APBRmetrics

[supersub15]

I wanted to see whether there was a correlation between winning an NBA championship and having players voted All-NBA First Team. I started with the 3-point era, i.e. from 1979 until now and found the following: In the 29 championships, only 6 teams did not have a player named to the All-NBA First Team, i.e. 80% of the teams have won when they had one of the top 5 players in the NBA. These 6 teams are:

2003-04 Detroit Pistons
1994-95: Houston Rockets
1989-90: Detroit Pistons
1988-89: Detroit Pistons
1981-82: Los Angeles Lakers
1978-79: Seattle Supersonics

Of those 6 teams, only the 1988-89 Pistons and the 1978-79 Sonics didn't have a player in one of the 3 all-NBA teams (mind you, the third all-NBA team was introduced for the 1988-89 season and the Sonics had Gus Williams, Jack Sickma, and Dennis Johnson, 3 legitimate candidates to the third All-NBA team, had it existed at the time).

Consequently, the rate of winning a championship goes up to 93% if you have a player among the top 15 players in the league.

I then went all the way back to the BAA and the start of the NBA, 1946-1947 to be exact, and tried to make the same correlation. In 61 championships, the winning team had at least one player in the All-NBA First Team 45 times (74%). However, if we account for teams having at least one player in the 2 or 3 All-NBA Teams, only 3 teams didn't have a player in the top 10 or 15, the aforementioned Sonics and Pistons teams and the 1978-1979 Washington Bullets team. That's 95% of the time.

Just something that intrigued me. Thoughts?

[basketballvalue]

[supersub15]

[Mountain]

Checking over last 10 years I see 4 teams with 2 first team all-NBA players and 2 of the 4 won championship or 50% chance. Too small a sample of course but maybe supersub will want to do it for a longer span. Phoenix missed this past season and 03-04 Lakers did also.

Chance of winning with 1 all-NBA player isnt far from 1 in 5.

Chance with one player on any of the 3 all-NBA teams is only about 1 in 12-13, at least from looking at last few years.

This was a strong article by Dennis Gallagher on the topic from 18 months ago. A similar article or post based on who teams have today and what their key players have achieved in top ranks would be a good read.

http://www.82games.com/dennis.htm

Of the 7 one-Star champions all were top 6 on defensive efficiency. The average on offense was about 8.5 but 3 were near or actually below league average. If you are one-star you better be top level on defense. Teams see the advantages of 2 stars but a number of them I think are mistaken in thinking their second star is good enough (not that commmon by these standards) especially if both on offensive stars. One star and a top defense seems more feasible and cheaper than 2 stars. It may give you a legit shot but still only a modest chance with probably at least 6 other legit contenders a year?

I think 7 teams (Spurs, Suns, Rockets, Celtics, Cavs, Nuggets, Heat and Pistons) have a tandem that meets these qualifications. Chicago and Mavs join the contenders for having one star and a top 6 defense. Jazz is very close Boozer almost top 8 on PER and Kirilenko was a top defender but with his game sketchy I don't feel like giving them the slight break. I don't think Nets or anyone else makes this top player historical based cut but I'll add one spot for any dark horse and you have essentially a 10 horse race. Don't need a study to make that kind of statement but that's where I end up with a quick updating of the study.

To win a championship takes some combination of top players and top team efficiency on offense and defense or differential. It might be illuminating to construct a study that boiled in that down to:

Yes or no?
1. Top player (by Gallagher standards or otherwise)?
2. Qualifying second star?
3. A qualifying sidekick? (perhaps with a bit looser definition)
4. Top ten on offensive efficiency?
5. Top ten on defensive efficiency?
6. Top ten on point differential?

How many of last x champions had 4, 5, 6 of these?
How many others did each year? Were they any champions with less than 3 yeses?

Looks like 07 Spurs were first team with 6 since Bulls dynasty. All other champions in that timespan have had 5.

Based on performance last season I think (moving quickly) the teams score this way: Spurs 6, Suns 4 (possibly 5 if sidekick criteria is stretched), Rockets 4 (5 if Mutumbo still counts or Battier), Celtics n/a, Cavs 4, Nuggets 6, Heat 3 Chicago 4 and Mavs 4 (possibly 5 if sidekick criteria is stretched), Detroit 5.

