Hoops Dynasty = Tale of 2 Halves Topic

I've been thinking this for the last 3-4 season but after yet another example with my Wisconsin team, decided to pot something.  I've talked this with a couple fellow coaches and they see the same thing.

When I read through my boxscores it seems like the 1st and 2nd halves are played by 2 entirely different teams.
MANY times a double digit underdog will play a dead even 1st half and then get blown out in the 2nd half.    Between my 2 teams I see this time and time again.  To the point where if I am even at half with an underdog, I am fiarly confident the 2nd half will turn around.

It's almost as if in the 1st half of games the engine has total variabililty on each posession and then in the 2nd half, the engine adjusts what the final stats should look like to have them make sense with the players' ratings.

I remember a thread some time this year where they talked about the engine adjusting to get total stats within what is expected (ie - an A FT player who goes 0-4 in the 1st half will most likely make all his FTs in the 2nd half).   I am guessing this is what I am seeing.

3/21/2013 11:07 AM
This is actually quite common in college basketball. You'll see a lot of poor teams keep games really close and then around 15:00 to go in the 2nd, the better team starts blowing them out.
3/21/2013 11:36 AM
What cj said. 

It's actually not that common for a better team to just steadily outscore you throughout the game. More likely, they go on a big run at some point. Same in WIS. If that big run happens in the second half, then you get this phenomenon. 
3/21/2013 1:57 PM
Posted by cjsweat2 on 3/21/2013 11:36:00 AM (view original):
This is actually quite common in college basketball. You'll see a lot of poor teams keep games really close and then around 15:00 to go in the 2nd, the better team starts blowing them out.
+1
3/21/2013 3:38 PM
But this isnt treal life college basketball.  This is a computer simulation where, statistically, the 1st half should be close to the 2nd half.    If this is on purpose, then WIS is intentionally adjusting the RNG in the 2nd half.    Which is really the point of this post.
3/21/2013 5:21 PM
Have seen this a lot where I'll be losing at half to a mediocre team then "turn it on" in the second half.  I've also been on the other end of that, leading at half time and then getting blown out in the second like we stayed in the locker room.
3/21/2013 5:24 PM
Posted by mullycj on 3/21/2013 5:21:00 PM (view original):
But this isnt treal life college basketball.  This is a computer simulation where, statistically, the 1st half should be close to the 2nd half.    If this is on purpose, then WIS is intentionally adjusting the RNG in the 2nd half.    Which is really the point of this post.
mully is absolutely right. the game was always designed without those human elements, momentum, the nature of teams comfortably ahead to soften up and give up big runs, etc... and thus, statistically, the first and second halves should be more tightly correlated than real life.

mully, you are also right about the fact that there are more "tale of 2 halves" type games than there used to be, and it all comes down to seble's feedback change. when a team dominates another in 1 half, its likely that the winning team has been overshooting, over rebounding, etc - and the losing team, the opposite. because feedback is a mechanism that kicks in over time, and it takes a while for imbalances to build up/for feedback to kick in, it is a lot more common in these days, to see wildly different 2nd halves - the feedback is pushing for it - resulting in much less of an RNG difference being required to see such a difference in performance by half.
3/21/2013 8:00 PM
Thanks for the confirmation.   Is this a byproduct of Seble's move to reduce the overall variation in the engine (ie = less upsets)?

If so I can live with that. 
3/21/2013 8:39 PM
Posted by mullycj on 3/21/2013 8:39:00 PM (view original):
Thanks for the confirmation.   Is this a byproduct of Seble's move to reduce the overall variation in the engine (ie = less upsets)?

