APBRmetrics

[HoopStudies]

I'm sometimes pretty oblivious to popular opinion, but not this much. NBA.com had this article on Allen Iverson

http://www.nba.com/features/inside_ai_050330.html

that suggested that he was having one of the best statistical seasons ever.

A couple questions:

1. What do you guys think about "best statistical season"? They have a lot of bar-talk numbers up there that are pretty good.
2. In the old APBR_Analysis message 4611, it was suggested that we may "safely predict another season of decline" for AI. I know my measurements don't show a decline, but my stuff is different than most. Do other systems show a decline? If not, why not? The reasons postulated last year for decline were age and ongoing injuries. Did rule changes trump this? Something else? Position change?
3. I haven't looked at all at AI's performance since the Webber trade. What do those say?
_________________
Dean Oliver
Author, Basketball on Paper
http://www.basketballonpaper.com

[jambalaya]

it is probably his best overall year for individual stats. generally in line with what he did in 2001-02 and 2002-03 with the only major difference being improved assist numbers.

i assume the handchecking rule has helped the shooting and scoring stats.
the more complete role change to point guard has helped the assist numbers though he started to show improvement there last year.

the only thing i saw since the webber trade is that the passing flow is now different and in the last ten games iverson is averaging 8.6 assists, up a bit, but turnovers are a league high 6.4, 60% more than any other player. i would guess they will eventually get use to each other.

i dont think philly is as talented or perhaps as well coached as they have been in the past so it isnt apples to apples to look at this year's modest win-loss record compared to past glory days. it might be his best overall year for individual stats and i think you are right to phrase it way. it would take more thought and analysis before you could go to any stonger statement about this year as it relates to past years.

[Kevin Pelton]

[HoopStudies]

[jambalaya]

that is a lot of praise for AI.

not critiquing your views but simply adding my own, i've never been a big AI fan. not a hater, just somewhat less impressed. i acknowledge he is perhaps one of the five best point guards in the league to have on your team but i dont find his game especially interesting and i dont feel any need to give him any special credit for being small. i put his game in the same category of steve francis and baron davis as very talented but who almost always dribble too much and often take too many shots instead of maximizing team productivity. AI's career shooting percentages at closer to the league average for all players rather than the level of past superstars. i believe his over-dominant role on offense contributed to philly not winning a championship when it was possible. they (he) is easier to guard playing that way. yes to some extent he allowed the inclusion of other defensive minded players in the lineup and compensated for their offensive shortcomings but i think they took that strategy too far. he isnt michael jordan from the outside or driving to the hoop. i respect his talent and effort but i would rank at least one to two dozen other current stars as better contributors, based on a comprehensive look at all stats good and bad, to their team's success and, partly because of this, more enjoyable to watch- to me. he is a rare talent with lots of impressive stats that rank well in the historical record as the article you cite indicates but he has been given an extraordinary opportunity to rack up those numbers. (i see far more dominique in him than wilt. thomas is the closest comparison and he was willing and able to both shine and share more successfully than iverson and got his championships because of that blend.)

at basketball reference apbr stat query, it shows a off rtg of 95.9 and a def rtg of 102.4 for him. (correction: apparently this is for 2003-04 season.)

[HoopStudies]

[jambalaya]

ok, thanks for the clarification. i wasnt sure on which season that data mark "04" referred to but i believe you are right that it was 2003-04 season data. have you run your rtgs for him this year to date? is his off rtg above his def rtg for this year?


82 games shows his off at 100.5 and def at 100.8 this year. last year the numbers were off 91.3 and def 93.3. it looks like pace may have changed and he improved modestly on a net basis on this measure.

he is currently 11th in league on PER.

i did see 82 games shows a stong net per because of his massive scoring edge. using per difference plus on/off +/-, iverson ranks 18th in the league. looking at "Game by Game plus/minus Won-Lost records" he isnt in top 30. his 55.6% personal +/- win % is far lower than the top guys. no philly lineup with iverson makes top 30 lineups in league on team +/-.


i also went to nba.com and found his simple linear, not that refined, "efficency" rating puts him 9th in the league this year. his play this year appears to be a significantly higher quality than last year. he was 36th on this measure then and was neighbors with mike bibby and andre miller.

his tendex is also up significantly. at doug steele's site he is #1 on off., #11 on defensive, #6 by traditonal tendex formula, but #1 by cumulative ranking of off + def rankings. of course his numbers are helped by playing league leading 42.5 minutes a game.

last year he wasnt in the league's top 20 in league on any tendex measure and would have been about 5th best among point guards using the traditional tendex formula for his overall contribution. (he appears to have been accidentially omitted on doug's list last year so i had to compare his numbers to the existing list to rank him.) and he would not have been even top 20 for either shooting guard or point guard judged on per 48 minutes basis. that is the iverson i knew til- today.

