A deeper look into simulated statistics in Association Mode....

[Mo_Magic]

Re: A deeper look into simulated statistics in Association Mode....

so, anyone else seeing this ? And what efect does it have on individuas as well as other team stats ? sepcifically : Are FT numbers too low ? (causing high FGs) What are the PPG numbers for teams ? Are there enough 3s ? (see FTs) What about shooting percentages ?

<!-- / message -->

Tnaks, brosef.

Well, the general effect is: one, it leads to WAY too many possessions and rebounds, assist, steals, blocks, points, etc. Everything is inflated to the nth degree. Generally, you should expect around 28-30 defensive rebounds in a game as an average from each team along with 8-12 offensive rebounds. The way the game is set up now, you get about 45+ defensive rebounds. This is the reason LeBron, Carmelo, Chris Bosh, etc, consistently average 10+ rebounds in my associations.

It's also why LeBron is literally averaging a triple double through the sim I'm going through now.

Basically, if your team shots are off, and by that much, it's going to affect the game... GREATLY and in a bad way.

Unfortunately, 2K's tendencies are STILL mucked up. If I were to get a realistic pace to the games, this would be an EASY fix. All it would take is literally adjusting for the right pace for every team, then making the coaching profile shooting tendencies average out to a certain set percentage point(so, if for example team A shoots 35% in real life, I'd review how fast into the shot clock they shoot and adjust all the coach profile shooting sliders to be higher, while still making sure the 3 takes up as close to 35% as possible). The result would be a completely realistic number of shots attempted from beyond 3.

Unfortunately, I can't get this. And as such, if somehow I were able to lower the shot attempts, I'd still need to do some HEAVY adjusting to get correct free throws, 3-pt shots, etc... In short, the ONLY reason that those stats might come out looking semi-correct, is because 2K lowballed those coach profile tendencies a TON and greatly overshot the general shots attempted statistic BY A MILE with their sim engine.

[nogster]

Re: A deeper look into simulated statistics in Association Mode....

lowering the game tempo coach profiles would make some difference wouldnt it mo?
i take it youve tried this but to no avail.

how low do you have to drop teams game tempo?
and does it hinder the actual gameplay?

i cant think of any other thing that will help.


man i was hoping i wouldnt have to do much if at all this season with 82games on board. damn frustrating.

and the competitions sim engine is even worse i hear.

what are the developers doing?
dont they test this kindve stuff. obviously not.
why is it always up to us to get it half right.

[Mo_Magic]

Re: A deeper look into simulated statistics in Association Mode....

lowering the game tempo coach profiles would make some difference wouldnt it mo?
i take it youve tried this but to no avail.

how low do you have to drop teams game tempo?
and does it hinder the actual gameplay?

i cant think of any other thing that will help.


man i was hoping i wouldnt have to do much if at all this season with 82games on board. damn frustrating.

and the competitions sim engine is even worse i hear.

what are the developers doing?
dont they test this kindve stuff. obviously not.
why is it always up to us to get it half right.

I've lowered every single team inside of ten and still had some teams shoot for over 7500. I can drop half the league inside of 6000, but there's another HALF that shoots over 7000 shots.

Because of this, it's IMPOSSIBLE to tell whether we got shafted on the ratings again... but it looks like we did in all likelihood and the game will need a massive rerating as far as shooting percentage ratings, rebounding ratings, passing ratings, etc... but this can only be done with any sort of accuracy when the game tempo problem is fixed.

Somebody PM Simballer or I will, because if you're like me and Nog here, this is a HUGE oversight.

From what I've seen this game plays right, so you may not have to do much on a tendency front. Games just feel right. This problem just needs to be fixed... and after they provide us the means, we just have to get it looking right from a statistics standpoint. By no means THAT big of a problem IF they fix it.

