ecruit generation +1

Weird. I don't think I've ever seen OR troll a thread before.

Not a bad effort, but in the future you should probably try to be a little more strident to really get a strong reaction out of folks.

Seble, I really do want an answer to OR's question.


And it really scares me how it's good coaches - the ones that this change is supposed to help - that are uneasy about it.

OR - I think you are looking at this the wrong way.   About about this view.

The game is built so that their is an inherent standard deviation from normal results built in.  (ie - the results are sometimes forced to stray from normal to reflect reality).

I think all seble is doing is adjusting this deviation range if the results in a game begin to stray too far away from "normal".

Again I ask and I dont think it has been answered. Was game testing done? Or is this a seat of the pants attempt at changing the game?

Whether this change ends up good or bad, the simple fact that they didn't utilize the beta world with a bunch of coaches to test it out is pretty mind-boggling.

Posted by mullycj on 8/29/2010 9:17:00 AM (view original):
OR - I think you are looking at this the wrong way.   About about this view.

The game is built so that their is an inherent standard deviation from normal results built in.  (ie - the results are sometimes forced to stray from normal to reflect reality).

I think all seble is doing is adjusting this deviation range if the results in a game begin to stray too far away from "normal".

mully / others - wrong way? - only comment I have made is I doubt it will amount to much of anything, but mathematically, it is a little odd to 'fudge' random number theory and I asked a question about comparing HD deviations to real life - given the nature of the change - how can any of that be construed as negative or criticism?

Are we now so afraid of the truth (after the recruit generation issue) that no coach is even allowed to ask a ? without getting criticized? - I mean read anton's little 4am attack dog beauty - what is that all about - no wonder noone posts anymore or cares?

Mully - your analysis of what seble 'is doing' may be right, but I will tell you from a strictly math sense - it is a little odd - if you think of a flow chart, it sort of says if I don't like the results so far I'll cheat moving forward so they turn out the way I want them to?

I will add real life I think this change is good to simulate - real life ref'ing tends to even out blow out games and odd things, and things like hot and cold shooting does tend to even out.  One of the last competitive games I played, a title game in an over 35 tourny, i made 6 straight 3's in the second qtr, I was 1 for 5 the rest of the game, and to boot, was 1 for 6 from the foul line in the 4th qtr (I was 48, fatigue may have set in)

See - there are rational arguements to be made without resorting to name calling, I just made two for the seble fanboy crowd.

From a sample night of one, maybe 5 OR-involved games, I only had one odd thing happen, my all world sr sf missed 4 straight 3 pointers in the last 30-60 seconds to lose a game by 2 - if the change had gone overboard, he probably would have made one of those - too bad really - LOL - there is a 3rd argument supporting the fudge factor engine change - there you go - criticize me for that - HA

The thing is, if anybody was designing a basketball sim from scratch, you'd make each shot attempt independent of the others.  To their credit, this is how HD worked before.  However, OTHER flaws in the engine led to greater variance than you'd see in real life.  Using shooting % as an example, presumably the flaw was that each shot attempt had greater variance from a guy's true skill, leading to inordinant 1-10 or 9-10 games.  Now, instead of fixing the variance in each individual shot attempt, Seble has turned something HD gets RIGHT (independent shot attempts) into something HD gets WRONG (an 0-6 shooter will automatically be hotter).  Perhaps this awkward fix will actually improve the bottom line on games.  But at best it will do so by having early stats skewed one way, balanced out by later stats jerry-rigged to skew the opposite way, NOT by actually improving the simulation on a possession-by-possession basis. 

Posted by mamxet on 8/28/2010 2:12:00 PM (view original):
what does "reduced single game variance" really mean?  does it mean that possessions are no longer independent events and that - for example - if a guy who should hit 80% of his FTs hits 8 in a row that the 9th will not have an 80% chance?

if it doesnt mean that, what could, what does it mean?

Seems to me like it means the game was too random....

Posted by jcfreder on 8/29/2010 10:37:00 AM (view original):
The thing is, if anybody was designing a basketball sim from scratch, you'd make each shot attempt independent of the others.  To their credit, this is how HD worked before.  However, OTHER flaws in the engine led to greater variance than you'd see in real life.  Using shooting % as an example, presumably the flaw was that each shot attempt had greater variance from a guy's true skill, leading to inordinant 1-10 or 9-10 games.  Now, instead of fixing the variance in each individual shot attempt, Seble has turned something HD gets RIGHT (independent shot attempts) into something HD gets WRONG (an 0-6 shooter will automatically be hotter).  Perhaps this awkward fix will actually improve the bottom line on games.  But at best it will do so by having early stats skewed one way, balanced out by later stats jerry-rigged to skew the opposite way, NOT by actually improving the simulation on a possession-by-possession basis. 

Best post on the topic so far.

Posted by oldresorter on 8/29/2010 10:25:00 AM (view original):

Posted by mullycj on 8/29/2010 9:17:00 AM (view original):
OR - I think you are looking at this the wrong way.   About about this view.

The game is built so that their is an inherent standard deviation from normal results built in.  (ie - the results are sometimes forced to stray from normal to reflect reality).

I think all seble is doing is adjusting this deviation range if the results in a game begin to stray too far away from "normal".

mully / others - wrong way? - only comment I have made is I doubt it will amount to much of anything, but mathematically, it is a little odd to 'fudge' random number theory and I asked a question about comparing HD deviations to real life - given the nature of the change - how can any of that be construed as negative or criticism?

