is there a relationship between a player's fg%, ponts-per-shots, and shots-per-game?

[superheadcat]

many are debating the shots certain players should take based on their pg% and pps relative to other teammates.

i am curious to know whether such a relationship exists, and if so, what kind of relationship it is. so i did some number crunching.

first, some preliminaries:
1. i chose 4 other teams to benchmark mavs: kings, tpup, spurs, and pacers. i exclude lakers for the obvious reasons.
2. obviously people r interested in offense numbers, thus i check only the players who r main offense powers for their teams (that is why only kings and mavs have 5 players in the study, pacers and spurs have 4, and tpup have only 3 to carry most of their offense)
3. i used only this season's stat, partly because i am lazy, but mostly because this maybe the most relevant and indicative stat set.

now, the raw data:
name------fg%--------pps--------------------spg
td----------513--------1.289051095---------17.125
manu------408--------1.194312796---------10.82051282
nesterovic470--------1.052219321---------8.704545455
tp---------- 431--------1.082452431---------12.78378378

bibby------444--------1.262295082---------13.3902439
divac-------498--------1.23655914----------9.073170732
bjax--------441--------1.161885246---------12.2
peja--------493--------1.445868946--------17.55
bmiller-----514--------1.398584906---------10.34146341

artest------423--------1.161538462--------15.11627907
jo-----------441--------1.116306954---------18.95454545
harrington445--------1.116232465---------11.34090909
rmiller------431--------1.333333333--------7.704545455

kg----------494--------1.193586698---------20.53658537
spree------426--------1.109955423---------16.41463415
cassell-----499--------1.250366032---------16.65853659

dirk---------442--------1.261217949--------16.42105263
nash-------433--------1.276785714---------10.92682927
fin----------433--------1.139261745---------15.68421053
aj-----------526--------1.299424184---------12.11627907
aw---------418--------1.004354136---------16.02325581

now, some observations:
for spurs, nesterovic has fg% second only to td, but he shots the fewest, if looking at pps and spg, the relation seems more positive correlated, except manu and tp's slight discrepency. notice nesterovic's pps is only 0.03 lower than tp's, yet he shot 30% less.

for kings, again, the relationship between pps and spg is more positive than that between fg% and spg. kings' top2 fg% shooter (divac and miller) shoot the least among the 5, the median pg% shooter (peja) shoot significantly more. the relationship is more positive between pps and spg. but again, miller's pps is 2nd best, yet shoot only 2nd fewest.

for pacers, their fg% are quite similar, yet the spg fluctuates significantly. in terms of pps, the relationship becomes almost negative, the one with the most shots (jo) has almost the lowest pps, and he took more than twice of the shots rmiller takes, who has the highest pps.

for tpup, fg% is perfectly positive correlated, for the pps, cassell is higher than kg, yet take 25% less shots.

for mavs, aw is a noticeable outliner. and the one who currently takes most shots (dirk) are not best shooters. the relationships among fg%, pps, and spg are ambiguous.

ok, now some conclusions/opinions:
is there a positive relationship between fg% and spg, and between pps and spg?
i think so. although the stat here seems to only weakly support this claim. and if the relationships do exist, i believe the relationship between pps and spg is stronger.

does this mean mavs' shooter selection need to be improved?
yes. but not necessarily mean it has to be aligned strictly by pps or fg%. (do you really want aj to shoot more than dirk?)

what does this mean for aw?
generally speaking, fg% or pps alone should not be used to argue how much shots one should take. but aw's fg% and pps contrasted with his spg constitute an obvious outliner in mavs' stat, and i believe his deviation is too significant to be ignored. my suggestion to aw: continue what you started doing recently: make high % shots, or don't shoot.

finally, an appendix:
suppose aw's offense won't improve any more this season, how will that hurt mavs?:
well, suppose his "ideal" number of shots are 8 to 10, everything else held constant, then he should cut 6 or 7 more shots from his current stat, let's say he "transfers" those 6, 7 shots to aj, that will result in about 2 ppg more for mavs.

so far this season, mavs average 103 points, second in the league. with thos 2 more points, mavs will tie with kings to be no. 1. unfortunately, mavs also allowed an average of 99.7 points, second worst only to magic by 0.1 points.

