Author Message
Johnny Slick
Joined: 15 Oct 2005
Posts: 45
Location: Seattle
PostPosted: Mon Oct 17, 2005 3:54 am Post subject: Fun with Standard Deviations Reply with quote
In order to do that semi-study of what "makes up" championship basketball in the expansion era correctly, I've figured out the standard deviations for every team in each of the six categories that I delineated. First off, doing this allows me to see whether defensive rebounds belong alongside defense or in its own rebounding category. Please bear in mind that no matter what the numbers say, I will continue to lump them in with offensive boards. So to answer the age-old question (and by "age" I mean "one day"), here are the correlations:
Offrebs/Defrebs: .034
Defshoot/Defrebs: -.398 (i.e. the lower the shooting percentage, the greater percentage of defensive rebounds a team gets)
Um. Okay, I stand corrected. I'm putting down my mat and bowing in the general direction of Dean right now. Actually, I'm amazed to see that there is basically *no* correlation between offensive and defensive boards. I steeled myself to this possibility when I tabulated these bad boys when I noticed that the worst defensive rebounding teams in the league were very often decent at crashing the boards on O but were nearly always just plain awful at preventing points from being scored. Still, you'd think, right? I mean, it's freaking rebounding both ways.
So, why standard deviations? Because the whole point of this study (eventually) is to see in which categories Finals teams are better than their opponents, and how much better. The fatal flaw in just looking at rankings is that it assumes that 4th place is equidistant from 5th and so on. As anybody who's watched the NFL knows, there's often as much difference between #1 and #5 as there is between #5 and $15. At the same time, raw stats are just going to make teams that played in offensive or defensive minded seasons look better than the rest when they really weren't.
Anyway, eventually I'm going to go back to that "what makes championship teams" thing with this. For now, here are the most dominant and most dominantly awful teams in each of these categories.
Offensive Shooting (Points / (FGA+.4*FTA) )
1. 2003-4 Sacramento Kings (55-27, 2.72 standard deviations above the mean)
2. 1996-7 Utah Jazz (64-18, 2.51 SDs)
3. 2004-5 Phoenix Suns (62-20, 2.45 SDs)
Best Shooting Bad (sub-.500) Team
2003-4 Seattle Sonics (37-45, 1.67 SDs). The 1995-6 Sonics are right ahead of them. Of course, they won a few more games...
Worst Offensive Shooting
1. 2002-3 Denver Nuggets (17-65, -3.00 SDs)
2. 2000-1 Golden State Warriors (17-65, -2.83 SDs)
3. 1992-3 Dallas Mavericks (11-71, -2.62 SDs and forever remembered for "IUZZOLINO!" in the original NBA Jam)
Worst Shooting Good (above .500) Team
1999-0 Philadelphia 76ers (49-33, -1.33 SDs)
****
Offensive TOs (TO / (FGA + .4*FTA)
1. 2003-4 Dallas Mavericks (52-30, 2.57 SDs)
2. 1996-7 Chicago Bulls (69-13, 2.50 SDs)
3. 1990-1 Minnesota Timberwolves (29-53, 2.45 SDs)
Best Handles on a Bad Team
The aforementioned T-Wolves.
Worst Handles
1. 1999-0 Chicago Bulls (17-65, -3.30 SDs)
2. 1996-7 New York Knicks (57-25, -2.57 SDs)
3. 1995-6 Toronto Raptors (21-61, -2.35 SDs)
Worst Handles on a Good Team
Them Knicks. Can you imagine if that team could hold on to the ball? Maybe the Bulls wouldn't have three-peated.
Best Offensive Rebound Percentage (OR/oppDR)
1. 1991-2 New Jersey Nets (40-42, 3.71 SDs) - This was a team that basically did absolutely nothing but get offensive boards. They were like the Damon Jones of offensive rebounds. And yeah, 40-42.
2. 1994-5 Dallas Mavericks (36-46, 2.93 SDs)
3. 2002-3 Golden State Warriors (38-44, 2.78 SDs).
Worst Good Offensive Rebounders
Er... let me get the best good O-rebounding team instead. The 1997-8 Nets were 43-39, but is that really all that great? The first elite team on the list is the 96-7 Chicago Bulls at 8th.
Worst Offensive Rebounders
1. 2004-5 Toronto Raptors (33-49, -2.3 SDs)
2. 2003-4 Toronto Raptors (33-49, -2.0 SDs) - That's called a bad GM, folks.
3. 1989-0 Golden State Warriors (37-45, -1.99 SDs)
Best Bad O-Rebounding Team
The 1999-2000 Knicks (50-32) were 5th from the bottom.
*********
Defensive Shooting
Best Defenders
1. 2002-3 Sacramento Kings (59-23, -2.50 SDs). Say whaaaaaat?