Wow... Live's is worse? How is that even possible? xD

[Whoman]

Re: A deeper look into simulated statistics in Association Mode....

i think you are confusing cause and effect. Shots taken have to be caused by sth else and not nescessarily game tempo. So we´d have to identify those to do the via the rosters or simply to give specific Feedback to VC. B
asically there´s (out of the top of my head) a couple possible reasons :

1) Too many possessions = Too many PPG = Too fast a game from the screens i saw, this isn´t the case, but could you check ? That´d be the easy one.
2) Too many shots and a bad FG% = normal PPG on way too many shots
3) Not enough FTs. Less FTs = More FGs when taking the same amount of possessions.
4) Not enough 3s and too many 2 point FGs = seeimingly normal PPG on way more shots (more 2s and less 3s = more Points per Shot attempt / more Shots for every Point)

or of course a mixture of those and others.

So one would have to check those stats first and foremost as to find out what is causing the discrepancy.

Still don´t have the game, release is Friday over here in Europe with an outside chance of thurday evening at my local store. That´s why i can´t check myself.

[Jukeman]

Re: A deeper look into simulated statistics in Association Mode....

I dont know, I wouldnt go and do a massive rereate as of yet....Kind of too early....

But I simmed 20 years and rookies who are now 16 year vets have a good looking ppg average.....As I mention before, Guys like LBJ(99shot tendency) is too high....The league actually got better after the default roster retire...This tells me "less is more" as I drafted a guy who was a 81OvR his rookie year and now sits at 89 in his 12th year.....If we lowered the rating/Shot tendency(only) across the board, the sim stats may be better, I dont think its the team profiles....Do we really need rating/tendencies in the 90's anyway?

Also Lou Williams averaging 25 ppg(90 shot tendency) since 2009-2010 but stayed between 69-71 in OVR.....

Oh yea and Bosh getting more than 10 rebounds a game is mainly do to the fact he is the best rebounder on the team by a large margin, infact his rebounding rating is overrated...

[Mo_Magic]

Re: A deeper look into simulated statistics in Association Mode....

i think you are confusing cause and effect. Shots taken have to be caused by sth else and not nescessarily game tempo. So we´d have to identify those to do the via the rosters or simply to give specific Feedback to VC. B
asically there´s (out of the top of my head) a couple possible reasons :

1) Too many possessions = Too many PPG = Too fast a game from the screens i saw, this isn´t the case, but could you check ? That´d be the easy one.
2) Too many shots and a bad FG% = normal PPG on way too many shots
3) Not enough FTs. Less FTs = More FGs when taking the same amount of possessions.
4) Not enough 3s and too many 2 point FGs = seeimingly normal PPG on way more shots (more 2s and less 3s = more Points per Shot attempt / more Shots for every Point)

or of course a mixture of those and others.

So one would have to check those stats first and foremost as to find out what is causing the discrepancy.

Still don´t have the game, release is Friday over here in Europe with an outside chance of thurday evening at my local store. That´s why i can´t check myself.

I've checked EVERY stat. Teams just run offense too fast when games are simmed. The turnovers are low for the incredulous amount of shots, as are the the free throws, but that's hardly the real issue here. I can see if plugging in all the right numbers for players works, but I severely doubt that that is the case and we are just dealing with a broken simulation engine.

Basically, from what I've seen, with the shot attempts jacked so high, you actually get the CORRECT number of 3PT attempts and fouls and the turnovers are not at all proportionate to the possessions, so you may be onto something... but keep in mind that we are dealing with over 50,000+ some odd extra shots when the game is on default coach profiles. I HIGHLY doubt there is anything we can do to remedy this problem.

Field goal percentages look okay for the most part, maybe a pinch high because 3PT percentages are higher than they should be.

[DC]

Re: A deeper look into simulated statistics in Association Mode....

Ok I did some testing. I put the following players at 48 mins and gave them all 99's in steals, blocks, and assists to see how it sims out

STL
Wade 3.7 - 41 mins
Paul 3.3 - 41 Mins
Rondo 3.2 - 40 mins

Block
Dwight 3.1 - 36 mins
Elton - 2.7 - 32 mins
Amare - 3.1 - 31 mins
Duncan - 3.3 - 37 mins

As-S
Nash - 9.8 - 36mins
CP3 - 9 - 40 ins
Deron - 11.5 40 mins


In case anyone was wondering. I did this to see some extremes.

