+- 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)
RE: New Changes to Regression - leviadan - 12-07-2021
I happily played SHL during 2020-2022, but one day HO removed the TPE component from my beloved Burlok’s player page, I cried myself to sleep, and on that day I realized what horrors centralized services can bring. I soon decided to quit. In 2023, searching for a new purpose in life, I discovered Bitcoin.
RE: New Changes to Regression - Carpy48 - 12-07-2021
12-07-2021, 03:16 PMbrickwall35 Wrote:12-07-2021, 03:05 PMlespoils Wrote: If my math is right you would become stagnant at ~630 TPE :Anyone willing to be an infinite 630 TPE BUG is a crazy person and outlier. Let them do it.
Reach 630
Regression by 30% (-190 TPE) = 440
Make up 190 TPE over the season which is the number they used I assumed it's some kind of average
Reach 630 again
Rinse and repeat
We might see infinite backup goalies but I don't see someone bringing in 190 TPE a season doing that either.
You're telling me there is chance?
RE: New Changes to Regression - PremierBromanov - 12-07-2021
shiny graphs
what was i talking about again?
RE: New Changes to Regression - nour - 12-07-2021
While things got a bit vitriolic at times, we do wanna thank everyone for the feedback they gave on regression prior to these changes. At the end of the day we’re just 6 guys trying to do as right by as many people as we can, and while we’re not always gonna release a perfect decision with perfect execution every time, we’re always going to at the very least listen to you guys and do what we can to make things right. Thank you for your patience over the last few days as we discussed these changes!
In addition to these changes, I’ve already said this before, but we do fully intend on tackling the FHM knowledge gap and shortening the season’s total length, which are both things that have been in discussion since the start of the season. I’ll have more information on both in this season’s State of the Union, which is coming in the near future!
RE: New Changes to Regression - CampinKiller - 12-07-2021
12-07-2021, 03:51 PMnour Wrote: While things got a bit vitriolic at times, we do wanna thank everyone for the feedback they gave on regression prior to these changes. At the end of the day we’re just 6 guys trying to do as right by as many people as we can, and while we’re not always gonna release a perfect decision with perfect execution every time, we’re always going to at the very least listen to you guys and do what we can to make things right. Thank you for your patience over the last few days as we discussed these changes!
In addition to these changes, I’ve already said this before, but we do fully intend on tackling the FHM knowledge gap and shortening the season’s total length, which are both things that have been in discussion since the start of the season. I’ll have more information on both in this season’s State of the Union, which is coming in the near future!
Honestly re: season length, go the PBE route and quick sim more in a day. Alternatively, do their playoff route of doing one round each sim
RE: New Changes to Regression - MyLittleHexx - 12-07-2021
12-07-2021, 02:47 PMluketd Wrote: 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
Out of spite I'm going to make that 31 people if I can help it
RE: New Changes to Regression - sve7en - 12-07-2021
12-07-2021, 03:15 PMRuggsy Wrote: I agree with however 7 ends up responding
They keep doing these while I'm at work gdi
RE: New Changes to Regression - Chevy - 12-07-2021
Applied. Hopefully someone will find use for Stroms corpse.
Also @"luketd", who gives out the 2k awards? I'm still waiting on mine from last season.
RE: New Changes to Regression - luke - 12-07-2021
12-07-2021, 04:23 PMGeekusoid Wrote: Applied. Hopefully someone will find use for Stroms corpse.
Also @"luketd", who gives out the 2k awards? I'm still waiting on mine from last season.
Tomen
RE: New Changes to Regression - RomanesEuntDomus - 12-07-2021
Looks reasonable. Even though I argued for the changes overall, it's good to see some of the pain that people felt being alleviated.
RE: New Changes to Regression - sve7en - 12-07-2021
Aight,
@nour, @"TommySalami", HO, I commend you guys for both sticking to you guns and hearing us out when we've had complaints. While I don't agree with the 9+D regression move, I can respect where you think it holds value and this change is a step in the right direction. To some extent this is just what I get for not applying for your job. The use of multiple transitional regression seasons was something I considered earlier today and I'm glad we are on the same page there.