this year his per 48 minute rankings are higher but put him as only 17th best in league by tranditional tendex, 3rd offense only, not top 20 on defense. very good, but not quite as high as with total production stats.


his 82 games on off +/- is up to +4 from last year's +1, but that is still modest as a positive and modest as a change.


your initial question about the quality of this season for him seems timely and is worth further review. i didnt know there was possibly a significant change in quality of performance from the past for him. based on these initial numbers i would upgrade my evaluation- of this year. but the different results (a mix of very strong and modest) shown on this measures make me uncertain of how far to upgrade the evaluation. i'll have to think and read more about it before deciding exactly how i judge it.


two more stats i found that cut against iverson. 82 games shows him with an effective FG% of less than 36% in the clutch (defined as final 3 seconds of shot clock) for all shots this year. and it is basically the same as last year. in the crunch (a similar sounding name but different measure defined as 4th quarter or overtime, less than 5 minutes left, neither team ahead by more than 5 points) it is 38% overall eff FG% and only 28% eFG% on jumpers. these stats have to be considered quite disappointing for a "superstar" (that is especially when they are supposed to deliver, where mere players, including pretty good ones, often don't) and certainly not helpful to his team's success. they are a pretty major counterbalance - i think- to consider along with his large overall individual stats.

[jambalaya]

most stats taken from 82 games.com

playmaker pts eff FG% eFG%crunch eFG% in clutch
iverson 30.4 0.445 0.36 0.38
payton 11.7 0.494 0.364 0.515
hinrich 15.5 0.453 0.421 0.361
l james 26.3 0.504 0.432 0.476
a miller 13.6 0.475 0.384 0.35
c billups 16.8 0.52 0.49 0.46
k bryant 28.1 0.481 0.459 0.351
j williams 10.9 0.515 0.491 0.621
d wade 24.2 0.484 0.44 0.462
s cassell 13.8 0.49 0.433 0.476
j kidd 13.6 0.467 0.443 0.565
marbury 21.8 0.501 0.508 0.367
s francis 21.4 0.434 0.378 0.414
s nash 16.1 0.563 0.573 0.566
bdavis(gs) 17.4 0.529 0.478 0.533
m bibby 19.5 0.498 0.457 0.48
t parker 16.7 0.508 0.441 0.316
g arenas 25.6 0.492 0.412 0.476

iverson: highest scoring average, but second lowest eFG%, lowest eFG% in crunch, 6th lowest eFG% in clutch.

snash, c billups and j williams have very strong shooting percentages


playmaker ast reb draw foul pass rtg hands rtg
iverson 7.8 4 14.50% 10.7 22.7
payton 6 3.1 9.40% 10.5 22.5
hinrich 6.7 4.1 6.20% 10.8 24.2
l james 7.2 6.9 13.50% 10.4 23.3
a miller 6.5 4.2 12.20% 15 27.6
c billups 5.9 3.6 12.50% 9.1 21.5
k bryant 6 6 16.50% 5.9 15.4
j williams 5.9 1.6 3.90% 13.8 28.6
d wade 6.9 5.3 19.20% 10.5 20.3
s cassell 5.4 2.6 6.50% 10.4 22.8
j kidd 8 7.1 4.90% 12.3 27.2
marbury 8.2 3.1 15.30% 10.7 24
s francis 7 5.9 15.00% 12.3 23.5
s nash 11.5 3.4 5.30% 24.6 44
bdavis(gs) 7.4 3.6 9.60% 13 33.5
m bibby 6.7 4.3 10.10% 9.4 23
t parker 6.1 3.9 9.50% 9.5 19.4
g arenas 5.2 4.6 14.60% 6.8 17

iverson: 4th highest assists, but genrally in the middle of the pack elsewhere- 9th highest rebounds, 6th highest drawing fouls, 8th highest pass rtg, 11 highest hands rtg.

d wade is the leader drawing fouls by a substantial margin. nash is galaxies ahead of most on passing and hands. j williams and b davis are surprisingly (to me) strong on these measures.


