
Posts: 17,030
And1: 3,662
Joined: May 30, 2007
Location: London, Ontario

Re: OWS distribution reveals championship style pg play 

Post#26 » by Hendrix » Mon Nov 5, 2012 7:03 am

30 years is an abysmally small sample size of data to come to this kind of conclusion in the way you are attempting to. Plus you made it even smaller by completely eliminating the years where Magic won the championship. So, your basically down to a sample size of 26.

Of that remaining 26 years, 3 of them actually did have PG's with a 7.5+ows. And, seeing as 7.5 ws's is a rather arbitrary line I think it's worth mentioning that multiple other championship teams had a PG in the 7-7.5 ows range. But, for arguments sake I won't count those.

So, now we are down to 23 years of a sample size. In those 23 years, 7 of them were won by teams that just happened to have the best SG, and SF in the league on the same team. Another 4 were won by teams that had the best SG, and best C in the league on the same team. When a sample size can be this heavily impacted by random occurrences like this, it's not enough. if you had a larger sample size there should be an even distribution of times when the best PG in the league happened to be paired up with the best C in the league. Or best PG league happened to be paired up with the best SF in the league. But over the small 23 year sample size this just hadn't happened.

I mean, hypothetically lets say Paton, and MJ just happened to be on the same team instead for a bunch of years in the early 90's, and then Payton+Shaq happened to be on the same team for a bunch of years, and then Chris Paul+Dwight Howard happened to be drafted onto the same team. Well, a great PG would have probably would have won a ton of rings over this sample size. In this hypothetical situation there would hardly be any championships won with great SF's on the team. So would you then be coming in here saying that you can't win with a great SF, and need a PG?

In Poker when evaluating EV vs. ROR, we consider a 40,000 hand sample size small. Usually once you were around the 250,000 hand range you would have something to work with.
oak2455 wrote:Do understand English???
User avatar
Rapcity_11
RealGM
Posts: 24,496
And1: 9,530
Joined: Jul 26, 2006
     

Re: OWS distribution reveals championship style pg play 

Post#27 » by Rapcity_11 » Tue Nov 6, 2012 12:05 am

trainwreckog wrote:the reason nash's style is not optimal is because good teams figure out the two-man iso game (pnr) in a 7 game series. that's why it has never won when it has been the staple of a team's offense. playoff games are usually won on just one or two possessions - the offensive rating of nash's offense in the clutch times of close playoff games is lower than it is at all other times.


Do you have data to support any of this?

the ows data over a sample size of 30 is pretty decent in my opinion. the game of basketball was invented in the early 1900's and at that time it was decided that the smallest guy is a pg and he runs the whole show. given the advancements in the game, this approach is obviously not the optimal way to play the game. and the fact that almost all the championship teams like the heat, mavs, lakers, bulls, rockets, 80's lakers, don't even use this antiquanted idea of a traditional pg only further supports what i'm saying. it's not an optimal way to play the game.

and given all the great, short ballhandlers there are nowadays and it is only increasing, they will one day be a dime a dozen and will be a minimal part of the game.


Like I said earlier, you're indicting a certain style of offense by looking at an overall result which factors in defense and rebounding. That doesn't make sense. You need to try and isolate offensive performance as much as possible in order to draw meaningful conclusions about offensive strategy.

The 04 Pistons won the title, but would you recommend their offensive style/system over the 05-07 Suns, which had much greater offensive success in the playoffs? Or the 03 Spurs?
Chicago76
Rookie
Posts: 1,134
And1: 228
Joined: Jan 08, 2006

Re: OWS distribution reveals championship style pg play 

Post#28 » by Chicago76 » Tue Nov 6, 2012 7:08 am

trainwreckog wrote:the data does support what i'm saying.

what i'm saying is that regular-sized pg's that have ows higher than 7.5 almost never make the nba finals (2 times in the last 30 years). this is a historical, statistical fact that cannot be argued.


I'm not going to argue with the fact that teams with a PG with a OWS of +7.5 have only comprised 2 of the last 60 finalists. That is a fact.

I am going to dispute your assertion that this fact means that the style of play generating this kind of output is the reason these teams aren't making the finals. What I will make note of is the following:

1) small PGs getting 7.5+ OWS is very rare. It occurs roughly 2-3 times per year on average. The odds of neither team in the Finals not having this type of PG randomly in a given year is roughly 86%. You can set up a binomial distribution to figure out over a 30 year period the odds of teams w/ a PG of +7.5 or greater making the finals X times.

Odds of it never happening: 1%
Happening once or zero times: 6%
Happening two or fewer times: 19%
Three or fewer: 38%
Four or fewer: 59%
Five or fewer: 73%

The odds of it happening only twice are somewhat below what you would randomly expect, but not by a statistically significant margin--especially when you figure in the strike shortened years (which makes it harder to get to +7.5 OWS with fewer games). Limiting n to 28, the odds of it happening two or fewer times = 23%. This isn't much different than flipping a coin 28 times and having it come up heads 12 or fewer times. 12 out of 28 isn't a whole lot different than the long run estimate of 14 of 28. Would you say this conclusively proves the coin is biased against heads? It's the relative scarcity of +7.5 PGs that is driving the infrequency.

2) Another problem w/ with the conclusion is that you are assuming that a high OWS player is making the offense inefficient and that they would be better if they somehow more evenly distributed the ball. The problem with this is that a lot of +7.5 OWS PGs were +7.5 players because the offense around them sucked. If you have 5 really awesome offensive players on a team, it will be difficult to hit +7.5, but I'm sure your chance of winning a title or getting to the Finals would increase. Conversely, if you don't, one player may need to pick up a lot of slack. The team is worse, and the results are worse, but this isn't done through some conscious design but rather out of necessity. Take a guy like Stockton. Sure he had Malone and he used him a ton. He had a Hornacek type as a third wheel, but who else was he supposed to concede more offense to? Malone was as used as he could be and Hornacek wasn't really a shot creator. Greg Ostertag wasn't going to step in to fill the void.

3) As others have mentioned, offense is only one half of the game. In many cases, the offensive production of +7.5 OWS PG-led teams was championship caliber, but their defense was poor. Drawing any conclusion in terms of Ws and Ls (which is dependent upon both offense and defense) would be like me saying great hitting player X in MLB can't get it done while overlooking his team's awful pitching. Or it would be like saying QB Y can't win the big one while ignoring his team's awful defense.
User avatar
Dr Positivity
RealGM
Posts: 59,694
And1: 15,486
Joined: Apr 29, 2009
       

Re: OWS distribution reveals championship style pg play 

Post#29 » by Dr Positivity » Tue Nov 6, 2012 7:57 am

PGs really haven't done that bad from 1980 to now. Magic obviously was a megastar (5 titles, 9 Finals). Isiah was the guy the Pistons built around and consensus franchise player (2 titles, 3 Finals). Billups was the leader of the 04 Pistons (1 title, 2 Finals). Payton and Stockton combine for 3 Finals. Kidd makes 2 Finals with the Nets. Nash doesn't make one but the 07 Spurs series was for all intents and purposes the Final. Then for star supporting PGs Kevin Johnson makes a Final with the Suns, Tony Parker wins 3 titles with the Spurs (2 as a star), Rondo wins a title and makes 2 Finals (1 as a star), Terry Porter was maybe the 2nd most important player on 2 Blazers Finals teams. Mo Cheeks makes 3 Finals (1 title). In fact the gauntlet of star PGs the last 30 years has been pretty loaded with Finals appearances. Also 78 and 79 had a star PG in the Finals in Gus Williams too extending just a few years back

Return to Statistical Analysis