Issues with player aging, goalie development and team spending

[Rand]

See this post (#35 in this thread) for the most recent data!

The data in the original post in this thread is outdated, and no longer relevant.





I've been doing some long term sims out to the 2040's to see how the player generation and development hold up and I think there are some areas of concern that should be looked at.

First and least important aging and decline:
I find that most hall of fame talent plays until their early 40's, with the odd player retiring into their late 40's.

In game results:
Current NHL: 3 players 40+. (Highest, 43)
10 players 40+ in 2035. (Highest 44)
9 players 40+ in 2040 (Highest 47)
13 players 40+ in 2045 (Highest 43)

It's not a huge problem but worth noting I think.

Second goaltender development:
Goaltenders definitely are not developing as well as forwards or defencemen, and it leads huge lack of talent at the position long term.

In game results:
Default database at the start of the game-
7 goalies of 4.5 Star or better. 29 of 4.5 Stars or better.
2035 sim-
0 goalies of 4.5 Star or better, 10 goalies of 3 Star or better.
2040 sim-
2 goalies of 4.5 Star or better , 11 goalies of 3 Star or better.
2045 sim-
1 goalie of 4.5 Star or better, 8 goalies of 3 star or better.

Fairly easy to see a dramatic decline in the quality of goaltending. In the 3 example years, all of them combined didn't amount to the same quality as in the starting database.

This does not seem to be an issue with forwards or defencemen, so the player development logic seems quite good.
At no point did forward talent vary by more then 5%! compared to the default starting database. Which is incredibly close, almost unrealistically consistent.
Defence varied from around 5-20% more talent then the starting database, consistently more but certainly well within a plausible range I'd think. So other positions are doing fantastic, really really good in my opinion.

And finally, player salaries and team spending:

With the latest patch it upped the salaries players were receiving to fix the salary depreciation that would occur long term.
There is still a problem though, but it's not so much with individual salaries I think... the number of players getting paid 7M+ or 10M+ in all tested years was about 20% fewer then reality. So a little low, but no big deal.

Except the number of players getting paid 5M+ ranged from 25% (36 players in a 2035 sim, compared to 132 in the starting database) the number in the current NHL, to about 44% as many at most. (64 players in a 2040 sim)
Which didn't seem to make sense since the high end salaries are reasonably close.

Until I checked team payrolls.... and discovered a large majority of teams are extremely reluctant to spend money.

Examples of this:
In the starting database-
All 30 teams are spending within 20M of the cap. (Starting db, simmed forward to opening day to get the most accurate season figures I could)
25 teams are spending within 10M of the cap.

Compare this to future years
2035 sim-
6 teams are spending within 10M of the cap.
And 9 teams are within 20M of the cap.

2040 Sim-
Only 2 teams were spending within 10M of the cap
and a better but still low 14 teams were within 20M of the cap

2045 sim-
Just 1 team was within 10M of the cap
Only 6 teams were within 20M of the cap.

Over the tested years most teams were around 30M below the salary cap. Which is way below the salary floor in real life. In the 3 years combined, only 9 teams came within 10M of the salary cap. That number is 25 in the starting database.

You've got a pile of teams that have 1 highly paid player, 1-2 around 3-6M and the rest sitting at 550K-2M. Teams in general are just really really reluctant to spend to the lowest amount their even allowed to spend in reality.
So you end up with a weird league where there are very highly paid players, and very low paid players with not much in between.

To some extent I understand this, the AI probably can't handle salary space as well as humans can so it needs a large buffer to ensure it doesn't make really implausible trades or buyouts.
But this seems well overboard that most teams are around 30M below the cap.
Some tweaks to increase teams willingness to spend money needs to be made IMO. At the very least the vast majority of the league shouldn't operate way below the salary floor.

Other then these points I think the sim does extremely well long term in maintaining consistently plausible results, and with some tweaks has the basic for a fantastic simulation of real world hockey.

All simulations were done with FHM3 version 3.0.56.
Sandbox mode, GM unemployed and let the game sim ahead without changing any settings from their default.

[JeffR]

Is this a custom sandbox game, or the full start (and, if so, which leagues active)?

[Rand]

Originally Posted by JeffR

Is this a custom sandbox game, or the full start (and, if so, which leagues active)?

Sandbox, all leagues deselected so it would sim faster.
No customization's made to the league or rule alterations. Everything was left at default.
GM was left unemployed the entire time.

