Posted by arssanguinus on 9/13/2010 7:39:00 AM (view original):

Posted by emy1013 on 9/12/2010 5:03:00 PM (view original):

Posted by arssanguinus on 9/12/2010 9:47:00 AM (view original):
Are those stats really going to be entirely meaningful in any world that isn't completely populated with new recruits though?  Even if they might be interesting.

Doesn't matter if the world is populated with new all new recruits or not, the numbers need to be correct.  We had a situation in earlier worlds where there were pre-potential recruits and post-potential recruits mixed together for four seasons (I guess five if you include a few ineligibles).  That was a much different and much more complicated situation than we are looking at here, yet the numbers during those seasons with mix and match recruits were inline.

To imply that the numbers shouldn't matter for roughly four seasons while the old recruits cycle out shows one of two things (and I'm sorry to have to say this) (1) you're kind of out of touch with how this whole thing runs, or (2) you too often blindly defend WIS without thinking things through.  As I said, sorry for saying that, and I'm not trying to start an argument here, but again, to imply that the stats are "meaningless" because there are mixed and matched recruits (in essence saying that these four seasons could be screwed up, but we should deal with it because of the changes, is just unacceptable).  I don't pay my money to just play "meaningless" seasons, they need to be right, regardless of what kind of "recruits" are being used.

Frankly, EMy, thats impossible, unless you want them to field a new engine each for their Freshman, Sophomore , Junior and senior class until they haver all graduated and been replaced. . . 

]And Emy, you can kinda stick it with the :"Blindly defending WIS' excuse you use _every_ time.  Its a valid point, whether you like it or not, that the new recruits and the old recruits sharing space together are going to make the stats be different.  I wasn't rude, I wasn't condescending, I just made a simple point that the mix of old and new is going to effect the stats and you reacted like a jerk.   You can disagree if you like , but that won't make it not true, regardless of your attacking my motives for saying it..

And you didn't say it would "effect the stats", you said it would make them "meaningless".  If you don't know the difference between those two terms, you should probably stick to slightly smaller, less complicated words.

He's also agreeing with me.  Nothing he said in his post disagreed with what I said in mine.  And if you want to accuse me of not understanding words, "Will they be entirely meaningful" isn't even close to the same as saying "Will be entirely meaningless".  In fact, the word meaningless is NOWHERE in the post.  SO at least have read the post before going off on a tangent, which you evidently did not.  Saying I made the wrong use of a term I didn't even make use of. . .

IN fact, I actually agree with the whole of both of BillyG's posts.

 

 

 

 

Posted by coach_billyg on 9/13/2010 12:49:00 AM (view original):

Posted by isack24 on 9/12/2010 5:15:00 PM (view original):

Posted by mullycj on 9/12/2010 8:48:00 AM (view original):
There used to be a pinned thread from Tarek that showed shooting %, 3pt%, and FT% for the three divisions vs NCAA.

I think the community would be interested in seeing those results for the current season accross worlds.  The general consesus is the current engine is out of whack and teams are shooting too high.     You obviously have a different view.

Please share with us the stats you see to support your view. 

I'm not sure that's a general consensus, nor do I think it needs to be completely in line with the real NCAA numbers.

i do think it is a general consensus. its important the sim engine behaves closely to real basketball so that real basketball strategy applies, etc... if there is a slight deviation from reality in the stats, i think that is acceptable, but it should definitely try to be minimized. there are also a number of factors that could push HD stats out of whack with reality - like, if people play - defenses more than + defenses, 3pta and 3pt% will reflect higher than the NCAA. but in general, we need the sim engine to mimick reality. in the extreme case, you could make it totally different, but i really struggle to believe anybody would play it then.

It's a general consensus that the stats are way out of whack?

I suppose it could be, but I haven't gotten that feeling generally, just from a select few vocal forum leaders.

Anyway, strictly mimicking real life only works if the the game itself strictly mimics real life, which it doesn't.  The strategy can still be real-life relevant when two evenly matched, man-to-man/motion human teams play each other, but when you have teams running uptempo fastbreak/press, why should any type of real-life results be expected, when that never happens in real ilfe?