2. 1988-9 Utah Jazz (51-31, -2.43). Now that one I can see. The Mailman and Mark Eaton combining to stop shots... man, I can't get the Kings at #1 out of my head.
3. 2003-4 San Antonio Spurs (57-25, -2.40 SDs)
Worst Good Defensive Squad
The 1993-4 Denver Nuggets were 34th with a 42-40 record (and a foul place in the hearts of all Sonics fans)... but the first actual under-.500 team doesn't come in there until the 2003-4 Toronto Raptors. You may remember them as the 2nd worst offensive rebounding team of this era.
Worst Defenders
1. 1992-3 Dallas Mavericks (11-71, -2.80 SDs). Anyone else remember how just plain awful these guys were? If memory serves, these guys started the year with Derek Harper, Jamal Mashburn, and Jimmy Jackson... and ran the triangle. Apparently the triangle in the wrong hands makes you suck at defense.
2. 1996-7 Boston Celtics (15-67, -2.63 SDs)
3. 1990-1 Denver Nuggets (20-62, -2.59 SDs). No suprises here. If you'll recall, this was the Paul Westhead run-and-gun Loyola Marymount offense Nuggets. The only thing amazing about this ranking is that there were actually two teams worse.
Best Bad Defensive Team
2001-2 Utah Jazz (44-38, -1.65 SDs). Guess Stockton and Malone were too old to play D.
********
Defensive Turnovers
1. 1997-8 Boston Celtics (36-46, 3.39 SDs). Rick Pitino's boys sure took to him his first year. Too bad about the whole gambling-on-steals-leading-to-easy-baskets thing (-2.24 SDs in defensive shooting).
2. 1993-4 Seattle Supersonics (63-19, 3.09 SDs). Okay, sure, all those steals meant jack squat when it came postseason time. But they sure were a fun team to watch.
3. 1996-7 Seattle Supersonics (57-25, 2.74 SDs). Actually 4 of the top 10 teams were George Karl coached Sonics squads.
Worst Good Stealing Team
Other than the Celtics? The 2002-3 Denver Nuggets were 2.28 SDs above the pack... and won 17 games all year long.
Worst Stealers
1. 1996-7 Denver Nuggets (21-61, -2.69 SDs)
2. 1999-0 Atlanta Hawks (28-54, -2.55 SDs)
3. 1989-0 Boston Celtics (52-30, -2.51 SDs). Coach, don't make Larry Bird swipe at the ball! He'll hurt his back!
Best Bad Stealing Team (really, this is about forcing all kinds of turnovers, not just steals, but "steals" is easier to type than "forcing TOs")
The Larry Bird Celtics. The only other better-than-.500 team in the top 12 are the 1993-4 San Antonio Spurs, who had Rodman and stuff.
********
Best Defensive Rebounders
1. 1997-8 New York Knicks (43-39, 2.67 SDs)
2. 1996-7 New York Knicks (57-25, 2.49 SDs)
3. 1992-3 New York Knicks (60-22, 2.42 SDs). See a trend? And people say Ewing was overrated...
Worst Good Rebounders
The 1995-96 Denver Nuggets were 18th and finished 35-47. The only other sub-.500 squad in the top 50 are the 1988-9 New Jersey Nets (26-56 and 21st).
Worst Defensive Rebounders
1. 1989-0 Golden State Warriors (37-45, -2.75 SDs)
2. 1992-3 Milwaukee Bucks (28-54, -2.58 SDs)
3. 1999-0 Dallas Mavericks (40-42, -2.45 SDs). That signing of Rodman doesn't seem so stupid when you look at it this way.
Best Bad Rebounder
The 1994-5 Golden State Warriors were 15th worst and finished 44-38, but last year's Suns are right behind them.
Back to top
View user's profile Send private message Yahoo Messenger
jambalaya
Joined: 30 Jan 2005
Posts: 282
PostPosted: Mon Oct 17, 2005 12:16 pm Post subject: Reply with quote
So for the top 3 in 6 categories (18 spots), only one championship team (1996-97 Bulls) makes the list as primo for modern times in specific areas? That is somewhat surprising. But I guess it is 17 championship teams competing with close to 500 non-championship teams over 17 years so 1 of 18 is 5.5% of the top slots with only about 3 % of the chances. Still, another argument for a good deal of balance and acknowledgement of the limits of where dominance in any one area will get you?
Back to top
View user's profile Send private message
Johnny Slick
Joined: 15 Oct 2005
Posts: 45
Location: Seattle
PostPosted: Tue Oct 18, 2005 12:52 am Post subject: Reply with quote
To a point, jambalaya, although one thing I've noticed is that the vast majority of really strong teams are good at defensive rebounding, no matter what their other strengths and weaknesses are. Don't want to make too much out of that, but it is the one big thing I've pulled out of this so far.
