07-16-2022, 04:51 AMMemento Mori Wrote:

07-16-2022, 02:05 AMgoldenglutes Wrote: I think this is the most important point. It feels like there should be a discussion for a middle ground where GMs aren't spending all their energy finding ways to perfect their team, but can still learn about the engine. Right now it feels like it won't be possible to learn anything about the engine unless you spend the next 10 seasons GMing to see a decent sample size of what works and what doesn't, or you play NHL saves in FHM 8 to familiarize yourself with builds/tactics/power levels that are probably not applicable to the SHL.


I don't have a perfect answer for how to reach that middle ground. Maybe there could be an unbiased sim testing committee (with people from multiple teams) that can do some general tests for builds/tactics and publish their findings for everyone to see?

This committee can also play a large role in steering the conversation around builds to ensure there's healthy variation. For example, instead of trying to find the one best build for all situations, they could try to answer questions like:

What are the attributes that contribute to playmaking?
How do you build a defensively responsible center?
Is there a viable way to build a speedy player that largely ignores defense?
Can you feasibly build a defenseman that specializes in puck possession and controlled zone entries?
What is the in-sim difference between a goalie focused on positioning and one that is more unorthodox?

Although the answers to the first two questions seem obvious, that's only with the context of how hockey works in real life. We all know that FHM did not make a perfect engine, and there are definitely going to be counter intuitive findings.


Back in the STHS days (before 40 passing), I remember there used to be build guides about common pitfalls to avoid and how to build viable players in each type of style. That's all I really want for the new sim engine.

Exactly. Just to draw attention to one of RED's points:

07-15-2022, 07:04 PMRomanesEuntDomus Wrote: That's a legitimate concern that I'll concede, but for other reasons than the ones you've mentioned. I see the risk here more in the fact that there clearly is a knowledge gap between teams/GMs at this point even without test simming, and it will be harder for the inexperienced GMs to catch up without test simming as one of their tools. That's why we need publicly available resources and essentially databases where people share their sim knowledge that are available to everyone, and I think it's fair to doubt that we'll get that anytime soon.

People sharing useful sim knowledge isn't doubtful, it's impossible to do so in a (legally) provable way. I will continue to ask my GMs for build advice - they know more about the sim than me, spent draft capital to acquire my player, I want the team to do well, etc. - and unless the SHL introduces a new rule saying its members are supposed to be deliberately unhelpful when other members ask for build advice, there are two potential (legal) strategies to follow: give build advice based on previous sim knowledge, or guess blindly given that the sim has changed.

I imagine most users/GMs will opt for the first option, and as series-based sports games rarely overhaul their engine in such a way that a large number of attributes go from good to bad or vice versa, this advice will probably still be largely good. Teams with a lot of sim knowledge will likely therefore continue to build according to the same logic - which is the actual reason teams like Hamilton were consistently good anyway, not test simming.

I dont know that I have a point here, but your first paragraph reminds me a lot of the first change to FHM6, which is undoubtedly an experience everyone in the league at the time had and also undoubtedly one we are trying -- in part or in whole -- to recreate. There are often users who want input on their build, and advice was usually based on a combination of real life hockey and an assumption for how the engine works. However, that gradually changed into advice based on "findings", and I put that word into quotes because these findings were often contradictory after a time. It was often said screening was useless. It was often said that a 2000 TPE goalie was not significantly better than a 400 TPE goalie. It was suggested that checking was very good and positioning was not. It was suggested that agility was better for shorter players but otherwise not as good as the other skating stats. Etc. Ad Nauseum.

However, I think this aggregate, social-based decompiling was a lot more fun to grapple with than knowing exactly the most effective way to build a team to win. We "knew", it seemed, what was good and what was bad based on thousands and thousands of sims split up between different teams and users, and of course the evidence that your best results came from copying ham strats 1 for 1.

Even the illusion of deterministic knowledge of the sim made the whole thing drab. Are we right in the confidence we have in what we know? It hardly matters. We can treat the fact or fiction of whether or not we had really cracked the code as the same, because the difference between us being right or wrong about that is immaterial to how the league feels about it. And, if it's not true, the pursuit of the REAL truth produces a league that isn't much fun to be in. Either you've made the best player, or you havent, but either way you'll know.

Obfuscation is a great thing, and being unsure that you're building a good player, or even being skeptical of advice you might get from anyone...well that's just part of the experience isn't it? If I wanted to solve a math problem, I'd be playing Factorio (which i am). I'm not here to see if I can score 55 goals or 60 goals (Whether I hit the + 5 or - 5 statistical chances), I'm here to run a long-term simulation with an agree-upon version of reality that matters, and where no other realities do. Is it cool to know that a team rolled the 20% chance to win the series? Kinda. But it's not as fun as knowing that your choices will have real consequences, that you can build a player that can be very very good or very very poor, and the driving force behind those consequences arent whether or not you played for one of the big 4 teams.