Spurs a leader in championship construction, no surprise. Denver might be a surprise but they are top 10 on all 3 team measures and have Iverson, Anthony and Camby. The 6 criteria system is still first cut and you have to have the right stars. Denver with more time together to develop chemistry will get at least one more chance.

I'd add a 7th criteria on principle- a top 10 coach but I don't think it would change anything looking at last season, except maybe with Mike Brown of Cavs but too early to say on him and hard to say independent of team results.

Moving from last season numbers to next season, who improved? Celtics of course, probably become a 5. Chicago might become a 5 if Deng advances up top player list. Maybe Mavs or Cavs do that too- maybe. That about all I see in terms of major changes. Spurs best positioned. I don't think there is a clear #2 or 3. Pretty wide open. If a solid by the book 5 is needed next year then the list of top contenders might end up at 6-7 instead of near 10.

(If you counted a top flight 1st or 2nd team awarded 2 way player as 2 stars instead of just one some of these team scores would go up throughout the prior discussion.)

[KeeneKaufmanWheeler]

Isn't this a study that might lose some validity as the composition of the league has changed (through expansion and the evolution of the salary cap)?

20 years ago there were fewer teams and therefore a higher percentage of the league was all-NBA players.

[Flint]

A few thoughts. That Supersonics team and the Rockets, were the third and fourth worst teams by differential to win an NBA title. Only the Celtics and Bullets in the 1970's were worse.

So you have two teams that simply weren't very good.

The 81-82 Lakers team had Magic Johnson. His season that year was possibly the best season ever in the three point era. His not being first team all - nba was as close to cosmic injustice as you are likely to experience in the field of basketball statistics. Gus Williams!?!

http://www.basketball-reference.com/fc/pcm.cgi?req=1&cum=0&p1=gervige01&y1=1982&p2=willigu01&y2=1982&p3=johnsma02&y3=1982

That leaves the three Pistons teams. They had Ben Wallace and Dennis Rodman on them, players unlikely to ever receive consideration from the the First Team All NBA committee. I don't know that those players were top five players in those years, but if not they were very very close.

I think what this leaves you with is the observation that you very rarely win a championship without an actual top five player, and that the All NBA team doesn't always feature the top five players.

Lol, ok, Captain Obvious, signing off

[Mike G]

Flint,

Why don't you just admit that you are Dave Berri?

1982 was far from Magic Johnson's best year. It was just his best rebounding year.
_________________
40% of all statistics are wrong.

[asimpkins]

Great illustration of the differences in thought here. If we throw out the two seasons he was injured and played less than half the season, his 81-82 season ranks 10th out of 11 according to PER. This is because that year he was:

10th out of 11 in Usage Rate. 10th out of 11 in TS%. 10th out of 11 in Assist Rate. 7th out of 11 in Turnover Rate.

But yes, 1st out of 11 in Rebounding Rate.

[bchaikin]

The 81-82 Lakers team had Magic Johnson. His season that year was possibly the best season ever in the three point era. His not being first team all - nba was as close to cosmic injustice as you are likely to experience in the field of basketball statistics. Gus Williams!?!

fwiw simulation shows magic johnson's 81-82 season as the 2nd best in terms of wins generated by a PG - on a 40 min/g and 82 game basis - over the past 30 years (since 77-78 and when turnovers were first tracked), bettered by only his 80-81 season...

however gus williams in 81-82 was also very good, generating wins at a rate of just about 4-5 less wins than magic's 81-82 season on a 40 min/g and 82 game basis...

[Flint]

Well, thanks Mike G, thats something of a compliment from my view, but I am definitely not David Berri. Just a fan of his and the Knicks. Do you honestly think I would be asking how to calculate team TS% if I were an economics professor?

Also, even if that wasn't Magic's best season, for the purposes of this discussion, I think it's pretty clear he was a better player than both of the guards voted first team all nba ahead of him. Or did Gus Williams have a higher PER than Magic that year?

[Flint]

Bchaikin - Are you Dave Berri? C'mon, fess up. We know it's you.

