## performance rating and trade value

Expand Messages
• i recognize that linear weight systems are a dime a dozen and are considered somewhat antiquated. i intend to read more about the winval discussion here. but
Message 1 of 66 , Mar 5, 2004
i recognize that linear weight systems are a dime a dozen and are
winval discussion here. but i wondered if i might share a revised
version of my linear weight performance rating system i developed
previously/elsewhere and get some feedback on it, if you care to
comment. if you can point me to specific past threads that deal with
certain elements/issues in more detail that would help me catch up
with the current discussion / consensus i'd appreciate it.

performance = (PTS - VOP*FGmsd - .5*VOP*FTmsd + 1.25*VOP*st + .5*as
+ .33*VOP*bl + VOP*reb - 1.25*VOP*to - .81*pf) / GamesPlayed

(where value of possession VOP is equal to 1.04. based on team point
and possession data at 82gams.com, the current league average is
1.038620421 points per possession, rounded to 1.04.)

i dont adjust for minutes played (what you actually did rather than
what you might be capable of) or pace (only a 5-7% spread in the
league?). but you could.

explanation of specific weights:

about 50% of misses FTs result in a change of possession as shown in
VORP formula: .5*-1.04= -.52 value of missed FT

steals or turnovers i assume
on half the team has improved prospects for points (on fast break)-
60% for a bucket and 20% for 2 foul shots:
.5 * [(.6*2 for bucket=1.2) + [(.2*1.5 points)average for two FTs)
=.3)] =.75

on the other half i assume it occurs in a way that leads basically
to a normal possession
.5*1=.5

.75+.5= 1.25 points expected per steal or turnover

rebound= 1 possession * value of possession (1.04)= 1.04

personal fouls: on average yield 1.0893 FTs per foul using league
stats *0.7475 league FT%= approx .81 points

assists (.5) and blocks (33 *1.04= .342) basically as previously
discussed.

techs and flagrants are left out as rare.

at the end of this post i list the top 50 players according to the
new performance rating. i also try to capture trade value with
factors to adjust that performance rating based on rarity of that
performace level (my thinking being that in trades guys with high
performance rankings are worth a premium greater than the difference
in the performance ranking from the average alone because they are
rare and stars win games), and that age and size also matter
independent of performance.

more specifically i estimated that players with a performance ranking
of 20 or above would command a 1.35 factor premium, over 15 a 1.2,
above 10 a 1.1, between 5 and 10 a normal value of 1 and below
that .8 (because those level players are common enough to be
discounted in value for trading purposes).

age was weighted as follows <22 1.3 (valuable because of youth but
still a little raw and learning) <27 1.4 (highest-young prime), <32
1.2 (older prime, worth about 15% less), <35 1.0, older than that .75.

size i divided into three groups because size earns a trade premium:
under 6-8 1.0, 6-8 - 6-11 and not really a power player 1.15, power
players 1.35.

(please give me feedback on these weights.)

