What simulating style do you all use?

[renojedi]

I chose "Own League 2D, rest classic". First season had very high goal numbers with the three league leaders having 62, 58, and 56 goals.

Now I have the 2022-23 NHL league goal leader (Mika Zibanejad) with 78 goals at the end of the season and the players behind him have 66 and 62 goals respectively.

I play out all my team games in the 2D simulator and those totals seem more realistic. Would I get better results with "Own team 2D, rest classic"?

[jumpSeat]

Originally Posted by renojedi

Would I get better results with "Own team 2D, rest classic"?

I'm playing on this setting right now, and I have to say my scoring totals seem a bit higher than the 'classic' numbers. Not ridiculously so, but in terms of goals and assists a good 10 to 20 points.

[skunt3m]

I just ran a handsoff sim for 3 seasons using the latest patch, current rosters and the "own league" 2D setting. All minor/foreign leagues deactivated. FHM 8 defaults to 5.94 goals per game.

Beyond what you're reporting, I've seen screenshots of McDavid, Matthews, etc with 70+ goals on Discord, so I assumed scoring would be too high, but my results were pretty reasonable:

21/22:
5.82 G/G.
League leaders in goals: 48, 47, 46.
League leaders in assists: 64, 63, 62
League leaders in points: 100, 99, 98

22/23
6.21 G/G
League leaders in goals: 54, 53, 51
League leaders in assists: 79, 70, 68
League leaders in points: 121, 117, 104

23/24
6.18 G/G
League leaders in goals: 53, 46, 45
League leaders in assists: 74, 72, 70
League leaders in points: 116, 115, 114.

A few thoughts:

--Are you using the current rosters? I know in FHM 7 one of the things they did to reduce scoring in the 2D engine was reduce ratings for some of the top players. That could account for the difference in our results.

--Are any of the big outliers on super teams? The truly crazy results seem to occur when people stack their teams with Panarin, McDavid, Matthews, etc. The AI can luck into that kind of situation too. In my test, Zibanejad scored 37, 35 and 45 goals:

21/22
Panarin: 82-37-62-99
Zibanejad: 65-37-43-80
Kreider: 80-31-39-70

22/23
Panarin: 72-34-49-83
Zibanejad: 75-35-47-82
Kakko: 80-26-43-69

23/24
Panarin: 80-45-70-115
Zibanejad: 82-45-69-114
Kakko: 76-45-45-90

Obviously RW is lagging behind a little. Add a Pastrnak to the mix and maybe you are looking at 60+ goals for Zibanejad.

--Scoring in FHM always jumps in the 2nd season. In the classic engine it would climb for a few years, plateau and eventually decline below whatever you set G/G at. I hadn't tested the 2D engine until today but the results seem similar. Is chemistry too strong/built too quickly? Are young players developing too quickly? Are old guys dropping off too slowly?

Anyway, not arguing with you or claiming things are perfect. Maybe not having a human in the mix or solely running the NHL skewed my test. I'll try again at some point and report back.

[renojedi]

Thanks for the detailed response.

I am using the latest update with current rosters and default settings. I added two expansion teams (beyond Seattle) to the regular NHL set-up so that may have affected things since rosters got watered down and jumbled between that and other AI initiated moves.

In that second year, Mika had 78 goals and 130 points. The rest of the team was more realistic and were the same players from the year before...Panarin (39 goals), Chytil (30), Lafreniere and Kreider with 24 and everyone else in the teens. Not sure what made Mika go to 78 as they were the one division without an expansion team to beat up on. League GPG was 6.12 and the Rangers were second in the league with 4.0 goals per game.

In year three, with an expansion team where no one had more than 25 goals in year one, and only one hit 33 in year 2, I have a newly signed MacKinnon who has 26 goals through 32 games which seems about right especially on a team without other options. But then McDavid has 39 in 30 so we'll see if we have another 70+ goal scorer.

I know none of this helps isolate the variables that might figure out what is going on but thought I'd share since you asked. Thanks again for your detailed test to show the baseline should be fine.

[skunt3m]

You know what, I just realized that not selecting a team with the "own league 2D, rest classic" option means that my test was probably entirely in the old engine. I'd delete my post in shame but you already saw it! I'll try again with "all games in 2D" later today and confirm/compare.

[crusadertsar]

I'm running 2D for NHL and classic for the other leagues. And honestly I am loving the fact that you can finally create some pretty high scoring tactics! I find that has always been my objective in all the previous game but never worked well until now.