Back to top
View user's profile Send private message Yahoo Messenger
Johnny Slick
Joined: 15 Oct 2005
Posts: 45
Location: Seattle
PostPosted: Tue Oct 18, 2005 3:49 pm Post subject: Reply with quote
I entered in all the playoff teams and winners and will post the average SDs of each level. First, all playoff teams:
Average record: 50.2-31.8
Off scoring: 0.48 SDs
Off TOs: 0.28 SDs
Off Rebs: 0.02 SDs
Def scoring: 0.47 SDs
Def TOs: 0.12 SDs
Def Rebs: 0.33 SDs
1st round losers:
Record: 45.8-36.2
Off scoring: 0.20 SDs
Off TOs: .08 SDs
Off Rebs: .06 SDs
Def scoring: .13 SDs
Def TOs: .20 SDs
Def Rebs: .17 SDs
1st round winners:
Record: 54.5-27.5
Off scoring: .67 SDs
Off TOs: .35 SDs
Off Rebs: -.01 SDs
Def scoring: .80 SDs
Def TOs: .03 SDs
Def Rebs: 0.49 SDs
2nd round winners:
Record: 57.6-24.4
Off scoring: .82
Off TOs: .45
Off Rebs: .12
Def scoring: .98
Def TOs: .01 (actually, .005)
Def Rebs: .66
Teams that made the finals
Record: 58.8-22.2
Off scoring: .69
Off TOs: .36
Off Rebs: .47
Def scoring: 1.14
Def TOs: .12
Def Rebs: .70
Champions
Record: 60.6-21.4
Off scoring: .59
Off TOs: .76
Off Rebs: .69
Def scoring: 1.39
Def TOs: .02
Def Rebs: .58
Interesting how preventing turnovers and getting offensive rebounds don't appear to mean much of anything at first, but then begin to become more important as the playoffs progress. By the time you get to Finals winners, both are actually more important than offensive scoring efficiency, which appears to be the single largest factor when it comes to getting into the playoffs in the first place (well, it's basically tied with defensive scoring).
Here's a graph of this (hope it works):
Back to top
View user's profile Send private message Yahoo Messenger
jambalaya
Joined: 30 Jan 2005
Posts: 282
PostPosted: Tue Oct 18, 2005 4:54 pm Post subject: Reply with quote
Great work.
But you say "Interesting how preventing turnovers and getting offensive rebounds don't appear to mean much of anything at first, but then begin to become more important as the playoffs progress. By the time you get to Finals winners, both are actually more important than offensive scoring efficiency"
That may be your opinion, and you can of course advance it. The numbers certainly say more strongly correlated but I am not sure "more important" is the necessary inference or yet sufficently demonstrated case to everyone's or my personal satisfaction.
Rebounding of both kinds does indeed rise in strength of correlation for the winner through the rounds. Supportive of your position as their spokesperson (and I also speculated).
Defensive scoring rise in strength of correlation measured in standard deviations to about 240% of what it was for all playoff teams, while offensive scoring only rose about 20% (also in line with the defense rank results for championship winners study and with my recent speculation on the subject in the another thread looking beyond Kevin's regression for all teams regular season results to the changing dynamics of playoff basketball thru the rounds).
Thanks for the additional research.
One minor point about the graph, it might have also been nice to see a second graph with a single data point for all playoff teams in place of 1st round losers and first winners but I can see a logical for displaying it that way too, as long as people notice the choice and dont think it is my graphing option rather than your choice.
Back to top
View user's profile Send private message
Johnny Slick
Joined: 15 Oct 2005
Posts: 45
Location: Seattle
PostPosted: Tue Oct 18, 2005 5:17 pm Post subject: Reply with quote
jambalaya wrote:
Great work.
But you say "Interesting how preventing turnovers and getting offensive rebounds don't appear to mean much of anything at first, but then begin to become more important as the playoffs progress. By the time you get to Finals winners, both are actually more important than offensive scoring efficiency"
That may be your opinion, and you can of course advance it. The numbers certainly say more strongly correlated but I am not sure "more important" is the necessary inference or yet sufficently demonstrated case to everyone's or my personal satisfaction.
Good point. "More strongly correlated" is a better way of saying it. Still, defensive scoring in and of itself doesn't really correlate all that well with defensive rebounding, so I'm not terribly sure that there isn't a causal link here. The next step, I imagine, is regression analysis of the six factors.
Back to top
View user's profile Send private message Yahoo Messenger
jambalaya
Joined: 30 Jan 2005
Posts: 282
PostPosted: Tue Oct 18, 2005 5:31 pm Post subject: reply Reply with quote
"The next step, I imagine, is regression analysis of the six factors."