[Whoman]

Re: A deeper look into simulated statistics in Association Mode....

Actually the roster tendencies do make a difference if the sim engine is anything like 2K9 (and i suppose it is). The engine does do player stats first from my tests with 2K9, not team stats that then get "divided" and broken down to player stats.

And the PPG ? High amount of shots, okay FTs, high 3Pt% and okay FG% should mean way too many PPG.

Can you just take the team stats sorted in those categories :

3PT attempts
3Pt Makes
FT attempts
FT makes
FG%

Ok I did some testing. I put the following players at 48 mins and gave them all 99's in steals, blocks, and assists to see how it sims out

STL
Wade 3.7 - 41 mins
Paul 3.3 - 41 Mins
Rondo 3.2 - 40 mins

Block
Dwight 3.1 - 36 mins
Elton - 2.7 - 32 mins
Amare - 3.1 - 31 mins
Duncan - 3.3 - 37 mins

As-S
Nash - 9.8 - 36mins
CP3 - 9 - 40 ins
Deron - 11.5 40 mins


In case anyone was wondering. I did this to see some extremes.

For Paul and the assist givers in general the problem might be their high PPG and in essence his high shot tendency or rather his teammates low shot tendency (just guessing from similar cases in 2K9). Try raising a couple of New Orleans starters in shot tendency and lower Paul by a bit maybe, normally you should see him closer to his real world averages for Points and Assists as he then isn´t the clear cut best scorer by as much.
Again, just guessing here.

I guess minutes are simply limited by stamina and/or foul problems.

[nogster]

Re: A deeper look into simulated statistics in Association Mode....

Ok I did some testing. I put the following players at 48 mins and gave them all 99's in steals, blocks, and assists to see how it sims out

STL
Wade 3.7 - 41 mins
Paul 3.3 - 41 Mins
Rondo 3.2 - 40 mins

Block
Dwight 3.1 - 36 mins
Elton - 2.7 - 32 mins
Amare - 3.1 - 31 mins
Duncan - 3.3 - 37 mins

As-S
Nash - 9.8 - 36mins
CP3 - 9 - 40 ins
Deron - 11.5 40 mins


In case anyone was wondering. I did this to see some extremes.

nice DC. keep the testing coming.
from these stats. i will take some educated guesses and say that just like last season the "commit fouls" tendency is too high for a lot of players.
last season as im sure you know. you had to lower the commit fouls tendencies scale in half to get the players to actually play near the minutes you allocate them in the rotations.

some of pauls team mates will need a passing attribute downgrade. [im assuming paul is 99 for passing yeah?]
same with nash to get their apg to the level they should be. [paul in particular]

i see the same issue i had with amare's blocking stats as last year [once shaq left phoenix] and the simple reason for that is there is no shotblocker on the suns. so amare gets the lions share of the sim engines blocks.
brand getting that high in 32 mins when dalemburt and to a lesser extent speights is on the team indicates to me that brands blocking attribute is very high. and dalemburts mpg are fairly low. [his commit fouls tendency must be high or something.]

it looks like all the same simming issues we had last year are back this year. sigh. i thought the 82games.com partnership wouldve had this sorted. maybe only for gameplay.
we are going to have to find a balance between sim stats and gameplay yet again.

[Whoman]

Re: A deeper look into simulated statistics in Association Mode....

yeah, passing attributes play a role as well obviously, but i´d really like to take a shot at changing shot tendency. Propably not worth the time with the default rosters though ...
Ah well, i´ll see what the next couple days/weeks bring in terms of anouncements by VC and then dive into it.
If you need help @ nogster feel free to ask.

I´ll definitely test once i get it.

[Mo_Magic]

Re: A deeper look into simulated statistics in Association Mode....

Actually the roster tendencies do make a difference if the sim engine is anything like 2K9 (and i suppose it is). The engine does do player stats first from my tests with 2K9, not team stats that then get "divided" and broken down to player stats.

