Neon1 v1.8 Rosters (360)....cracking the code on shot selection.

[Kiba]

Re: Neon1 v1.8 Rosters (360)....cracking the code on shot selection.

PIP 78? maybe you have to improve your defensive attitude man.

lol yeah i'm not give more eye on D...cuz i want to see how's cpu play the ball...
ok i'll try again later with another team see what happen if i play more D..
cuz i'm feel good w/ this rosters set...thanx neon

[Neon1]

Re: Neon1 v1.8 Rosters (360)....cracking the code on shot selection.

Neon1 Roster Update 2-23-09
On 2k Share.

Updated jersey numbers, minutes, fixed rotations (without injuries), FIXED 96 Bulls minutes and coach profile.

Updated Coach Profiles. Slightly lowered Defensive pressure to create some seperation. 1st attempt will make further adjustments as long as it doesnt bring the
3pt bug back, so im lowering little by little.

[DrUrsus]

Re: Neon1 v1.8 Rosters (360)....cracking the code on shot selection.

Neon1 Roster Update 2-20

On 2K Share.

- Trade Deadline moves
- Waived Mikki Moore, Sam Cassell, Re-added S.Marbury.
- Continued with the raised overall shot tendency scale
- Updated Rookies 76ers thru MIA
- Lowered scale for Zone Defense tendencies.
- 96 Bulls attached

Rules Settings -
Charge:65
Block:100
Reach-In:60
Shooting:75
Loose: 100

Roster Update Notes:
- 2009 All Star Uniforms unlocked
- Rebounding completely fixed all teams.
- 1996 Chicago Bulls with tendencies for Jordan,Pippen,Rodman,Harper, Longley, Phil Jackson coach Profile.


This roster has a few things untested/unfinished as i did not have the time to glance through things and check for errors. 2k Auto downloaded and froze my 360 up while i had autosave turned off for speed. I had to redo half of this thing so its a little rushed.

Ill clean up the jersey numbers, minutes, positions, outdated rookies and team tempos this weekend.

Neon and everyone,

A link to this file in xchange format is up in my blog.

DrUrsus

[Neon1]

Re: Neon1 v1.8 Rosters (360)....cracking the code on shot selection.

Neon and everyone,

A link to this file in xchange format is up in my blog.

DrUrsus

Well that file is old. Theres a newer one up today.

[dhowell8]

Re: Neon1 v1.8 Rosters (360)....cracking the code on shot selection.

Hey DrUrsus, do you have any updated rosters for 2k8 that I can down load with exchange 360. When I try rosters from 2k share my 360 freezes.
thanks D.Howell

[KG#5beantown]

Re: Neon1 v1.8 Rosters (360)....cracking the code on shot selection.

Hi Neon,

Mikki Moore to Boston (Moore will wear #7. source celtics.com)...marbury will be next on friday...

thanks for adjusting my man!

[Oraeon1224]

Re: Neon1 v1.8 Rosters (360)....cracking the code on shot selection.

Problems with points per game: I have run 10 simulation games CPU vs CPU and 4 games player (me) against CPU

I have posted this on a separate thread but wanted to put it here since Neon seems to comment more here. Points are not actually what should be reviewed (percentages and PIP change this) to see if clock management is right, but rather total attempts per game. I have stats through the midseason for all NBA teams and here are the average attempts per game:

Overall Total attempts per game (per team): ~100 (98.6 real)
Bulls: 100.3
Lakers: 102.1
Timberwolves: 99.9
Heat: 100.1

The gentleman who published his games has a correct score but attempts of:
Bulls: 130
Lakers: 99
Timberwolves: 115
Heat: 109

The main issues are from my test and the above posted screens is:
1) Players do not use enough clock (not correctible)

2) CPU frequently shoots too soon in the playclock when playing a player, but interestingly gets approximately the exact correct number of shots per game if playing against another CPU (usually ~100 in all 10 of my test). This is why CPU testing will not show this problem.

3) Even if you try to play a slower paced game as the player (and I do with the correct number of PIP) then the game still has too many attempts and thus to many points per game. This is almost always corrected by using 11 minutes (~100 points per game) but will have to be individualized by the players method of play (i.e drivers may need 10minute quarters).

4) Why were the points in these games not horribly desperate from real world numbers given the excessive number of shots and the good FG%? It is the lack of free throws, which add points but reduce attempts. The average number of free throws per game int eh NBA is 24.8 (the lakers average 27). All posted games have <10 (one exception). Again interestingly the CPU vs CPU games have the correct # of FT per game according to the sliders posted by Neon. This is easily correctibly by keeping the neon suggested foul slider set, which is quite good, and changing commit fouls to 75 for player and 65 for CPU. Unfortunately this will again depend on the playing style of the human and will have to be individualized by the player. What is critical is that the slider NOT be the same for the player and CPU since the player seems to always foul less than the CPU. (please also note that increasing fouls helps the attempts and points problem but also stops the clock for those points resulting in no significant time correction).