That being said, while these changes are immensely appreciated, they don't exactly solve the whole issue.
I've compared the old plan and this new one on three similar metrics, sum of the percent TPE relative to the league's max over three, five, and seven seasons to track how this change affects peak (3 season max), plateau (5 season max), and time at the top (7 seasons, which comes out to 90% league max or ~2080/2280 on the old system and 83% league max or ~1740/2805 on the new). In the old plan you can clearly see the disadvantages applied to S53 and S54 and the subsequent rebound to the new norm that we called out as unfair. The peak players enjoyed (100/99/97) was skipped for these classes, this omission was felt at the longer scales as well obviously, but in those longer scales we can see the positive effect of less dominance over the 5 and 7 season time frames. The revised scale starts to amend this (while also marginally boosting the future arcs ironically due to the 9% allowing for essentially 100/100/97 seasons instead of 100/99/97), but the effects aren't incredibly impactful.
While my idea isn't perfected yet (low time to turn this back around, etc.) this shift in peak, plateau, and time at the top can be better managed still, and I hope we're not going to stop with the slight changes that we're making.
OLD
![[Image: unknown.png?width=1440&height=515]](https://media.discordapp.net/attachments/509021444429053954/917894957086429194/unknown.png?width=1440&height=515)
NEW
![[Image: unknown.png?width=1440&height=515]](https://media.discordapp.net/attachments/509021444429053954/917894981535019018/unknown.png?width=1440&height=515)
MY IDEA (forgive the conditional formatting at the bottom)
![[Image: unknown.png?width=1440&height=515]](https://media.discordapp.net/attachments/509021444429053954/917895026569248768/unknown.png?width=1440&height=515)
RE: New Changes to Regression - sve7en - 12-07-2021
At the cost of double posting, I think managing the shift in those metrics and ensuring every class gets that 100% are the core priorities of this. Those constraints are obviously well managed by an extra transition season, though I also agree that doing more than that leads us to an awkward path of class specific regression curves as we look to balance numbers that I think is best to avoid.
RE: New Changes to Regression - Symmetrik - 12-07-2021
12-07-2021, 04:04 PMCampinKiller Wrote:12-07-2021, 03:51 PMnour Wrote: While things got a bit vitriolic at times, we do wanna thank everyone for the feedback they gave on regression prior to these changes. At the end of the day we’re just 6 guys trying to do as right by as many people as we can, and while we’re not always gonna release a perfect decision with perfect execution every time, we’re always going to at the very least listen to you guys and do what we can to make things right. Thank you for your patience over the last few days as we discussed these changes!
In addition to these changes, I’ve already said this before, but we do fully intend on tackling the FHM knowledge gap and shortening the season’s total length, which are both things that have been in discussion since the start of the season. I’ll have more information on both in this season’s State of the Union, which is coming in the near future!
Honestly re: season length, go the PBE route and quick sim more in a day. Alternatively, do their playoff route of doing one round each sim
PBE still allows teams to make changes between games (there's a discord channel for simmer + playoff GMS) where teams can inform the simmer between games and so you can get through a whole round, and changes can be made. No time for extra testing though or anything, which also kind of pairs nicely with trying to cut down on how much teams can fight for FHM advantages.
I'd be curious to see how much the mid series adjustments lead to wins and if by playing the games in a shorter timespan actually helps parity.
RE: New Changes to Regression - Aleris - 12-07-2021
My goal is to now force some poor GM to play me for 25+ seasons
RE: New Changes to Regression - LordBirdman - 12-07-2021
12-07-2021, 03:13 PMTommySalami Wrote:12-07-2021, 03:11 PMLordBirdman Wrote: Glad feedback was taken to heart by HO. I think it is a step in the right direction for improving player turnover which hopefully will lead to more roster parity, but these changes won't meaningfully improve actual parity because it doesn't address the FHM knowledge gap issues.
I do want to say we aren't ignoring this issue and we are trying to work towards a solution on the FHM side of things.
Very glad to hear this.