+- Simulation Hockey League (https://simulationhockey.com)
+-- Forum: Community (https://simulationhockey.com/forumdisplay.php?fid=17)
+--- Forum: Announcements (https://simulationhockey.com/forumdisplay.php?fid=24)
+---- Forum: SHL Announcements (https://simulationhockey.com/forumdisplay.php?fid=27)
+---- Thread: New Changes to Regression (/showthread.php?tid=120657)
New Changes to Regression - Sean - 12-07-2021
Hello SHL,
We've taken into consideration some of the additional feedback we received and have tried to soften the blow of our new regression implementation. Fundamentally, we still stand by the decisions we made to move regression up one season but we did agree with some of the issues that were brought forward. The first change is our new permanent regression schedule will be changing slightly by lowering the first season of regression from 10% to 9%. The new scale is as follows:
Code:
10th season (Draft + 9): 9%
11: 12%
12: 15%
13: 15%
14: 15%
15: 18%
16: 20%
17: 22%
18: 22%
19: 25%
20: 25%
21: 25%
22: 25%
23 and higher: 30%We had some discussion with GMs about making regression even harder in the seasons after 23 but have decided to leave it as-is for a few reasons. The first is we don't want to completely kill players through regression, some people in the past have managed to play for 25+ seasons in the SHL and we don't want to completely kill the dream of someone wanting to do that now. Our goal was to make it so a team would be at a better advantage to call up a rookie than play an aging veteran which we feel this scale accomplishes. The second is we don't really have an issue with people playing too long with our current, less harsh scale. We'll revisit this if it becomes an issue but history has shown it hasn't been.
The more substantial of the two changes is we will be spreading out the transition to the new scale to 2 seasons instead of one to slightly soften the blow. The other benefit of this change is through our calculations, S54 and future classes will receive their spot at the top of the TPE leaderboard.
S63 Regression will be the following:
Code:
S55: 0%
S54: 7%
S53: 17%
S52: 18%
S51: 18%
S50: 18%
S49: 21%
S48: 22%
S47: 24%
S46: 24%
S45: 25%
S44: 25%
S43: 25%
S42: 25%
S41 and later: 30%S64 Regression will be the following:
Code:
S55: 9%
S54: 14%
S53: 17%
S52: 17%
S51: 17%
S50: 20%
S49: 20%
S48: 22%
S47: 22%
S46: 25%
S45: 25%
S44: 25%
S43: 25%
S42 and later: 30%The effects on the classes can be seen in these two charts: *These are using current class leader TPE values, actual TPE amounts will be slightly higher at the time of regression. Also should be "effects" but I'm not going back to change it.
![[Image: Old_Regression.PNG]](https://media.discordapp.net/attachments/905882280810119179/917840756021010442/Old_Regression.PNG)
![[Image: New_Regression.PNG]](https://media.discordapp.net/attachments/905882280810119179/917840755760979998/New_Regression.PNG)
We understand that this is still an aggressive plan to fix regression but that's been our intention all along. Hopefully, this helps ease the transition and rights some of the wrongs that we missed in our original implementation. We're not expecting to please our largest critics, but we did see some valid points that needed addressing and we feel these changes do that. I'll put a slightly more detailed chart in a reply below so you can see specific changes between the old proposed scale and the new proposed scale in one chart.
Thanks for working with us through this process!
On Behalf of SHL HO,
Tommy
RE: New Changes to Regression - Sean - 12-07-2021
![[Image: Capture.PNG]](https://media.discordapp.net/attachments/905882280810119179/917845611708031056/Capture.PNG)
RE: New Changes to Regression - steveoiscool - 12-07-2021
Ah shit, here we go again
RE: New Changes to Regression - boom - 12-07-2021
Good stuff. Glad to see HO respond to feedback.
Now I'm gonna go grab my popcorn.
RE: New Changes to Regression - Mazatt - 12-07-2021
Getting in on this one early so I can feel special
RE: New Changes to Regression - luke - 12-07-2021
The first is we don't want to completely kill players through regression, some people in the past have managed to play for 25+ seasons in the SHL and we don't want to completely kill the dream of someone wanting to do that now.
Tbf, there is really only like 30 people ever to play 20+ seasons so you don’t have to worry about it
RE: New Changes to Regression - sköldpaddor - 12-07-2021
I continue to acknowledge that I don’t know enough about the math to know whether the numbers look good so I will leave that for other people to comment on, but I very much appreciate that y’all took the time to consider user feedback and make changes here.
RE: New Changes to Regression - Spartan - 12-07-2021
At a glance you've remediated the issues the S53/S54's had with the regression changes, which is nice to see, although I may not fully understand the impact of the changes as someone not really impacted immediately. At the least, I hope this doesn't put people off of the league. I do respect the efforts to positively address and fix issues brought up by the community though, nice to see.
RE: New Changes to Regression - NamelessNate - 12-07-2021
So uhhhh, what was the thought process from changing it from 10% to 9%? Is that even a significant number of TPE? At 2k that's like 20 tpe lmao
RE: New Changes to Regression - luke - 12-07-2021
12-07-2021, 02:48 PMsköldpaddor Wrote: I continue to acknowledge that I don’t know enough about the math to know whether the numbers look good so I will leave that for other people to comment on, but I very much appreciate that y’all took the time to consider user feedback and make changes here.
Big number goes down, player maybe plays worse?
RE: New Changes to Regression - WannabeFinn - 12-07-2021
Strongly disagree with regression plateauing, but that’s always been a gripe of mine. At least now it’s up to 30%.
Glad we could work on finding a compromise in regards to the top end of the scale.
RE: New Changes to Regression - brickwall35 - 12-07-2021
No idea if this change will appease people, but extremely happy to see HO take feedback and work on a plan to try and accommodate reasonable issues.
RE: New Changes to Regression - Nhamlet - 12-07-2021
Is it possible to get the Old/New Proposed graphs in the same size/resolution @"TommySalami"
I made this and tried to resize it but it's not quite 1:1
![[Image: ezgif-2-9825862f1524.gif]](https://im2.ezgif.com/tmp/ezgif-2-9825862f1524.gif)
RE: New Changes to Regression - Sean - 12-07-2021
12-07-2021, 02:52 PMNamelessNate Wrote: So uhhhh, what was the thought process from changing it from 10% to 9%? Is that even a significant number of TPE? At 2k that's like 20 tpe lmao
Really, it was just a slight necessary change. When going from 2200 to 2000 TPE, it's a 22 TPE difference in TPE lost in regression, adding the 1% difference on top of that (making a 42 TPE difference) helps bring class peaks back into line.
RE: New Changes to Regression - Zema - 12-07-2021
First