Teams are going to shoot higher overall percentages in HD because there are more mismatches, and more unrealistic strategies being run.  If the stats in those games followed closely with real life, it would in fact make the stats in other games less realistic, because you'd be adjusting them to fit in with unrealistic strategies.

Posted by ike1024 on 9/13/2010 10:33:00 AM (view original):

Posted by coach_billyg on 9/13/2010 12:49:00 AM (view original):

Posted by isack24 on 9/12/2010 5:15:00 PM (view original):

Posted by mullycj on 9/12/2010 8:48:00 AM (view original):
There used to be a pinned thread from Tarek that showed shooting %, 3pt%, and FT% for the three divisions vs NCAA.

I think the community would be interested in seeing those results for the current season accross worlds.  The general consesus is the current engine is out of whack and teams are shooting too high.     You obviously have a different view.

Please share with us the stats you see to support your view. 

I'm not sure that's a general consensus, nor do I think it needs to be completely in line with the real NCAA numbers.

i do think it is a general consensus. its important the sim engine behaves closely to real basketball so that real basketball strategy applies, etc... if there is a slight deviation from reality in the stats, i think that is acceptable, but it should definitely try to be minimized. there are also a number of factors that could push HD stats out of whack with reality - like, if people play - defenses more than + defenses, 3pta and 3pt% will reflect higher than the NCAA. but in general, we need the sim engine to mimick reality. in the extreme case, you could make it totally different, but i really struggle to believe anybody would play it then.

It's a general consensus that the stats are way out of whack?

I suppose it could be, but I haven't gotten that feeling generally, just from a select few vocal forum leaders.

Anyway, strictly mimicking real life only works if the the game itself strictly mimics real life, which it doesn't.  The strategy can still be real-life relevant when two evenly matched, man-to-man/motion human teams play each other, but when you have teams running uptempo fastbreak/press, why should any type of real-life results be expected, when that never happens in real ilfe?

Teams are going to shoot higher overall percentages in HD because there are more mismatches, and more unrealistic strategies being run.  If the stats in those games followed closely with real life, it would in fact make the stats in other games less realistic, because you'd be adjusting them to fit in with unrealistic strategies.

sorry if that was unclear, i was saying its a general consensus that there should be strong correlation between the NCAA stats and HD stats.

i guess i did say that wrong. are the engine stats out of whack with reality? probably, but i wasn't really trying to comment on that at the time, at least not without looking at it a little deeper. but now that i mis-spoke and already commented on it, i guess i may as well. i think there is a pretty significant difference today in fg% and 3pt% from what we are used to, which supposedly was in line with NCAA stats. but maybe its just a different curve now, and the better teams will see better numbers and the worse teams will see worse numbers. i don't remember if tarek was correlating anything but averages, i don't seem to remember he was, but to do it properly, i think you would. not that i expect it or anything...

Posted by dcy0827 on 9/13/2010 8:56:00 AM (view original):

Posted by arssanguinus on 9/13/2010 7:39:00 AM (view original):

Posted by emy1013 on 9/12/2010 5:03:00 PM (view original):

Posted by arssanguinus on 9/12/2010 9:47:00 AM (view original):
Are those stats really going to be entirely meaningful in any world that isn't completely populated with new recruits though?  Even if they might be interesting.

Doesn't matter if the world is populated with new all new recruits or not, the numbers need to be correct.  We had a situation in earlier worlds where there were pre-potential recruits and post-potential recruits mixed together for four seasons (I guess five if you include a few ineligibles).  That was a much different and much more complicated situation than we are looking at here, yet the numbers during those seasons with mix and match recruits were inline.

To imply that the numbers shouldn't matter for roughly four seasons while the old recruits cycle out shows one of two things (and I'm sorry to have to say this) (1) you're kind of out of touch with how this whole thing runs, or (2) you too often blindly defend WIS without thinking things through.  As I said, sorry for saying that, and I'm not trying to start an argument here, but again, to imply that the stats are "meaningless" because there are mixed and matched recruits (in essence saying that these four seasons could be screwed up, but we should deal with it because of the changes, is just unacceptable).  I don't pay my money to just play "meaningless" seasons, they need to be right, regardless of what kind of "recruits" are being used.

