Post by eric on Jan 30, 2016 19:17:14 GMT -6
So I wanted to figure out how no-shoot players and scoring options affected everyone else's usage. To that end, I set up this DC:
1. starting point guard
2. starting shooting guard
3. starting small forward
4. starting power forward
5. starting center
6. backup shooting guard and point guard (roster SG)
7. backup center, power forward, and small forward (roster C)
8. third string small forward, shooting guard, and point guard (roster SF)
9. third string center and power forward (roster PF)
10. not on the depth chart (roster PG)
11. not on the depth chart (roster SF)
12. not on the depth chart (roster SF)
I then ran a sample (2639 seasons) (why 2639? because f!@# your rules that's why) for every person as the third option except the last two, who were always first and second options because they never played. Then I did the same thing for the first ten players as second options, then as first options, keeping the other options on non playing players.
.
First finding: setting any of 8 9 or 10 as third option didn't change starter usage at all.
Second finding: therefore, here are the baseline usage values for inside/balance/outside offense, given in (player uses / team uses) * ((team minutes / 5) / player minutes) * 10:
Third finding: setting 6 or 7 as third option DID change starter usage, so if you're trying to absolutely maximize a particular starter's usage by burying other scoring options you'd better go deep.
Fourth finding: the amount of usage granted by being a particular option is NOT a function of base usage, but a function of position. However, if a particular position in a particular offense gets more usage as a first option than another, it will also for second and third. For demonstration, here are all three options for inside offense.
Note that SG has the highest gain and the lowest usage, then SF is a bit lower from a bit higher, but then PG is the lowest (instead of the second lowest) with the second highest usage, and PF/C gains a medium amount with by far the highest usage. Here are the values for first option for all three offenses, and you can see the pattern of lack of pattern continue:
Fifth finding: The question of usage vs. efficiency is contentious IRL. In sim life, it is a fact: extended usage does not harm a player's efficiency. On average a starter saw +.001 pts/tsa by being promoted from a non option to a #1 option, and only two starters saw decreases. I do not know how to calculate the ± on these measurements, but I am reasonably sure that there is no usage/efficiency tradeoff in the software at all and this very tiny positive effect is just noise.
Sixth finding: Going from non-option to primary option increased every starter's MP/g by an average of 0.2 (about 14 seconds a game). Because there are no offensive fouls in the software a player trying to score more would never get in more foul trouble. It is conceivable that players trying to do more on offense would try less on defense and therefore get in less foul trouble, but as a spot check the center and power forward actually got quite a bit more blocks playing as the #1 option. It's a small effect anyway so I'm not going to put too much effort into researching it.
Seventh finding: These additional touches have to come from somewhere: there is, and will always be, only one ball. It turns out they are taken from two sources: non-starters and the other starters. The following chart is given for inside scoring, and should be read that when the column is the #1 option, the rows lose that amount of usage.
This is another function of position, and a separate one from the scoring option one. We can also see that the touches are mostly taken from other starters, only about a quarter are taken from reserves.
Eighth finding: We can adjust only a player's usage (and not efficiency) by adjusting their Jumping attribute. There are no shot clock violations in the software, teams always get a shot off. I whipped up a little analysis of scoring efficiency by shot clock using the play by play logs. Unfortunately the only time stamp is on possession start, so a miss will seem to have taken longer than a made shot (because rebounds take time to corral), but I used a moving three second average and I'm pretty sure that sorted it:
Bottom line is that going deeper into the shot clock isn't a terrible thing. Anyway, here are the changes in usage for an inside offense when only the center has decreased Jumping, and when both bigs have decreased jumping:
This mercifully follows the pattern of uses taken away by the big being a scoring option, which makes sense because subtraction is just addition with a sourpuss. Most of the touches go to starters, some go to non starters. The "both bigs non scorers" option multiplies by two pretty well down the line for starters, but doesn't add as much to the other non scoring big as expected, and the only other place for those touches to go is to non starters and so they do: instead of C and PF being at +17 - 61 = -44 they're at -48, so instead of others getting 10 + 10 = 20 touches they get 10 + 10 + 4 + 4 = 28 = 27 within rounding error. That is all to say that playing non scoring starters has diminishing returns when it comes to generating more touches for the other starters.
.
What Does It All Mean?
There are three factors for how much a player wants to shoot: the offense, how much they gain from being an option, and how much every other player wants to shoot. None of these are functions of the others, so you've really got to do math homework to predict how your guy's going to change. In the interests of brevity (too late), here is a table you can use to get a rough guess. The first column is the position's base value, the second is how much they gain from being a #1 option, the third is how much they lose from someone else being a #1 option (#2 and #3 options are roughly 75% and 40% of the effect in each case). Therefore, if you set the given position as your #1 option and put your #2 and #3 options as third stringers, you obtain the max. If you instead set three other starters as scoring options, you obtain the min.
1. starting point guard
2. starting shooting guard
3. starting small forward
4. starting power forward
5. starting center
6. backup shooting guard and point guard (roster SG)
7. backup center, power forward, and small forward (roster C)
8. third string small forward, shooting guard, and point guard (roster SF)
9. third string center and power forward (roster PF)
10. not on the depth chart (roster PG)
11. not on the depth chart (roster SF)
12. not on the depth chart (roster SF)
I then ran a sample (2639 seasons) (why 2639? because f!@# your rules that's why) for every person as the third option except the last two, who were always first and second options because they never played. Then I did the same thing for the first ten players as second options, then as first options, keeping the other options on non playing players.