First season had some crazy scoring on my own team (New York Rangers) with Panarin getting 79 (too bad was only one short of 80 ) and Mika 64. But I am running a pretty crazy Swedish Torpedo tactic. with my 1st line using 4 forwards and one defenceman.

Was inspired by this:
https://en.wikipedia.org/wiki/Torpedo_system
https://www.ricktraugott.com/post/20...torpedo-system

[crusadertsar]

Originally Posted by skunt3m

You know what, I just realized that not selecting a team with the "own league 2D, rest classic" option means that my test was probably entirely in the old engine. I'd delete my post in shame but you already saw it! I'll try again with "all games in 2D" later today and confirm/compare.

But isn't that the option by default? I never even knew that there was option to choose how you want to sim so I never made any changes in my game. But I when I checked my save the other day I saw that it was pre-selected "own league 2D, rest classic". Thats probably what devs want us to use by default anyway otherwise the load times will eventually become too crazy long.

[rmt1982]

Originally Posted by crusadertsar

I'm running 2D for NHL and classic for the other leagues. And honestly I am loving the fact that you can finally create some pretty high scoring tactics! I find that has always been my objective in all the previous game but never worked well until now.

First season had some crazy scoring on my own team (New York Rangers) with Panarin getting 79 (too bad was only one short of 80 ) and Mika 64. But I am running a pretty crazy Swedish Torpedo tactic. with my 1st line using 4 forwards and one defenceman.

Was inspired by this:
https://en.wikipedia.org/wiki/Torpedo_system
https://www.ricktraugott.com/post/20...torpedo-system

How do you manage the 4 forwards idea, just drop a 2 way forward into defence, or use a more attacking minded D man?

[oldfatbaldguy]

I'm running 2D on everything with all leagues active, and it's slow, but it's not THAT slow. Sometimes I get a little bored waiting for the next day to start and I check a couple of things on my phone. It's worth it to me for the presumed uptick in realism. I'm playing with a 3.6GHz four-core processor (Ryzen 5 2400g) and 8 GB installed RAM.

[crusadertsar]

Originally Posted by rmt1982

How do you manage the 4 forwards idea, just drop a 2 way forward into defence, or use a more attacking minded D man?

I use my best two-way forward. So for the first line, my blitzkrieg torpedo line i use Lafreniere and Kukko as power forward torpedoes and a two-way forward replaces one of the defenders.

[skunt3m]

Originally Posted by crusadertsar

I'm running 2D for NHL and classic for the other leagues. And honestly I am loving the fact that you can finally create some pretty high scoring tactics! I find that has always been my objective in all the previous game but never worked well until now.

I agree, I like the idea of being rewarded for a crazy tactic or assembling perfectly complimentary linemates. It is also hard to gauge what could be possible since NHL coaches are typically very conservative. With that said, nearly 80 goals from a guy who has never scored more than 32 seems excessive to me, no matter how you have him playing. Accounting for era it would be the greatest goal scoring season of all time!

Originally Posted by crusadertsar

But isn't that the option by default? I never even knew that there was option to choose how you want to sim so I never made any changes in my game. But I when I checked my save the other day I saw that it was pre-selected "own league 2D, rest classic". Thats probably what devs want us to use by default anyway otherwise the load times will eventually become too crazy long.

Yes, you are right. I was intending to run a quick test of the 2D engine specifically. The game runs faster if you are unemployed. It just didn't occur to me that setting myself as unemployed for my test would put the entire NHL in the "rest classic" category.

[rmt1982]

Originally Posted by crusadertsar

I use my best two-way forward. So for the first line, my blitzkrieg torpedo line i use Lafreniere and Kukko as power forward torpedoes and a two-way forward replaces one of the defenders.

That makes a lot of sense to be fair.

[skunt3m]

I ran a set of tests - just 21-22 so far (2D is slow, lol), in both Classic and 2D. I tried to see if some of the old soft spots in the simulation still existed in 2D/FHM8.

Scoring leaders:


Ice-time:


Shots leaders:


Where did the assists go?


Keep in mind that this is just one test. Hopefully others can weigh in with their results. Some conclusions:

-Top players score too many goals and not enough assists.

-Some of the shortfall in assists could be explained by ice-time distribution amongst d-men. Top guys not playing enough, bottom guys playing too much.

-Everyone shoots too much, especially top d-men.

[skunt3m]

22-23 just finished simulating in 2D.