If so I can live with that. 
yes, thats what im saying - feedback is the mechanism that accomplishes what you describe. meaning, a guy who shoots 50%, after say 5 straight misses to open the game, now his 6th shot, he might be shooting 60%, not 50% (just for example) - so the shooting function uses feedback from the previous functions, to determine the chance of a makes/misses.

so, pretty much any time you have a fairly balanced match up, and one team has a killer first half and takes a big lead, feedback is going to be working against them, pushing them to shoot lower, rebound poorly, etc - in an effort to reduce deviation from the statistical mean.

on the other hand, if you have a team 20ppg better than another, and they are up by 10 at the half, then roughly speaking, feedback wont be on anyone's side
3/21/2013 9:22 PM
Posted by mullycj on 3/21/2013 5:21:00 PM (view original):
But this isnt treal life college basketball.  This is a computer simulation where, statistically, the 1st half should be close to the 2nd half.    If this is on purpose, then WIS is intentionally adjusting the RNG in the 2nd half.    Which is really the point of this post.
Totally disagree. 

Yes, they may be adjusting the RNG in the 2nd half. Seems like there are some indications from reliable people that they are. 

But even if they weren't, I'd expect to find a lot of tale of two halves type games. Statistically, a big run by a superior team is likely to come around at some point. Statistically, a stretch of even play by an inferior team is going to come at some point. If the latter comes in the first half and the former in the second half, you have what you're talking about. 

Are they helping it along with adjustments? Probably. But it never really seemed that strange to me in general. 
3/21/2013 10:51 PM
billyg, I'm not sure that I agree with your line of thinking here. Even with a totally random set of events -- coin flip, whatever -- you're going to very often have runs like you see here. Flip a coin 20 times and you could get 14 heads, flip it 20 more times and it comes up 14 tails. And this game is obviously way more complicated that, with more moving parts and more things that can impact the outcome. With all of that going on, one half is a tiny sample size, and seeing frequent variation from one half to the next is something that I'd expect.
3/21/2013 11:41 PM
Case and point:

http://whatifsports.com/hd/GameResults/BoxScore.aspx?gid=9272023
3/22/2013 6:59 AM
Posted by hobobilly on 3/22/2013 6:59:00 AM (view original):
Case and point:

http://whatifsports.com/hd/GameResults/BoxScore.aspx?gid=9272023
I don't consider that remotely weird, unusual, surprising or off putting.
3/22/2013 7:05 AM
Definitely with girt. 

I had a game last night in Iba that I thought made my point pretty well. Was playing a much inferior team, went down seven midway through the first half. Went on a big run, took a 12-point lead into the break and ended up winning by 21. 

As it stands, that game seems pretty normal. Have a few missed shots in a row, get a deficit, but, as the better team, eventually start getting hot, outscore them by 19 in an eight minute stretch, and then cruise through the rest of the game, building the lead much more slowly from 12 to 21.

If that seven-point deficit is there at the end of the half instead of the under eight timeout, and the run is at the beginning of the second half instead of the first, it pops out in memory as one of those "tale of two halves" things. My suspicion is that this happens all the time and that we just notice it most often when it breaks along half boundaries. 

Same thing happened to me today in Crum. Down 2 at the under eight, go on a run, up 19 at the half. Happens all the time. Nobody notices. 

3/22/2013 7:36 AM (edited)
Posted by girt25 on 3/21/2013 11:41:00 PM (view original):
billyg, I'm not sure that I agree with your line of thinking here. Even with a totally random set of events -- coin flip, whatever -- you're going to very often have runs like you see here. Flip a coin 20 times and you could get 14 heads, flip it 20 more times and it comes up 14 tails. And this game is obviously way more complicated that, with more moving parts and more things that can impact the outcome. With all of that going on, one half is a tiny sample size, and seeing frequent variation from one half to the next is something that I'd expect.
+1.
3/22/2013 10:09 AM
12 Next ▸
Hoops Dynasty = Tale of 2 Halves Topic

Search Criteria

Terms of Use Customer Support Privacy Statement

© 1999-2024 WhatIfSports.com, Inc. All rights reserved. WhatIfSports is a trademark of WhatIfSports.com, Inc. SimLeague, SimMatchup and iSimNow are trademarks or registered trademarks of Electronic Arts, Inc. Used under license. The names of actual companies and products mentioned herein may be the trademarks of their respective owners.