.
First finding: setting any of 8 9 or 10 as third option didn't change starter usage at all.
Second finding: therefore, here are the baseline usage values for inside/balance/outside offense, given in (player uses / team uses) * ((team minutes / 5) / player minutes) * 10:
pos ins bal out
PG 184 225 278
SG 176 204 249
SF 179 212 189
PF/C 228 175 151Third finding: setting 6 or 7 as third option DID change starter usage, so if you're trying to absolutely maximize a particular starter's usage by burying other scoring options you'd better go deep.
Fourth finding: the amount of usage granted by being a particular option is NOT a function of base usage, but a function of position. However, if a particular position in a particular offense gets more usage as a first option than another, it will also for second and third. For demonstration, here are all three options for inside offense.
pos 3rd 2nd 1st
PG 43 76 100
SG 58 101 132
SF 56 99 130
PF/C 47 79 104Note that SG has the highest gain and the lowest usage, then SF is a bit lower from a bit higher, but then PG is the lowest (instead of the second lowest) with the second highest usage, and PF/C gains a medium amount with by far the highest usage. Here are the values for first option for all three offenses, and you can see the pattern of lack of pattern continue:
pos ins bal out
PG 100 97 64
SG 132 127 85
SF 130 127 118
PF/C 104 133 121Fifth finding: The question of usage vs. efficiency is contentious IRL. In sim life, it is a fact: extended usage does not harm a player's efficiency. On average a starter saw +.001 pts/tsa by being promoted from a non option to a #1 option, and only two starters saw decreases. I do not know how to calculate the ± on these measurements, but I am reasonably sure that there is no usage/efficiency tradeoff in the software at all and this very tiny positive effect is just noise.
Sixth finding: Going from non-option to primary option increased every starter's MP/g by an average of 0.2 (about 14 seconds a game). Because there are no offensive fouls in the software a player trying to score more would never get in more foul trouble. It is conceivable that players trying to do more on offense would try less on defense and therefore get in less foul trouble, but as a spot check the center and power forward actually got quite a bit more blocks playing as the #1 option. It's a small effect anyway so I'm not going to put too much effort into researching it.
Seventh finding: These additional touches have to come from somewhere: there is, and will always be, only one ball. It turns out they are taken from two sources: non-starters and the other starters. The following chart is given for inside scoring, and should be read that when the column is the #1 option, the rows lose that amount of usage.
pos PG SG SF PF/C
PG x 15 15 12
SG 17 x 23 19
SF 19 25 x 21
PF/C 21 30 30 x
total 22 33 33 29This is another function of position, and a separate one from the scoring option one. We can also see that the touches are mostly taken from other starters, only about a quarter are taken from reserves.
Eighth finding: We can adjust only a player's usage (and not efficiency) by adjusting their Jumping attribute. There are no shot clock violations in the software, teams always get a shot off. I whipped up a little analysis of scoring efficiency by shot clock using the play by play logs. Unfortunately the only time stamp is on possession start, so a miss will seem to have taken longer than a made shot (because rebounds take time to corral), but I used a moving three second average and I'm pretty sure that sorted it:
Bottom line is that going deeper into the shot clock isn't a terrible thing. Anyway, here are the changes in usage for an inside offense when only the center has decreased Jumping, and when both bigs have decreased jumping:
pos C PF and C
PG 8 17
SG 14 28
SF 12 25
PF 17 -48
C -61 -48
total -10 -27This mercifully follows the pattern of uses taken away by the big being a scoring option, which makes sense because subtraction is just addition with a sourpuss. Most of the touches go to starters, some go to non starters. The "both bigs non scorers" option multiplies by two pretty well down the line for starters, but doesn't add as much to the other non scoring big as expected, and the only other place for those touches to go is to non starters and so they do: instead of C and PF being at +17 - 61 = -44 they're at -48, so instead of others getting 10 + 10 = 20 touches they get 10 + 10 + 4 + 4 = 28 = 27 within rounding error. That is all to say that playing non scoring starters has diminishing returns when it comes to generating more touches for the other starters.
.
What Does It All Mean?
There are three factors for how much a player wants to shoot: the offense, how much they gain from being an option, and how much every other player wants to shoot. None of these are functions of the others, so you've really got to do math homework to predict how your guy's going to change. In the interests of brevity (too late), here is a table you can use to get a rough guess. The first column is the position's base value, the second is how much they gain from being a #1 option, the third is how much they lose from someone else being a #1 option (#2 and #3 options are roughly 75% and 40% of the effect in each case). Therefore, if you set the given position as your #1 option and put your #2 and #3 options as third stringers, you obtain the max. If you instead set three other starters as scoring options, you obtain the min.
inside
pos base gain loss max min
PG 184 100 14 284 154
SG 176 132 20 308 133
SF 179 130 22 309 132
PF/C 228 104 27 332 170
balanced
pos base gain loss max min
PG 225 97 16 322 191
SG 204 127 24 331 152
SF 212 127 26 339 156
PF/C 175 133 22 308 128
outside
pos base gain loss max min
PG 278 64 23 342 229
SG 249 85 30 334 185
SF 189 118 25 307 135
PF/C 151 121 20 272 108
![[Account Deleted] Avatar](http://storage.proboards.com/5589790/avatar/ZhaFVHTxVPwmyuOiOKEd.png)
