05-20-2023, 07:49 PMCitizen of Adraa Wrote: So what I see is what I could probably gather from the league myself anyway - we kinda don't have mid table teams at all. Half the league at any given moment is in "fighting for cup" caliber, half the league is pretty much openly tanking or at least freshly out of the position, and we get absolute minimum of sort of "ok, but like round 2 ok" teams, which isn't fantastically realistic compared to the NHL. The way GMs here operate is that we are far more likely to just explode a middling team on the spot, because there is very little punishment for not performing unlike a real GM could face while trying to rebuild/retool, and that then translates into a state where teams are on either extremes of performance, and very rarely in the middle. That then creates the issue of playoffs being seemingly very uncompetetive in round 1, and a lot of blowouts - it's effectively as if you attempted to mash 10 NHL playoff teams and 10 AHL teams and tried to pretend there is ever going to be any competition between them. I think it's more of a symptom of the way the league is, rather than an outright issue with the sim itself, and I think it's going to be very hard to properly fix unless we actively put in rules that would for example allow HO to replace GMs underperforming for large periods of time continuosly, therefore making GMs far less likely to explode a team even if the rebuild hasn't exactly worked to at least hit some .500+ seasons, and to not get fired for underperforming.

05-20-2023, 07:49 PMCitizen of Adraa Wrote: So what I see is what I could probably gather from the league myself anyway - we kinda don't have mid table teams at all. Half the league at any given moment is in "fighting for cup" caliber, half the league is pretty much openly tanking or at least freshly out of the position, and we get absolute minimum of sort of "ok, but like round 2 ok" teams, which isn't fantastically realistic compared to the NHL. The way GMs here operate is that we are far more likely to just explode a middling team on the spot, because there is very little punishment for not performing unlike a real GM could face while trying to rebuild/retool, and that then translates into a state where teams are on either extremes of performance, and very rarely in the middle. That then creates the issue of playoffs being seemingly very uncompetetive in round 1, and a lot of blowouts - it's effectively as if you attempted to mash 10 NHL playoff teams and 10 AHL teams and tried to pretend there is ever going to be any competition between them. I think it's more of a symptom of the way the league is, rather than an outright issue with the sim itself, and I think it's going to be very hard to properly fix unless we actively put in rules that would for example allow HO to replace GMs underperforming for large periods of time continuosly, therefore making GMs far less likely to explode a team even if the rebuild hasn't exactly worked to at least hit some .500+ seasons, and to not get fired for underperforming.

05-20-2023, 07:46 PMspooked Wrote: Looks like I was expecting, 10% of all games being won by 6+ goals is a lot, 18~% when you drop it down to 5+ seems a bit much too. from what I can tell NHL only had 4 games with 8+ goal differential, and 5 at 7, with more games played total. So we probably have a pretty large bias toward blowouts.

05-20-2023, 07:59 PMluke Wrote: Yes but the NHL also doesnt have this many teams just tanking.

This season there is a 100 point difference from 1st in the league and last in the league. It isnt like that in NHL. This is something where you just can't compare it to the NHL. Its asinine because the incentives are different from NHL to SHL. Also we are on an engine we aren't playing it on the ice. The players pride isnt on the line. Also the average Goals Per Game is completely different between the NHL and SHL, because in FHM we set the goals per game

and again, tell me the last time a NHL team has won 6 games in 1 season. Like toronto did this year

05-20-2023, 08:00 PMhockeyfan Wrote: Should we get rid of teams?

05-20-2023, 07:59 PMluke Wrote: Yes but the NHL also doesnt have this many teams just tanking.

This season there is a 100 point difference from 1st in the league and last in the league. It isnt like that in NHL. This is something where you just can't compare it to the NHL. Its asinine because the incentives are different from NHL to SHL. Also we are on an engine we aren't playing it on the ice. The players pride isnt on the line. Also the average Goals Per Game is completely different between the NHL and SHL, because in FHM we set the goals per game

and again, tell me the last time a NHL team has won 6 games in 1 season. Like toronto did this year

05-20-2023, 09:37 PMspooked Wrote: Ottawa won 10 in their first season in existence tbf.

And yes we are not the NHL, but that doesn't mean we can't look at OUR results and think about maybe we could adjust something so game results aren't quite so one sided just for pure entertainment of OUR league. I have no idea where the GPG setting is at right now, but maybe it needs a tweak atm.