[Oraeon1224]

Re: Neon1 v1.8 Rosters (360)....cracking the code on shot selection.

Problems with PIP & FG%:

I have run 10 CPU vs CPU simulations and 4 player simulations (me). In real world statistics PIP is usually around 40 per game. As you can see from the screen captures posted (which are similar to what I witnessed), the CPU frequently scores >50 points in the paint and has a FG% of >50%. This breaks down into several areas of concern with both default rosters and neon rosters:

Problems:
1) Players are not at good at preventing cuts to the basket. Even with this said the CPU sim games still have far too many PIP.

2) The game designers and neon's rosters did not take into account real world FG% from different areas of the court as well as attempts per area per game. The real NBA statistics are published and the game statistics are visible from the post game hot spots screen:

Close range (paint): 0-5 feet (layups, dunks and under basket shots); FG% = 57%
Short 2's: 6-11 feet (area next to the paint); FG% = 37%
--Close plus short 2's ~60 attempts per game
Midrange 12 feet - 3pt line = 38%
--~20 attempts per game
3pt = 36.1%
--~20 attempts per game (18 to be exact)

What are the default neon rosters results in CPU vs CPU (and about the same in player vs player) with 100 attempts per game:
Close range (paint): FG% = 68-72%, 60-80 attempts per game
Short 2's: FG% = 42-50%, 10 attempts per game
Mid range: FG% = 20-33%, 9-12 attempts per game
3pt: FG% = 27-47%, 14-22 attempts per game (nearly spot on for individual teams)

What this means:
1) The inordinately high close attempts combined with the unbelievably high FG% success rate results in extremely high points in the paint each game and unrealistic FG% in both CPU vs CPU and player vs CPU games

2) 3pt statistics and attempts are pretty accurate (against player d this may change, I had to knock mine down 3 pts)

3) To have accurate rosters with correct points in the paint and FG% the following has to happen:

--Midrange shots tendency has to be increased to have about 20 per game (currently can be created by increasing midrange tendency on game sliders to 67, this corrects the midrange shots to correct # attempts and fixes excessive close shots attempts and seems have limited impact on 3pt attempts)

--Midrange FG% has to be increased to 38% (corrected currently by increasing midrange shooting on game sliders to 53) this will be in keeping with real world statistics and will balance out the decrease in close FG%.

--Close shot FG% must be decreased to reduce unrealistic FG% and PIP. This can be achieved primarily by reducing the success of layups. Currently this can be fixed by decreasing the close slider to 46 (still want those short 2's to have correct FG% as not PIP) and decrease layup success rate to 43 and dunk in traffic to 40.

--Player sliders should not use the above changes as play style (i.e. drives to the basket and bad shots) will change the required alterations. I use default sliders for the player with realistic stats.

Using the above slider change as well as 11min quarters as per the prior post I have played two games of magic vs calves that came down to teh last minutes with me winning one and loosing one with the following stats on my last game (similar for both) along with real world stats; I played the Magic:
Magic 102 Cavs 106
Shot attempts 101 (104) 109 (98.3)
FG% 46% (46%) 44% (48%)
3pt attempts 26 (26) 24 (21)
3pt% 41% (40%) 46% (37%)
FT attempts 24 21 (using my foul sliders)
PIP 42 38

My final question is could Midrange tendency and success vs close range success and layup success be adjusted for each player in the roster without wrecking simulation statistics?

[Oraeon1224]

Re: Neon1 v1.8 Rosters (360)....cracking the code on shot selection.

Problems with defense in CPU vs CPU simulation 10 times and CPU vs player 4 times:

In posted screen captures shows offensive rebounds of 12,10,4, 9 and steals of 12,5,12,8. I have found in my simulations that rebounds are roughly correct in total. However, if you correct the excessive shots per game (which necessarily reduces rebounds) to real world values of ~100 the number of rebounds is low largely from the FG% being to high (discussed above). If FG% is corrected then the correct number of rebounds per 100 rebounds is accurate, however, offensive rebounding is atrocious.

NBA offensive rebounds per game: 11.1
Screen shots rebounds per game: 12,10,4,9 and this is with well over 100 shots per game. Projected correction applied to screen rebounds: 10, 8, 3, 7
My simulation offensive rebounds per game: 5-11

If the number of shots per game is corrected then offensive rebounding is woefully low even with the correct FG% games that occured. I have increased my offensive rebounding slider to 65 and decreased defensive rebounding to 40 and this has helped but I am still ironing these out as FG% variation makes this difficult to project (I don't know if this is true but players such as D. Howard are 99 in defensive rebounding and offensive rebounding and I don't know if increasing offensive will help them get more Oreb or hurt them by improving lesser players, that is why I am toying with lower defensive rebounds currently which could keep dominant players dominant).