Actually I made a mistake. That was not Magic's best season according to Berri. That would be 88-89 where he had a WP48 of .538 and produced 32.4 wins

http://www.wagesofwins.com/MagicCareer.html

But what you wrote is very interesting though. Thanks for vaguely agreeing with me. That would possibly be a first on this site. What do your numbers show for Gervin actually? And for that matter what do they show for Rodman and Wallace?

[bchaikin]

Are you Dave Berri? C'mon, fess up. We know it's you...

ya' got me...

actually i clicked on your link and he has a stats called PAWS. i've got cats for pets so anyone creating a stats called PAWS is ok by me...

Actually I made a mistake. That was not Magic's best season according to Berri. That would be 88-89 where he had a WP48 of .538 and produced 32.4 wins...

let's compare magic's 88-89 and 80-81 stats:

year------min---pts/g---ScFG%---reb---ast----st----to---bs----g
8081---1371---21.6----.570-----320---317--127--143--27---37
8889---2886---22.5----.610-----607---988--138--312--22---77

to compare apples to apples let's normalize to minutes played (multiply his 80-81 stats by 2886/1371):

year------min---pts/g---ScFG%---reb---ast----st----to---bs----g
8081---2886---21.6----.570-----674---667--267--301--57---77
8889---2886---22.5----.610-----607---988--138--312--22---77

and if you want to be a bit more technical, adjust for pace (80-81 lakers got 103.9 poss/g, the 88-89 lakers 101.6 poss/g, so multiply his 80-81 stats by 101.6/103.9):

year------min---pts/g---ScFG%---reb---ast----st----to---bs----g
8081---2886---21.1----.570-----659---652--261--294--56---77
8889---2886---22.5----.610-----607---988--138--312--22---77

what i see here is a player who got 261 - 138 = 123 more steals, 312 - 294 = 18 less turnovers, and 34 more blocks in 80-81 vs 88-89. estimating say 55% of all blocked shots are retrieved by the defense as defensive rebounds, that'd be 123 + 18 + 34x0.55 = 123 + 18 + 19 = 160 less zero point possessions for magic in 80-81 vs 88-89 (not to mention he got more rebounds in 80-81)...

so the question is for magic's 88-89 season to generate more wins than his 80-81 season, how does he make up for this difference of 160 less zero point team possessions in 77 games or 2.1 zero point possessions per game. after all a steal is the vast majority of the time the end of your opponent's possession, and thus they do not score, and a turnover for your team most often results on a team possession where your team did not score...

to make up for 160 zero point team possessions - not to mention 52 less rebounds - he would have to have accounted for, say, more defensive stops outside of steals, such as by forcing more misses or forcing more turnovers that are not steals. a forced miss is not a defensive stop unless your team gets the defensive rebound, so approximately 160/0.66 = 242 more forced misses would have to have been forced by magic in 88-89 vs 80-81 (with no more forced turnovers), or some combination of less forced misses and more forced turnovers (the 80-81 lakers got 64% def rebs and the 88-89 lakers got 69% def rebs so i used 66% as an average)...

you could point out that magic shot better overall in 88-89, but just because magic didn't score on a certain team possession (and isn't personally responsible for a team zero point possession) doesn't mean his team didn't score on that possession. plus you could point out that magic in 88-89 got 988 - 652 = 336 more assists than in 80-81. both surely make up for some of those 160 zero point possessions and 52 less rebounds. the question is how much?...

however a PG can have significantly less assists and even shoot worse than another PG but still generate as many if not more wins through his other stats. this was pointed out in another thread:

In Brandon's career year of '96, he was apparently as good as Stockton was in any year.

some will point out that magic's better shooting and more assists in 88-89 more than make up for these zero point possessions, but simulation shows that on a 40 min/g and 82 game basis, magic's 80-81 season generates slightly more wins than his 88-89 season, 2 to 3 more wins depending which team you run the simulation on. both seasons however are among the very best in generating wins by a PG in the last 30 years...

But what you wrote is very interesting though. Thanks for vaguely agreeing with me.

you are vaguely welcome...