05-20-2023, 10:09 PMluke Wrote:

Here I did the math. About 7200 games for each league. Basically, we have a lot less 1 goal games, a slightly more 4,5,6,7 goal differential games. I dont see the problem. NHL has more 1 goal games because you are playing in the real world. In FHM, its not the real world, its a simulation.


When you look at the SHL dataset



Which is 7,260 games, having those numbers is good! We are talking about 46% are two goal games, another 35% are 3/4 goal differential games(which lines up with the NHL at 33%). We just see more 5+ goal games(1304 of them) because you have 1600 TPE average teams going against 800 TPE average teams. Its a large spread of really good teams, and really bad teams

05-21-2023, 08:14 AMRomanesEuntDomus Wrote: Thanks for your post luke and for bringing some more data into this. I think it's also useful in clarifying some of the language around the topic. We keep talking about "parity" when you are right that this isn't actually really the core of the issue anymore, there are enough different teams winning at this point and not just singular powerhouses who stay on top forever. The core issue as I would define it is "predictability", the fact that at the start of a season you already pretty much know instantly who will have a chance to compete and who won't, and the latter category isn't just limited to the tanking teams, but includes the majority of the midfield as well. The surprise runs and upsets are so rare not just in #1 vs #16 series, but when the #4-#6 seeds are playing the #10-13 seeds as well, which is when thing should start to become wide open, but they don't.

CoAs observation that there isn't really a midfield anymore is a direct result of that, because naturally GMs have started to adapt to this competitive landscape. Over the last few years they have slowly learned that there isn't really a point in hanging around in that 7th to 12th place range, play a solid regular season and then hope to do some damage in the playoffs if you manage to run hot. They know that it's boom or bust, you either are among the absolute top of the league or in a rebuild, and everything between those two extremes is a waste of time. I don't think that's a good situation overall and having almost half of the league tanking/rebuilding at the same time doesn't seem like something we'd want to see.

Also, an additional point that we could take from your data, is considering changes to playoff seeding. I don't know how many options FHM gives us there, but maybe the fact that the same teams keep running into each other at the same points in the playoff tree are part of the problem. I have definitely heard some people talk about how they think it's boring that they are playing the same teams again and again in the playoffs, that their first and second round series are all against the same teams etc. It's not an issue I've looked at in-depth at all, but maybe there is some potential here to make things better in a rather non-invasive way.

05-21-2023, 01:27 AMspooked Wrote: If you look at it from the other side, it's not that good really though. the NHL ratio would see only ~4 games per team (80 games) end at 5+ goal diff, we have avg of ~12 per team (240 games), which is a pretty significant chunk of the season when you consider that is the average. I understand the spread will be larger to some degree cause the way things are right now, but my whole feeling is that it's kind of a turn off to see teams getting slaughtered in a pretty significant chunk of games, and it's not just the TPE mismatch games either. NEW lost to PHI in a couple crazy games this season too 8-1 and then 9-6, which only shakes out at a 3 diff, but that wouldn't even look bad in this metric despite clearly being an insane scoreline.

If you want to attribute the weird results to TPE mismatches fine, it doesn't take long to find examples that disprove that being the sole reason, but fine. But if that is the case, why wouldn't we want to raise the replacement level players ability to compete so games are so stretched out to a completely unrealistic level? Losing is one thing, having be a complete slaughter is another.

And if you look deeper beyond just diff, why are some games ending at like 15+ goals in the first place? Where do we have the goal slider at where this is happening? I just think there are a lot of weird abnormalities in the results the sim has been spitting out in specific games that are either completely insane, or doesn't even feel close to realistic for the matchup, even if one team "had an off game" the scorelines are insane. If the league has no appetite to simulate a professional sports league, fine, but to me that was kind of the whole point.. and the engines we use are trying to do that as well, so it kind of makes sense we are getting wonky outcomes when we aren't using the engine close to how it was designed.

Quote:Second - As many of you have noticed, scoring overall, particularly in individual games has increased tremendously to a level we don’t think is realistic (and definitely isn’t enjoyable for goalies, who admittedly already have a rough go of it in FHM). We are still gathering data on the reason for this, and ways to address it, but for S68 we will be decreasing the average number of goals per game from the current value of 6.27 to 6.1 (this is the value used by FHM for NHL leagues).