And the PPG ? High amount of shots, okay FTs, high 3Pt% and okay FG% should mean way too many PPG.

Can you just take the team stats sorted in those categories :

3PT attempts
3Pt Makes
FT attempts
FT makes
FG%

Yeah, that shouldn't be too hard.

As an aside, Pared's shooting sliders are awesome, but the only reason they really need to be jacked is that most of these players have over rated shooting attributes in comparison to what the sim engine and actual game puts out. I've tested and tinkered and a 40% shooter from beyond the arc falls at around 83 as opposed to 85 last year.

Basically, EVERYONE is over rated in regards to their 3PT rating and what the sim engine will get out of that rating. For example, LeBron James is a 34% 3PT shooter. To achieve the desired effect of his mediocre shooting from beyond 3, you will need to put him at a 73 as opposed to what I used last year to get accurate statistics(76).

It's only three points, but it really does make a WORLD of difference. Especially in the mid range game.

If the mid range ratings were done right, we wouldn't even have to touch them.

[Jukeman]

Re: A deeper look into simulated statistics in Association Mode....

Yeah, that shouldn't be too hard.

As an aside, Pared's shooting sliders are awesome, but the only reason they really need to be jacked is that most of these players have over rated shooting attributes in comparison to what the sim engine and actual game puts out. I've tested and tinkered and a 40% shooter from beyond the arc falls at around 83 as opposed to 85 last year.

Basically, EVERYONE is over rated in regards to their 3PT rating and what the sim engine will get out of that rating. For example, LeBron James is a 34% 3PT shooter. To achieve the desired effect of his mediocre shooting from beyond 3, you will need to put him at a 73 as opposed to what I used last year to get accurate statistics(76).

It's only three points, but it really does make a WORLD of difference. Especially in the mid range game.

If the mid range ratings were done right, we wouldn't even have to touch them.

I'll go to the extreme a say ALL of the ratings are a bit overrated...something Ive been saying for weeks....No way Bosh should even be sniffing a 80 in either rebounding rating when only averging under 10 a game for his career....for a 6'10 big...

[Whoman]

Re: A deeper look into simulated statistics in Association Mode....

I think that is the fallout from the wider gap between ratings, stars "take up" more of the stats as the difference between them and their teammates is greater this year.
If a player is 40 points higher in shot tendency than the average of his teammates he´ll have more PPG than when he has a 30 point higher one.

the scale for the shot tendency is too wide as it is and seems to be gotten worse, was the same in 2K9. Upon adjusting i had the PPG leaders and also the team balance down pretty good last year.

There were outliers of course due to team structure, but that´s good as it´s realistic in a way. F.e. did Durant jump from 29 PPG on a stacked team to 37 PPG on a team loosing the 2nd best scorer to FA and the 3rd best being injured for a long time. (and boy, Durant was a load a couple seasons in, my nemessis in 2K9, my whole lineup was designed to stop him. Like teams would get big bodies to stop Shaq, i got lanky guards/forwards and agile big guys to stop him)

Same for rebounding or assist ratings a i guess, although it seems that assists are affected negatively by other factors and are too low for the best passers instead of way too high like it seems with scoring and certain rebounders.

Although assist imo have a lot to do with the Shot Tendency as well, maybe guys like Paul or Parker (to an extent) have to high a shot tendency.

How this would screw up gameplay is a question, but imo shot tendency might be the key to a lot of stat related things.

[eeastfan]

Re: A deeper look into simulated statistics in Association Mode....

I don't think these poor generated rookie's stats are due to their tendencies, but to the combination of both low shooting tendencies and the minutes they're getting.
From what I recall, in the previous two games, most/all rookies (even Top picks) had between 0 and 5 minutes of playing time.

[CWSapp757]

Re: A deeper look into simulated statistics in Association Mode....

I feel you eeastfan. So many of these highly rated rookies receive less than 10 minutes, their morale drops to 0 and their overall drops to the 60s or 50s. I've just accepted the fact that we are going to have to do a lot of tweaking to incoming rookie ratings. This is why 2k share is so important to have this year.