I know in the past they've striven for accuracy in the percentages, but I don't see how they could compare it by offensive set.  How many teams run a true Triangle?  How many teams run a true motion?  Nobody runs the same FCP that HD features.  

They compared the stats in HD by set, not real life sets.  It was more of a comparison between HD sets.  For example (and I'm going off the top of my head) it was found that zone gave up 8% O boards but had a better FGA% (1% lower) than m2m.  Press caused more TOs but fouled more.  They had numbers.  It was good.  Changes were made based on the self-evaluation.

I'm also curious as to the comment that seble had in regard to upsets in the dev chat.  He said there are less than before.  Why play the game?  What is an upset in HD?  Using the spread isn't the best way since those are all over the board.  What are the expectations?

did someone at HD actually do this?

I remember a few years ago I looked at all the D1 teams in three worlds and posted the totals.  Totals from all the HD worlds would be much more complete than the 1000 teams I used.

i get the feeling that its not individual sets that require the tweaking right now, more so all the sets

Posted by Iguana1 on 11/17/2010 8:58:00 AM (view original):
did someone at HD actually do this?

I remember a few years ago I looked at all the D1 teams in three worlds and posted the totals.  Totals from all the HD worlds would be much more complete than the 1000 teams I used.

I have asked seble on a forum thread to provide us such info when the new engine came out, he ignored me.

I am not even concerned that it is all that encompassing, just some comparisons would be interesting. 

Might be he has the numbers and is working to get them better b4 releasing them?  Were I him, that would be my major focus with the engine, while trying to similtaneously increase functionality, like depth charts vs defense type, same as shot distro's are done.

One thing I recall about data, tarek once published something that I thought gave away too much, it was not intentional - it was related to -5 actually, but I used that one thing he posted ages ago more than any one tool in this game, and it was related to this sort of thing, so if seble is trolling on this post, I would warn him to be somewhat careful about what he releases, there is a limit where it could go too far.

Originally posted by Admin (TK) 3/13/2006  Google Internet Archives Wayback Machine for more whatif archives.

weena - you are the man!

I've asked Seble about this and he says he has no intention of sharing this data with the current engine. I dont understand why not but I'm sure he has his reasons.

maybe it's embarrassing?

also, why didnt tarek show blocks...?

Posted by Iguana1 on 11/17/2010 8:58:00 AM (view original):
did someone at HD actually do this?

I remember a few years ago I looked at all the D1 teams in three worlds and posted the totals.  Totals from all the HD worlds would be much more complete than the 1000 teams I used.

I haven't seen anything recent that is why I bring it up.  I'm guessing the engine has had many tweaks to it.  I'm curious as to how it's playing out.  I'd suggest using a 10-day window with the most recent engine tweak.  That would give them about 40,000-60,000 games since; having a small time frame is key because looking at data from a previous engine tweak is not applicable.  Is something you'd be willing to do seble?  And by division too?

I think giving us something would yield peace of mind as it did when it was released by tk a few years ago.  The info weena provided was confirmation of that, but TK also provided data per defensive set in order to try and dispel the "press is magic" conversations.  While the press was found to commit a few more fouls, and yielded a slightly higher FGA% and slightly lower def. rebound %,  the turnovers generated more than made up for it.  Perhaps that is what you were referring to, Ol'.

As I think of it, I think he did it twice (once to discuss defensive positioning and the other to compare sets)

I think TK was pretty forthcoming in talking about randomness et al.   He talked about upset % and that about 1/4 of games are upsets irl (games where there is a 5 point spread or something like that).  Can't remember how he defined them, but I'm curious as to what's currenty taking place (other than a heckuva lot of tip ins by a kid who just missed a fadeaway from the free throw line) in the way of upsets.  Why play the game if upsets are not in line?  Of course the "line" is much more accurate irl than in HD.