does aw's reducing shots also somehow help mavs d'up in reducing opponents' points? if so, i'll say his reduction in shot numbers is so much more urgent and worthwhile.

but my overall felling is, mavs' biggest problem right now is its overall defense, which is everyone, including aw, 's responsibility. if that area can be improved significantly, i don't think it is such a huge damage to mavs, as claimed by some posters, if aw won't reduce his shots any more.

although i definitly hope aw can help upping mavs' ppg by an extra 2.

edit: better format

[dude1394]

Superheadcat. Great stuff. The numbers don't take into account ft's right?

[superheadcat]

Originally posted by: dude1394
Superheadcat. Great stuff. The numbers don't take into account ft's right?

fg% does not, but pps does.

[dude1394]

What about spg?

[superheadcat]

Originally posted by: dude1394
What about spg?

no, it does not include ft.

[mercury_rev]

so far this season, mavs average 103 points, second in the league. with thos 2 more points, mavs will tie with kings to be no. 1. unfortunately, mavs also allowed an average of 99.7 points, second worst only to magic by 0.1 points.

does aw's reducing shots also somehow help mavs d'up in reducing opponents' points? if so, i'll say his reduction in shot numbers is so much more urgent and worthwhile.

A handful of less missed 3's per game may translate into one or two less easy transition buckets for the opposition. Don't have any numbers for this, but might it shave a couple of points off that 99.7?

[grndmstr_c]

Nice work, SHC.

So here's the next question. If Walker did cut down his shot attempts and set up other guys for buckets, and the Mavs did net 2 points a game, how much of a difference would that really make?

It seems to me the best way to get an idea of how much those two points would help would be to look at the average margin of victory for other teams and find out what kind of record tends to be associated with MOV's in the range of what Dallas has (+3.3), and what they "could" have if we got those two points (5.3). We can take it season by season.

2003-4 (margin of victory, record, winning percentage)
Dallas: +3.3, 27-16, .628
Minnesota: +5.3, 29-12, .707

2002-3
SA: +5.4, 60-22, .732
NJ: +5.2, 49-33, .598
Indiana: +3.5, 48-34, .585

2001-2
Minnesota: +3.4, 50-32, .61

Now, that's a limited sample size (not sure where to find the info for previous seasons), and the results could vary somewhat if we went farther back, but if we average what we've got, we find that an average margin of victory in the range of 3.3 has been associated with an average winning percentage of .608, whereas a MOV in the range of 5.3 has been associated with a winning percentage of .679. Over the course of an entire season that's worth about 6 wins. Over the rest of this season, that's about three extra wins, and that's without improving in any phase of the game except giving some of Walker's shots to someone who's a better shooter. No defensive improvement, no chemistry, no Dirk's shooting coming around, nothing outside of giving a few of Walker's attempts to somebody else, and the team could expect to get 3, give or take, extra wins. As tight as the race for playoff seeding figures to be, I'm not sure we can afford to not have Walker cut down on his shooting.

[jayC]

Im sorry but just two words tired head.

[thebac]

grandmaster, your analysis doesn't hold water.

Even assuming that Jamison remains as efficient with more shots (unlikely) and assuming that other team dynamics stay constant, let's look at the top 6 teams in point differential the past two seasons and six years ago:

2002-2003
1 Dallas +7.8, 60 wins
2 Sacramento +6.5, 59 wins
3 San Antonio +5.4, 60 wins
4 New Jersey +5.2, 49 wins
5 Detroit +3.7, 50 wins
6 Indiana +3.5, 48 wins

2001-2002
1 Sacramento +7.6, 61 wins
2 LA Lakers +7.1, 58 wins
3 San Antonio +6.2, 58 wins
4 Dallas +4.3, 57 wins
5 New Jersey +4.2, 52 wins
6 Minnesota +3.4, 50 wins

1996-1997
1 Chicago +10.8, 69 wins
2 Utah +8.8, 64 wins
3 Seattle +7.7, 57 wins
4 Miami +5.5, 61 wins
5 Atlanta +5.4, 56 wins
6 Detroit +5.2, 54 wins

What's immediately obvious is that greater point differential (pd) does NOT necessarily equal more wins. As a matter of fact, you had one example yourself, NJN in 02-03, where the pd is more than what the Mavs might hope to get if Walker took fewer shots (even if the Mavs started to average 2 more points from now on until the end of the season, their pd would increase by less than 1 ppg), and they had only 49 wins. On the other hand, DET and IND in 02-03 and MN in 01-02 had approx. the pd the Mavs have now and they won an average of 49+ games. Not much of a difference.