I don't know how this can be fixed in the roster without wrecking simulation rebounding statistics.

[oChaos_Nine]

Re: Neon1 v1.8 Rosters (360)....cracking the code on shot selection.

Can't get these rosters to work. I save them, then overwrite the original "Roster" file then delete the 2kshare rosters and I'm still locking on the team select screen. BTW How do you turn off the auto-download of the online rosters?

[DrUrsus]

Re: Neon1 v1.8 Rosters (360)....cracking the code on shot selection.

The latest Neon file is posted in Xchange format in my O.S. blog.

[laker8la]

Re: Neon1 v1.8 Rosters (360)....cracking the code on shot selection.

Problems with PIP & FG%:

I have run 10 CPU vs CPU simulations and 4 player simulations (me). In real world statistics PIP is usually around 40 per game. As you can see from the screen captures posted (which are similar to what I witnessed), the CPU frequently scores >50 points in the paint and has a FG% of >50%. This breaks down into several areas of concern with both default rosters and neon rosters:

Problems:
1) Players are not at good at preventing cuts to the basket. Even with this said the CPU sim games still have far too many PIP.

2) The game designers and neon's rosters did not take into account real world FG% from different areas of the court as well as attempts per area per game. The real NBA statistics are published and the game statistics are visible from the post game hot spots screen:

Close range (paint): 0-5 feet (layups, dunks and under basket shots); FG% = 57%
Short 2's: 6-11 feet (area next to the paint); FG% = 37%
--Close plus short 2's ~60 attempts per game
Midrange 12 feet - 3pt line = 38%
--~20 attempts per game
3pt = 36.1%
--~20 attempts per game (18 to be exact)

What are the default neon rosters results in CPU vs CPU (and about the same in player vs player) with 100 attempts per game:
Close range (paint): FG% = 68-72%, 60-80 attempts per game
Short 2's: FG% = 42-50%, 10 attempts per game
Mid range: FG% = 20-33%, 9-12 attempts per game
3pt: FG% = 27-47%, 14-22 attempts per game (nearly spot on for individual teams)

What this means:
1) The inordinately high close attempts combined with the unbelievably high FG% success rate results in extremely high points in the paint each game and unrealistic FG% in both CPU vs CPU and player vs CPU games

2) 3pt statistics and attempts are pretty accurate (against player d this may change, I had to knock mine down 3 pts)

3) To have accurate rosters with correct points in the paint and FG% the following has to happen:

--Midrange shots tendency has to be increased to have about 20 per game (currently can be created by increasing midrange tendency on game sliders to 67, this corrects the midrange shots to correct # attempts and fixes excessive close shots attempts and seems have limited impact on 3pt attempts)

--Midrange FG% has to be increased to 38% (corrected currently by increasing midrange shooting on game sliders to 53) this will be in keeping with real world statistics and will balance out the decrease in close FG%.

--Close shot FG% must be decreased to reduce unrealistic FG% and PIP. This can be achieved primarily by reducing the success of layups. Currently this can be fixed by decreasing the close slider to 46 (still want those short 2's to have correct FG% as not PIP) and decrease layup success rate to 43 and dunk in traffic to 40.

--Player sliders should not use the above changes as play style (i.e. drives to the basket and bad shots) will change the required alterations. I use default sliders for the player with realistic stats.

Using the above slider change as well as 11min quarters as per the prior post I have played two games of magic vs calves that came down to teh last minutes with me winning one and loosing one with the following stats on my last game (similar for both) along with real world stats; I played the Magic:
Magic 102 Cavs 106
Shot attempts 101 (104) 109 (98.3)
FG% 46% (46%) 44% (48%)
3pt attempts 26 (26) 24 (21)
3pt% 41% (40%) 46% (37%)
FT attempts 24 21 (using my foul sliders)
PIP 42 38

My final question is could Midrange tendency and success vs close range success and layup success be adjusted for each player in the roster without wrecking simulation statistics?

I would just like to say that this dude is on point. I have tested out and tried your slight altercations, and I must say that I am impressed. Why these tweaks are not getting more recognition is beyond me. Good job Oraeon, and thanks Neon for the best roster out there! My love for the game is reborn.

[ProfessaPackMan]

Re: Neon1 v1.8 Rosters (360)....cracking the code on shot selection.

So are these rosters the best ones to use or is there an updated version of these?

[cam21224]

Re: Neon1 v1.8 Rosters (360)....cracking the code on shot selection.

So are these rosters the best ones to use or is there an updated version of these?

Anyone who knows please post.

Is big baby Davis updated as well(his med range game has improved by a great margin).

Also, Neon did you use the ft and sig shot changes in your roster?

[DC]

Re: Neon1 v1.8 Rosters (360)....cracking the code on shot selection.

He is done with updates