Are we now so afraid of the truth (after the recruit generation issue) that no coach is even allowed to ask a ? without getting criticized? - I mean read anton's little 4am attack dog beauty - what is that all about - no wonder noone posts anymore or cares?

Mully - your analysis of what seble 'is doing' may be right, but I will tell you from a strictly math sense - it is a little odd - if you think of a flow chart, it sort of says if I don't like the results so far I'll cheat moving forward so they turn out the way I want them to?

I will add real life I think this change is good to simulate - real life ref'ing tends to even out blow out games and odd things, and things like hot and cold shooting does tend to even out.  One of the last competitive games I played, a title game in an over 35 tourny, i made 6 straight 3's in the second qtr, I was 1 for 5 the rest of the game, and to boot, was 1 for 6 from the foul line in the 4th qtr (I was 48, fatigue may have set in)

See - there are rational arguements to be made without resorting to name calling, I just made two for the seble fanboy crowd.

From a sample night of one, maybe 5 OR-involved games, I only had one odd thing happen, my all world sr sf missed 4 straight 3 pointers in the last 30-60 seconds to lose a game by 2 - if the change had gone overboard, he probably would have made one of those - too bad really - LOL - there is a 3rd argument supporting the fudge factor engine change - there you go - criticize me for that - HA

I do want to point out again (I did it earlier too) ... I have absolutely no problem with asking if the MOV and Fouls, Rebounds, Shooting %, and everything else (for that matter) are compared to Real Life.  I think the game should strive to have means and deviations that similar to real life for stats ... of course taking into account things like running FCP Defense or FB Offense with slow guys, etc.  (The FCP might add 1-3 fouls per game to the mean for those teams, etc.).

OR, I also don't think any of my comments were directed toward you except to answer your question. (Which I thought I answered with malice).

I do want the game to have similar stats to real teams ,,, we are simulating College Basketball so it should be similar.

I also am not blatantly supporting this change without monitoring and tweaks (if necessary).  I am only asking that we assume seble has the best interests of HD at heart, that he is trying to obtain suggestions and make changes that he thinks will be good for the game, and that is knowledgeable enough to do this without the first reaction being everyone thinks the game is ruined without the first game even being played.  (And OR, I am not saying you DID the things in this paragraph .. you in fact, did not do them.  Others in this thread, did, however).

If this adjustment is too much, they can cut it back a bit ... if it is not enough, they can adjust it to add more feedback.

Let us not forget that in an actual game that is being played, there is very little "Randomness" in what is happening.  When John Wall is driving against Jermaine Beal and puts up a shot, there is no RNG involved ... only the skill of both players.  Therefore if the RNG is not getting it right, then it might need to be adjusted.  The adjustment would be (in theory) to correct results that are not what is expected.  It is like the head coach recognizing that the guy is not performing correctly and talking to him to correct his shooting technique (or calling him out for being out of position on defense or not blocking out).  The guy is not performing up to his potential, he is obviously doing something wrong, so he is getting correction from the staff, and he is making the recommended adjustments.  Sounds normal to me.

That is not to say that I would support the change if a guy who shoots 35% from 3 and misses 5 shots would then have a 100% chance of hitting the next 3 shots to making him 3 of 8 for the statsheet.  But, if he is 0-5, I would not have an issue adjusting his percentage to 36%-38% for the next 3 shots.  (Of course, if he is outmatched on D,  I would want the corrected % to be used, not 35%.  If the defender makes it so he should shoot 30% (instead of his normal 35%, and he is shooting 0% instead of 30%, then adjust it to 32% ... but if he is shooting 70%, adjust it down to 28% instead of 30%)  Those numbers are just pulled out ... the point being, minor adjustment only ... and monitor the games closely after the update (and compare the deviation to real life and HD expected results).

So, OR, I have absolutely nothing against you or your question.  If it seemed I did, I apologize, it was late and my tone may have been more than I wanted to convey.

There is a classic old comedy called No Time for Sergeants staring Andy Griffith as a rube who is forced to join the military. The movie primarily focuses on his exploits in Basic Training. In one scene, he is required to separate a chinese ring puzzle. Instead of solving it the traditional way he bends one of the rings straight to separate it from the other ring. The test monitor (Don Knotts) goes ballistic because he didn't do it the correct way. I think that is analagous to the situation here in that the arguments seem to question the method rather than the results. If the outcome proves satisfactory, does the method of arriving at it really matter all that much?

Posted by girt25 on 8/29/2010 10:52:00 AM (view original):

Posted by jcfreder on 8/29/2010 10:37:00 AM (view original):
The thing is, if anybody was designing a basketball sim from scratch, you'd make each shot attempt independent of the others.  To their credit, this is how HD worked before.  However, OTHER flaws in the engine led to greater variance than you'd see in real life.  Using shooting % as an example, presumably the flaw was that each shot attempt had greater variance from a guy's true skill, leading to inordinant 1-10 or 9-10 games.  Now, instead of fixing the variance in each individual shot attempt, Seble has turned something HD gets RIGHT (independent shot attempts) into something HD gets WRONG (an 0-6 shooter will automatically be hotter).  Perhaps this awkward fix will actually improve the bottom line on games.  But at best it will do so by having early stats skewed one way, balanced out by later stats jerry-rigged to skew the opposite way, NOT by actually improving the simulation on a possession-by-possession basis. 

Best post on the topic so far.

I don't disagree with the gist of this post, but I really think you guys are assuming a much greater effect than it will actually be.

Isack.. My point is directly relayed to yours. If it hasn't been tested. No one knows...