Of course, it's hard to compare across teams, anyway. If your team tends to be in blowouts (i.e., a higher percentage of the games, both wins and losses, are blowouts) as opposed to being in close games, and that team happens to be winning team, the pd of the one in blowouts is much greater than the one of the team with many close games, even with the wins being equal. There is no particular number of wins attached to a given pd. Other factors matter--is it a defensive-minded team (note that MIA in 96-97 won more than SEA despite a much lower pd, and SAN won much more than NJN in 02-03 despite comparable pd's--of course the Mavs in 01-02 are a counterexample as they won five more games than NJN even though their pd was virtually the same, and I don't think it's because the Mavs were more of a defensive team than NJN)? Is it a team that is good at winning close games (probably not as important as in football)?

In short, the problem with pd as a statistic is that it's not that meaningful. It's hard to compare across time, because scoring has decreased so much in the NBA (as recently as 1994-95, the average team scored 101.4 ppg), so the blowout issue comes up. You also can't look at cross-sections because there are many team-specific components that determine pd. Thus, it's hard to make any meaningful predictions based on what the pd might be. If the Mavs win each of their remaining games by 1 point, they end up 66-16, but their pd will have gone down to just 2.2 or so. On the other hand, if they traded wins and losses until the end of the season (starting with a win), with each win by 15 points and each loss by 5 points, their pd at the end of the season would be more than 6, but they would end up just 47-35. The pd is such a highly aggregated statistic, it has lost most of its meaning.

Having said all that, AW should shoot less.

[grndmstr_c]

thebac, do you really think I was not aware that NJ was an outlier in the stats I presented. The fact that NJ's record last year did not fit a perfect curve on a graph does not undermine my argument. It only re-emphasizes a point that I had already tried to make in my original post: that I was analyzing a very small sample size and that small sample sizes inherently carry with them an increased danger of unreliable information.

The fact is that pd is quite obviously positively correlated with winning percentage, as is evident in your own examples. Whether or not there are potential outlliers (like NJ), and whether or not there is clear evidence for causality in the relationship, it remains a verifiable fact that winning and losing can be predicted from pd. Doesn't mean that the predictions are perfect, but it also is a far cry from the argument that they don't hold water. Your points are good ones, but just because I didn't spend the time making them, don't assume that I wasn't aware of them.

[superheadcat]

regardless a team's offensive power, it is quite a dangerous practice if the team allows opponents 3-digit scoring every game, no matter what.

is it worth it that aw continue to cut down his shots? absolutely, definitely, without a doubt.

it is just that is not the #1 concern for mavs right now.

what would mavs' #1 priority be? dirk says it best on the top right corner of page 21 of here

[Murphy3]

Of course PPS should play a role in Walker's FGA's whenever he's playing on a team full of players that are much more efficient than AW. It's just common sense. However, FGA's shouldn't strictly be based on PPS. Some players get most of their FGA's on cleaning up other player's misses. That's how Najera was able to shoot over 50% a couple of years back. Should he have been looked for offensively more often? No, of course not.

No, adding a point or two to your total doesn't help the mavs defensively. However, it helps to counter their defensive deficiencies.

[PubaNWO]

Put it this way!!!

It's easier for AW to get Dirk and others easier shots......opposed to AW since he isn't an efficient scorer....

simple....get our better scorers shots and preferably easier ones...enhancing your strengths and minimizing your weaknesses...

[chumdawg]

I read in today's DMN that the Mavs are 13-1 when the Big Five all score in double digits. How's that for an argument that getting everyone involved in the offense is a very, very good thing?