Player New Performance Rarity Age Size Trade TRank
garnett,kevin 27.55 1.35 1.2 1.35 60.25 1
webber,chris 23.31 1.35 1.2 1.35 50.98 3
duncan,tim 23.12 1.35 1.2 1.35 50.56 4
brand,elton 21.76 1.35 1.4 1.35 55.53 2
stojakovic,predr20.95 1.35 1.4 1.15 45.54 5
o'neal,shaquille20.54 1.35 1.2 1.35 44.92 6
nowitzki,dirk 19.17 1.2 1.4 1.35 43.48 7
mcgrady,tracy 18.80 1.2 1.4 1.15 36.32 14
miller,brad 18.77 1.2 1.2 1.35 36.48 13
boozer,carlos 18.33 1.2 1.3 1.35 38.60 9
wallace,ben 18.04 1.2 1.2 1.35 35.06 16
randolph,zach 17.97 1.2 1.4 1.35 40.75 8
marion,shawn 17.92 1.2 1.4 1.15 34.62 17
bryant,kobe 17.39 1.2 1.4 1.15 33.61 18
kirilenko,andrei17.36 1.2 1.4 1.15 33.53 19
martin,kenyon 16.95 1.2 1.4 1.35 38.45 10
abdur-rahim,shar16.81 1.2 1.4 1.35 38.12 11
malone,karl 16.73 1.2 0.85 1.35 23.04 32
marshall,donyell16.43 1.2 1.20 1.35 31.95 20
o'neal,jermaine 16.36 1.2 1.4 1.35 37.10 12
ming,yao 15.87 1.2 1.4 1.35 35.99 15
dampier,erick 15.86 1.2 1.2 1.35 30.83 23
redd,michael 15.76 1.2 1.4 1 26.48 24
allen,ray 15.62 1.2 1.2 1 22.50 34
kidd,jason 15.13 1.2 1.2 1 21.78 39
stoudemire,amare15.09 1.2 1.3 1.35 31.77 21
finley,michael 15.02 1.2 1.2 1.15 24.87 30
gasol,pau 14.96 1.1 1.4 1.35 31.11 22
odom,lamar 14.86 1.1 1.40 1.15 26.32 25
maggette,corey 14.81 1.1 1.4 1.15 26.24 26
pierce,paul 14.80 1.1 1.4 1.15 26.20 27
cassell,sam 14.75 1.1 1.00 1.00 16.23 x
marbury,stephon 14.56 1.1 1.4 1 22.43 35
lewis,rashard 14.23 1.1 1.4 1.15 25.21 28
iverson,allen 14.13 1.1 1.2 1 18.65 51
jefferson,richar14.05 1.1 1.40 1.00 21.64 38
bibby,mike 13.93 1.1 1.4 1 21.45 40
thomas,kenny 13.89 1.1 1.2 1.15 21.09 43
davis,baron 13.79 1.1 1.4 1 21.24 41
brown,p.j. 13.76 1.1 1 1.35 20.43 45
jamison,antawn 13.63 1.1 1.2 1.15 20.69 44
richardson,jason13.39 1.1 1.4 1 20.62 46
james,lebron 13.37 1.1 1.30 1.15 21.99 36
miller,andre 13.13 1.1 1.20 1.00 17.33 x
camby,marcus 13.12 1.1 1.20 1.35 23.39 31
vanhorn,keith 12.97 1.1 1.20 1.15 19.68 47
harpring,matt 12.95 1.1 1.2 1.15 19.65 48
anthony,carmelo 12.90 1.1 1.30 1.15 21.22 42
ilgauskas,zydrun12.84 1.1 1.2 1.35 22.88 33
walker,antoine 12.84 1.1 1.2 1.15 19.49 49
artest,ron 12.57 1.1 1.4 1 19.36 50
wallace,rasheed 12.26 1.1 1.2 1.35 21.84 37
magloire,jamaal 12.22 1.1 1.40 1.35 25.42 29
mashburn,jamal 12.12 1.1 1.2 1.15 18.40 52

the top 7 on performance rating and trade value are the same except
brand moves up a bit because he is so young. in the next group,
young bigmen pass by a some of the perimeter players with slightly
higher performance rankings. that seems reasonable to me.

feedback?
• I will agree that Manu has been gambling since about mid-season when he moved to the bench more than he did in the beggining of the season (in fact much more).
Message 66 of 66 , Mar 23, 2004
I will agree that Manu has been gambling since about mid-season when
he moved to the bench more than he did in the beggining of the season
(in fact much more).

I actually thought he cut down on the gambling A LOT from last
season, but there was a game in New Orleans, when he came off the
bench and did it at a crucial point of the game and might have been a
huge reason the Spurs lost the game. He usually never made those
type of mistakes down the stretch of games.

Either way, I think he is generally a very solid defender. Even
though he gambles a bit these days, he still creates a lot of
turnovers via charges and steals.

And about 2001, I know Bowen was not around. But I was trying to
kind of hint towards the TWIN TOWER approach maybe being the primary
reason an SG/SF always is in the top TENDEX defensive ratings since
2001 (maybe even before that) if we were to check.

If Bowen is the reason for the SG's being held so low, then why is
the SG production PER at a net positive at +2.5? And negative at the
SF position (or was once close to it)? He certainly doesn't put up
LARGE PER and Turkoglu altough is scoring more these days and
grabbing rebounds, I would have to assume Manu, Turk, AND Bowen are
all playing solid defense. If there is any weakness in the D its
with Parker slightly, and even MORE SO at the backup PG position for
the Spurs.

Of course the two 7 footer approach helps them all become a little
more risky etc..... But I don't see how Manu and Hedo cannot being
doing solid defensively, when the PER, defensive ratings are solid
across the board? Isn't it safe to conclude the Spurs are getting
reasonable production on both ends with the THREE MAN ROTATIOn at the
SG and SF? And the reason the SF might be low is solely because of
Bowens lack of scoring at times?
Your message has been successfully submitted and would be delivered to recipients shortly.