What do your numbers show for Gervin actually?

that same 81-82 season george gervin scored 32 pts/g for the spurs. simulation shows that - on a 40 min/g and 82 game basis - he generated wins at the rate of about 12-13 more than some of the worst win generating SGs in the league that year (like chris ford of the celtics, phil smith of the clippers/sonics, and jim brogan of the clippers), and that year i have gervin rated as a poor defender outside of steals, blocks, and defensive rebounding. but that season simulation also shows sidney moncrief as generating 4-5 more wins than gervin did on a 40 min/g and 82 game basis, and about 16-17 more wins than the worst win generating SGs. that year i have moncrief rated as an excellent defender outside of steals, blocks, and defensive rebounding...

And for that matter what do they show for Rodman and Wallace?

ben wallace's best year statistically was probably his 01-02 season (his first DPOY award). simulation shows - on a 40 min/g and 82 game basis - wallace generating as many wins as shaquille o'neal did that year (when shaq scored 27 pts/g and wallace managed less than 8 pts/g). that season wallace got 13 reb/g, 1.7 st/g, and 3.5 bs/g, but while committing less than 1 to/g and shooting a ScFG% right at the league average...

possibly rodman's best year statistically was 91-92. that year he generated - on a 40 min/g and 82 game basis - as many wins as both karl malone and horace grant...

[Flint]

Thanks for the response. And I did appreciate your response.

I am not a WOW expert. And I don't know exactly how he calculates wins produced. But I understand his logic I think.

For simplicity sake, here are the numbers I am looking at. Per 48.

http://www.basketball-reference.com/fc/pcm.cgi?req=1&cum=0&p1=johnsma02&y1=1981&p3=johnsma02&y3=1989

In Berri's book, a block, assist, ft or personal foul is worth .5 a rebound, point, steal, or shot attempt.

Sometimes I mess up my numbers, I don't have a spreadsheet set up like everyone has here. Have to figure out how to do that.

In non-scoring stats, Young Magic has an advantage of

+ 1.1 rebounds
+ 2.1 steals
+.25 blocks
+.2 turnovers
- .3 pfs
- 2.65 assists

+ .7


On the scoring end, Young Magic scores 27.9 points but he takes 20.6 shots and 7.9 fta's. So dividing the ft's by 2, this works out to a deduction of 24.55. His net from his points is 27.9 - 24.55, or 3.35

On the Scoring End Old Magic scores 28.8 points on 23.6 shots, for a net of 5.2

So Old Magic has a scoring advantage of 1.85

Overall, the 88-89 version has a 1.15 advantage, per 48.

Now, that is just how it works out with win score. REally, thats some gunky analyis. As I said, I don't know how Wins Produced works exactly. From the end notes of his book it seems like a fairly involved process. But I think the basic logic I employed above probably approximates his reasoning for why the later Magic was better. I think Wins Produced is adjusted for pace, and there is a team defensive adjustment in there as well, so that might make a difference, but I think that's the gist of it.

A much shorter way of saying it would have been that a 4% difference in TS% is a big difference in shooting efficiency. And 5 more assists per 48 is a big difference. Those things more than make up for Young Magic's non-scoring stat superiority.

I had no idea Gervin or Williams were that good. Both were before my time. I thought Gervin was Iverson-esque, but clearly I was wrong. But, it still seems like Magic was the better player than them in 81-82. So, I feel my original point stands, basically, that sort of is one less team that won without a 1st team All-NBA player.

Re Wallace -

I don't know what the exact numbers for Wallace are, but I would think he was a top five player at least one of those years 2000-5. Your model sees to be a lot more in line with Berri's conclusions than others I have read about here. I am not familiar with what you do. Do you have a link to a thread explaining it?

And re Rodman

I don't know what his best year was. Berri says in the book he was the best power forward in the league in 93-94, when, incidentally, he averaged 8.4 offensive rebounds per 48. He seems to have been pretty good in 95-96 on the best team of all time also. I think Berri rates him at a .405 that year, although he didn't play 82 games that year, and that's just off the top of my head.

But who knows, Berri probably doesn't have it right. According to his PER of 13.6 he was a below average player and the seventh best player on that team, despite averaging 18 rebounds per 40.

Lol, shouldn't say stuff like that around here Thanks for the post, good night.

[gabefarkas]

[HoopStudies]