playmaker off def win% off/def rtg in crunch
iverson 100.5  100.8 55.6 103/101
payton 102.7 101.8 57.6 93/96
hinrich 93.8 93.7 51.6 102/93
l james 96.6 93.8 55.2 99/97
a miller 98.6 96.5 55.9 110/98
c billups 92.1 87.8 63.6 98/86
k bryant 100 100.5 54.5 105/100
j williams 96.1 94.4 53.3 96/97
d wade 102.7 95.7 74.6 121/95
s cassell 98.4 95.4 51.1 113/82
j kidd 95.1 92.5 55.4 111/90
marbury 99.3 99.5 51.5 92/120
s francis 101.3 100.8 51.6 107/113
s nash 116.3 104.2 81.3 127/103
b davis gs 103 99 64.7 106/79
m bibby 104 100.6 60.3 107/109
t parker 96.9 86.5 73.9 99/119
g arenas 102.1 100.3 59.1 119/104

note the superior team +/-'s when d wade, s cassell, j kidd, s nash b davis and g arenas play in the crunch. iverson isnt strong or that bad. bad crunch performing players/teams include marbury/knicks and surprisingly t parker/spurs. (of course this is heavily influenced by the contributions of the other four guys on the floor.)


iverson's most similar comparisons i would say may be with l james, k bryant, g arenas, d wade, b davis and steve francis. (lead playmakers not necesarily point guards.) i'd say lebron james and d wade are better by the numbers overall because of shooting numbers and roughly equal on passing and hands. b davis probably is as well right now- at golden state. it isnt as easily to say better than for bryant and arenas. they have both pluses (generally shooting) and minuses (passing and hands) compared to iverson. i am more willing to say steve francis isnt because his shooting isnt that strong but it is fairly close.

nash is a different type (not as much scoring) but may be the best overall point guard or the most valuable to his team. kidd and cassell still have a game, especially in the crunch.

[jambalaya]

i tried to quickly assemble data for various measures of productivity and effectiveness. i havent tried to carefully and precisely weight the various data, crunch it and solve the question based on all their values and others that could be added. i've just discussed the data and shared some partial judgments about it. a comprehensive (pre-built and hopefully neutral) database query system could obviously handle it better than ad-hoc balancing. what do other systems say about allen iverson this year compared to his other playmaking counterparts or the rest of the top players in the league?

[jambalaya]

one more comment. (really same general viewpoint expressed differently.)

i think philly needed a superstar, believed AI was that person, paid him to be a superstar, played him like one, and partly because of his high minutes and high possessions used got results they interpreted as being superstar-like, which was re-inforced by the finals appearances.

if iverson's minutes and possessions used were adjusted downward to that of other "star" playmakers, i think the case for him being "exceptional" would be weakened. i dont have the '05 possessions used numbers handy (someone else probably does. last year i believe he used 15-50% more possessions per 48 minutes than many of his playmaking counterparts), but just looking at FG attempts this year AI is getting close to 24 a game, L James about 20, D Wade 17, B Davis only 14. but of course AI is playing 42.5 minutes and taking all those shots so you cant totally take that reality away from him. that is a big part of his story. but maybe that is somewhat apart from "greatness".

[Ben F.]

I know the case Dean made in his book was that Iverson's high usage rate helped the other players on his team maintain high levels of efficiency. The blow he took to his efficiency because of the high usage rate wasn't nearly what the team would have taken had the other peripheral players have had to step up.
_________________
XOHoops - A New Kind of Fantasy Basketball

[Carlos]

Good question. He certainly seems to be enjoying his finest statistical season so far. Using the Basketball-reference and Knickerblogger stats you see that his PSA is at 1.05 (up from 0.96 and his highest since his second season), his PER is at 23.1 (second highest in his career). His assists are up too, same as his scoring (which is almost like in his younger days) even though his FGA per minute are down. He is shooting less, getting more assists and making a higher percentage of his shots. The move to point guard has certainly helped him.

[Kevin Pelton]

[jambalaya]

i looked at iverson's teammate's crunch and clutch eFG%s.
it wont end the argument about whether his taking many of these shots helps the team or helps the other guys look better by letting them off from taking more of the tough shots, but, based on what they do take, he has got 5 guys doing pretty well at these times. better percentages than iverson. maybe he could cultivate them a little more.

crunch eFG% clutch eFG%

iverson 0.36 0.38

davis 0.4 0.5
korver 0.516 0.51
dalambert 0.439 0.5
iguodala 0.471 0.474
mckie 0.551 0.75
jackson 0.355 0.294
webber 0.372 0.278
salmons 0.437 0.25
green 0.379 0.262

williamson 0.381 0.367
thomas 0.478 0.333
skinner 0.386 0.25


so far webber isnt an improvement on this measure over the three guys he replaced. but most teams can't effectively go inside in the crunch or clutch and the shooting percentages tend to be lower there. many of iverson's perimeter options look competent enough to take more share depending who was the best look.

[Ben F.]

What does what you looked at have to do with anything? The point is that without Iverson using so many possessions, you'd force the role players to use more possessions than they are comfortable with, forcing the team's efficiency down.

That has nothing to do with clutch time.
_________________
XOHoops - A New Kind of Fantasy Basketball