Yep. Again I hope it will someday be done with playoff data.
A stepwise graph of changes for these categories based on regular season data is nice / useful (I assume that your dataset remains the regular season at this point), but I'd really like to see if the changes in the playoffs actually proceeds in similar fashion for the changes in strength of correlation. I am assuming it does but it may be dampened, perhaps significantly, both on offense and defense by the presence of better teams on average than is reflected in the regular season data.
But if that is the case, it is part of the explanation and you can compare the results of the regular season to the playoff data and perhaps make some useful comments about how much tougher that play is than regular season.
You might find that some of the correlations more more or less than others in the playoff / regular season comparison so until we have that, or better than correlation, the playoff regression, it is too early to make any definitive statements about the importance of the 6 categories, or the other parts hidden in the realtively large categories labelled offensive and defensive efficiency. Or at least wise to do so and feel confident the data will confirm. Until we have the playoff regression it is still a little sketchy and debatable.
It is still early in the discussion but I remain curious if anyone has run a playoff data, team/opponent regression for these categories, perhaps not for the full time period or for six parts. If such a regression has not been run, to this group's knowledge, it seem even more valuable to do so. It would seem of even greater importance than the serch for a single player evaluation formula, since "basketball is a team game" and "it is all about the playoffs baby".
Last edited by jambalaya on Thu Oct 20, 2005 9:08 am; edited 5 times in total
Back to top
View user's profile Send private message
gabefarkas
Joined: 31 Dec 2004
Posts: 1313
Location: Durham, NC
PostPosted: Tue Oct 18, 2005 5:31 pm Post subject: Re: Fun with Standard Deviations Reply with quote
Johnny Slick wrote:
Um. Okay, I stand corrected. I'm putting down my mat and bowing in the general direction of Dean right now. Actually, I'm amazed to see that there is basically *no* correlation between offensive and defensive boards. I steeled myself to this possibility when I tabulated these bad boys when I noticed that the worst defensive rebounding teams in the league were very often decent at crashing the boards on O but were nearly always just plain awful at preventing points from being scored. Still, you'd think, right? I mean, it's freaking rebounding both ways.
Perhaps you could look at what player designations are most responsible for offensive and defensive boards, respectively?
Back to top
View user's profile Send private message Send e-mail AIM Address
jambalaya
Joined: 30 Jan 2005
Posts: 282
PostPosted: Tue Oct 18, 2005 5:54 pm Post subject: Reply with quote
The ratio of offensive and defensive rebounds are probably fairly consistent by position with some outliers, for especially tall, quick or strong leapers for their position guys.
In general the rank of responsibility would seem to be frontline then out thru the perimeter, with caveats on a play by play basis that the shooter has a heightened opportunity and responsibility on his missed shots that are drives and on defense the defender might deserve a little allowance for being preoccupied and less ready to rebound, except maybe on shots in close.
The presence of an especially strong or weak rebounding teammate (offensive or defensive) can explain a lot of the variation at other positions as has been discussed elsewhere, so you could argue for some form of adjusted rebounding rating accounting for at least teammates (and also position if you wanted one scale).
Back to top
View user's profile Send private message
jambalaya
Joined: 30 Jan 2005
Posts: 282
PostPosted: Fri Oct 21, 2005 5:13 pm Post subject: Avg. stats top 4 teams in playoffs last 5 years Reply with quote
4 teams in playoffs last 5 years vs all playoff opponents faced
...............FG% 3p% OFF-R DEF-R APG PPG
Top 4 44.1% 35.4% 11.8 31.6 20.1 95.0
..Opp 42.5% 33.3% 12.0 30.1 19.1 91.5
In rough terms it seems that the top 4 playoff teams have about a 4-5% advantage in each of these categories compared to the average of all their playoff opponents, except offensive rebounding. (They also had about .6 more blocks and .6 less fouls.)
(Since my playoffs database at the moment only includes 20 team observations, I assume the correlation and regression results would not be that meaningful, right? If all 80 playoff teams in this period were included it might be more meaningful but I have no immediate plans to compile that and I would need to refresh on sound interpretation of the results.)
Champion compared to other 3 conference finalists
FG% 3p% OFF DEF TOT APG PPG
offense 99% 100% 107%100%102% 96% 100%
opp faced 98% 86% 97% 97% 97% 88% 95%
On average, the last five championship winners have distinguished themselves from the other 3 conference finalists on offense only on offensive boards, while they are on average superior across the board on defense of playoff opponents faced with especially large differentials on pressing and reducing the three point shot % and assists allowed.
Back to top
View user's profile Send private message