Frankly, EMy, thats impossible, unless you want them to field a new engine each for their Freshman, Sophomore , Junior and senior class until they haver all graduated and been replaced. . . 

]And Emy, you can kinda stick it with the :"Blindly defending WIS' excuse you use _every_ time.  Its a valid point, whether you like it or not, that the new recruits and the old recruits sharing space together are going to make the stats be different.  I wasn't rude, I wasn't condescending, I just made a simple point that the mix of old and new is going to effect the stats and you reacted like a jerk.   You can disagree if you like , but that won't make it not true, regardless of your attacking my motives for saying it..

And you didn't say it would "effect the stats", you said it would make them "meaningless".  If you don't know the difference between those two terms, you should probably stick to slightly smaller, less complicated words.

emy, FWIW, i don't see where he called them meaningless. at least not in what you quoted. he asked if they would be entirely meaningful. and the answer, frankly, is no. they are certainly meaningful to a degree, but even with a completely calibrated engine before and after the changes, the engine in transition is going to be slightly uncalibrated during the changes. particularly 2 and 3 seasons into them. i personally don't see that as being a serious issue as long as the impact is fairly small.

now, whether or not the engine today is close enough to reality is a completely different issue. but i honestly didn't take it that aintheb was commenting on that there.

But again, billy g, why should there be a strong correlation to the real world when a moderately high percentage of HD teams run offenses, defenses, and overall strategies that real life teams never do?

I agree that games involving real life defenses should, in the long run, produce real life results.  But overall percentages shouldn't necessarily correlate because the teams and gameplans aren't realistic.  Forcing them to produce realistic results, then, seems illogical.  It's why, when uptempo/press was producing ludicrous results after the new engine's release, I suggested relegating those to combo status, and removing them as primary options.  You simply can't advocate for realism when realism isn't a possibility.

Posted by isack24 on 9/13/2010 11:23:00 AM (view original):
But again, billy g, why should there be a strong correlation to the real world when a moderately high percentage of HD teams run offenses, defenses, and overall strategies that real life teams never do?

I agree that games involving real life defenses should, in the long run, produce real life results.  But overall percentages shouldn't necessarily correlate because the teams and gameplans aren't realistic.  Forcing them to produce realistic results, then, seems illogical.  It's why, when uptempo/press was producing ludicrous results after the new engine's release, I suggested relegating those to combo status, and removing them as primary options.  You simply can't advocate for realism when realism isn't a possibility.

i see what you are saying, but real life still needs to be the standard we are shooting for whenever possible. so, for example, seble could take all the motion/triangle/flex vs man/zone games and come up with the stats for that. then, he could also state the average +/- and tempo setting, which would explain some discrepancy from real life depending on the outcomes and the average settings. we could even restrict game to those with +/- of 1 or 0, normal tempo both ways. you then need to examine the other cases not included to make sure they fit in with the core model - but i don't think you make those numbers like up (like fcp, or trying to make -5 correlate to anything except the rest of HD).

another good way to look at things is to take the games between top 50 rpi teams only. these teams are usually reasonably well setup and most abnormalities are balanced out somewhere else. i guess the question is, is that data available for the NCAA? but my guess would be yes, i mean they seem to have every stat imagineable and then some, so i would think they would have metrics like that. 

another reason things might not line up is because we have d1, d2, d3, all running on the same engine. i don't expect all 3 to align perfectly with real life, i think if d1 lines up and d2 and d3 are reasonably close, that is probably good enough.

again, i am not advocating an exact match between RL numbers and HD numbers. but, RL absolutely needs to be the model the sim engine is trying to achieve. one way to measure how successful the sim engine models real life is to compare the stats. of course, they won't line up in all cases, but if there are some core cases where they are close, I would be fine with that. also if there is a significant discrepancy, and it comes with a reasonable explanation, that is fine too. the community could probably provide those explanations in many cases too. i don't think seble needs to run these numbers and then go change everything to make them match. but it is an extremely important exercise to go through nonetheless, and if there are major issues, i do think it is important to figure out why and then depending on your findings, make adjustments. it is exactly what i would do if i made my own HD, i really couldn't imagine not doing it actually. i would bet that seble has looked at this to some extent and was reasonably comfortable with the results, even though he hasn't published them. i mean it would literally be insane to rewrite the engine from scratch and NOT re-sim it against the old a million times to make sure the results were consistent.