G/G crept up to 6.20. Overall, scoring for top 10 and 25 forwards was about the same as in 21-22. I'll spare you guys the table this time. Too many shots, too many goals, not enough assists, overall points about right.

One big outlier was Nathan MacKinnon, who scored 73 goals in 74 games. This make me wonder if chemistry effects are too strong with elite players (does this line start with "great" chemistry?), as MacKinnon's attributes have not changed.

Landeskog: 79-34-73-107
MacKinnon: 74-73-52-125
Rantanen: 66-36-46-82

[skunt3m]

In my 22-23, MacKinnon scored his 73 goals in 309 shots, giving him a shooting percentage of 23.6. The only guys to shoot higher than 20% with 300 or more shots are Wayne Gretzky, Mario Lemieux, Mike Bossy and Brett Hull.

With that in mind, I decided to check if shooting percentages were out of whack (see table). They were not for forwards, at least on the whole. Excess goals are coming from excess shots. Defencemen are not shooting well. They are taking more shots and scoring around the same number of goals.

One of my gripes with OOTP's historical mode is that the best base stealers are caught too much. The game sets out to generate a certain number of attempts and a certain number of successful steals. The overall numbers are always spot on, but the AI won't run its crappy stealers. That leaves a lot of caught stealings without a home... the game has to assign those somewhere, so the guys who are running, the top stealers, end up getting caught too much.

Assuming FHM's engine has some things in common with OOTP, I could see this excess of low efficiency shots in FHM causing similar ripple effects throughout the simulation.

-The game sees the expected number of shots but not the expected number of goals.

-The game has to find those goals somewhere, and it makes sense to me that it would disproportionately find them among the top forwards.

-FHM's simulation being so point shot heavy might also account for the missing assists for top forwards (see my previous post).

The simulation is complex and obviously more questions come to mind (how do the excess shots affect goalies? what about shot blocking? are top forwards just feasting on the bottom pairing dmen who are playing 2 mins too much per game?)... but maybe this explains the high goal scoring numbers?

[renojedi]

Great insights that make a lot of sense. Hopefully Jeff and Co. are listening. I'd love to see this tweaked soon. Just finished the 23-24 season and McDavid had 79 goals. I had MacKinnon on my team this year and he scored 62 in 68 games on the back of a few 4 goal games. He would have been close too.

[crusadertsar]

I really don't get it, you guys don't want to see high scoring in this game? Why?
To me it's the opposite when I see high scoring from my own team then its a sign that I am doing something right tactically. I really hope that the devs don't nerf this.

[jumpSeat]

Originally Posted by crusadertsar

I really hope that the devs don't nerf this.

Agreed. We can already go into league edit to adjust the goals per game ratio, so a dev nerf seems somewhat redundant. (Better they spend their limited time on something we can't already adjust).

[skunt3m]

To be clear, this is not an overall scoring issue, it's a distribution of scoring issue. Overall goals per game in my 2D simulation of 21-22 were only 1.5% higher than in the real NHL in 18-19 - however the top 10 goalscorers in the league scored about 20% more and the top 25 about 10% more (and 17.4%/16.4% fewer assists) than in the real NHL.

Additionally, the fact that players on AI controlled teams are scoring a goal per game indicates a problem IMO. Yes, I think humans should be rewarded for tactical genius. If it was just human players getting those results, fine - but the AI???

[crusadertsar]

Originally Posted by skunt3m

To be clear, this is not an overall scoring issue, it's a distribution of scoring issue. Overall goals per game in my 2D simulation of 21-22 were only 1.5% higher than in the real NHL in 18-19 - however the top 10 goalscorers in the league scored about 20% more and the top 25 about 10% more (and 17.4%/16.4% fewer assists) than in the real NHL.

Additionally, the fact that players on AI controlled teams are scoring a goal per game indicates a problem IMO. Yes, I think humans should be rewarded for tactical genius. If it was just human players getting those results, fine - but the AI???

Fair enough. I get your point. But I just don't understand the need to imitate the NHL stats exactly. It's a fantasy hockey game after all where the world can diverge from real NHL depending on the kind of choices that user or AI managers make. If the AI put three superstar forwards on same line and they have great chemistry then why shouldn't they score lots?
Or if you fill your team with tough guys and tell them to hit and fight like crazy then that should also be reflected in the game. Even if in real life using enforcers is not really a thing anymore and the game has gotten more mundane.