Posted by mullycj on 9/12/2010 8:48:00 AM (view original):
There used to be a pinned thread from Tarek that showed shooting %, 3pt%, and FT% for the three divisions vs NCAA.

I think the community would be interested in seeing those results for the current season accross worlds.  The general consesus is the current engine is out of whack and teams are shooting too high.     You obviously have a different view.

Please share with us the stats you see to support your view. 

every time i have tried to argue this, other coaches tell me it's because too many coaches play a minus defense and that skews the WIS stats vs real life stats, although my counter argument would be "who's to say that real life coaches don't play minus defense more often either."  since there's no way to know exactly for sure, it's best to at least get the engine to get the WIS shooting %s near RL %s.

i would tend to agree with you that shooting %s still seem too high to me.

Still, so far, my position lines up perfectly with Billyg.

i honestly think this is a case of things sounding differently than intended. seems to happen all the time, whether its a text, instant message, or forum post. i really don't think you guys have much difference in opinion, as best i can tell. its just like, if you've ever had a fight with your girlfriend over text message, you know just how severely  things can be lost in translation. and if you've never had a fight with your girlfriend over text message, well, for the love of god, don't  =)

Okay, in order to keep this thing on topic and not turn this into a name calling, mud slinging, who can top whom, insult contest, I'll step up and apologize.  As you can see by the times that I posted, it was early in the morning and I hadn't yet been to sleep.  Not an excuse, but it's the best I've got.  Arss, I'm sorry that I had to stoop to using insulting words towards you in my post.  Apology accepted?

If you want to try to discuss this in a civil manner, I'm all for it, seems like it's become a bit of a lively topic.  So, without trying to come across as being a wiseass, I just think that regardless of whether all the new recruits have been cycled into the system or not, the stats that HD produces should be as "meaningful" (and yes, I misread that earlier, you got me there)  :^) as is possible.  Woud you not agree?  And if not, why not?

And by the way, I haven't yet read much of page 2's posts, so forgive me, gentlemen, if my question has already been addressed.......

Posted by summerteeth on 9/13/2010 1:33:00 PM (view original):

Posted by mullycj on 9/12/2010 8:48:00 AM (view original):
There used to be a pinned thread from Tarek that showed shooting %, 3pt%, and FT% for the three divisions vs NCAA.

I think the community would be interested in seeing those results for the current season accross worlds.  The general consesus is the current engine is out of whack and teams are shooting too high.     You obviously have a different view.

Please share with us the stats you see to support your view. 

every time i have tried to argue this, other coaches tell me it's because too many coaches play a minus defense and that skews the WIS stats vs real life stats, although my counter argument would be "who's to say that real life coaches don't play minus defense more often either."  since there's no way to know exactly for sure, it's best to at least get the engine to get the WIS shooting %s near RL %s.

i would tend to agree with you that shooting %s still seem too high to me.

I can counter that counter argument.  The definition of a 0 defense is that it is average.  That average would be roughly equivalent to the average of all real-life teams.  So if every team played at a 0 defense and we played a fairly realistic cross-section of offenses and defenses as compared to real-life and THEN the shooting percentages were still off, that would imply that they were in fact skewed.  But if in fact sim players trend towards - defenses - which in my limited experience does seem to be the case, at least in D3 - then we should expect to have somewhat elevated 3% and depressed 2% as compared to RL.

Posted by mullycj on 9/12/2010 8:48:00 AM (view original):
There used to be a pinned thread from Tarek that showed shooting %, 3pt%, and FT% for the three divisions vs NCAA.

I think the community would be interested in seeing those results for the current season accross worlds.  The general consesus is the current engine is out of whack and teams are shooting too high.     You obviously have a different view.

Please share with us the stats you see to support your view. 

please don't include me in your "general" consensus.