If it makes any difference I'll do a long term sim with all leagues enabled after the full release version comes out.
No point starting one now though, as I probably wouldn't have simmed enough years by the time a new version is released to have any firm conclusions, and updating mid way through a sim may throw things off as I have no idea what changes may be included in it.

[JeffR]

Yeah, wait for the release build tomorrow before you do that. Having all leagues deselected will have some influence on overall development.

[no way]

Originally Posted by JeffR

Yeah, wait for the release build tomorrow before you do that. Having all leagues deselected will have some influence on overall development.

I can say developement seems ok with sandbox and all leagues selected, more or less same amount of ability as in curent NHL even after 16 years, jsut the Goaltenders seems to be lacking in talent. Seems to me, that with custom game the lack of depth might be afecting the developement a bit.

As a side note in July of 2033 I found exactly 3x 39 years old and 6x 38 years old and not a single player 40 years old (5 of them goaltenders). Since it's start of season those will be 1 year older at the end of season

[Wan6Tent]

Have you tested this in the final game yet Rand??

[Rand]

Originally Posted by Wan6Tent

Have you tested this in the final game yet Rand??

I am but it takes awhile.
Simming with all leagues enabled is much slower, and I want to sim out from the beginning of the game to the mid 2040's a couple times to get a large sample size to judge from.
Plus I have to check all the results myself, and count the numbers etc.

Maybe Tuesday night, more likely Wednesday I'll have results

[Rand]

Going to do a few more days of testing on this because I'm getting some really weird results that don't make a lot of sense.

Aging seems to be slightly impacted by number of leagues simulated, development doesn't seem impacted much at all.... but of all things, it seems like spending levels are directly correlated to leagues simulated.
Which at least from the outside doesn't seem to make any sense, so I'm going to run a few more sims to make sure this is accurate.

[Nino33]

Thanks for your efforts Rand!

[MXD]

Originally Posted by Rand

Going to do a few more days of testing on this because I'm getting some really weird results that don't make a lot of sense.

Aging seems to be slightly impacted by number of leagues simulated, development doesn't seem impacted much at all.... but of all things, it seems like spending levels are directly correlated to leagues simulated.
Which at least from the outside doesn't seem to make any sense, so I'm going to run a few more sims to make sure this is accurate.

...Actually, it does make A LOT of sense if there is a few high-level leagues. There's shouldn't be a crazy correlation in spending levels, but there should be one nonetheless.

[Nino33]

Since it's a simulation of the IRL Leagues, I don't understand why there'd be any correlation between number of Leagues simulated and spending levels.....I'd think the spending levels would be exact (when possible) or reasonable approximations of the IRL spending levels

[Sebastian Palkowski]

I´m working hard on the next update, so I have no time to go to deep into this interesting thread but if you talk about spendings (as in hoh much player ask for/get and teams therefor spend): there was a rather crucial bug which now is fixed. Player significantly asked for less then they should. So at least for this area you should wait for the next update to look into it.

Anyways, thank you for the great reports, something like this is really helpful to further improve the game!

[Rand]

Will be done running the tests for this build by tomorrow.
Currently have 5 sims out to 2045 with all leagues enabled, and 5 with all leagues disabled besides NHL/AHL.
Just running finishing running with NHL/AHL/Canada/Sweden for a partial leagues test.

Sorry it's taking so long but I ran into some oddities so I wanted to be extra thorough to ensure the results were valid.


I'll start again whenever the next build is released since that apparently has some fixes for the financial side.

[Rand]

This is now a little outdated, since a new version was released today but my results from version 3.0.59. AKA the original release version.
All tests done using sandbox mode, GM unemployed and no settings adjusted besides number of leagues enabled.

At the beginning of the game, with the default database, simmed forward to opening day-
Defencemen rated 3 Star or above: 61
Defencemen rated 4.5 Star or above: 9

Forwards rated 3 Star or above: 178
Forwards rated 4.5 Star or above: 20

Goalies rated 3 Star or above: 29
Goalies rated 4.5 Star or above: 7

Team Cap space
Teams with 10M or less in cap space: 25
Teams with 20M or less in cap space: 30

So the above are the baseline conditions with the starting database. I think that hopefully things would tend to stay within +/- ~25% of these figures aside from the occasional outliers. That would imo leave a fairly large amount of variance but still ensure a league that remained recognizable and plausible.


I ran five separate simulations out to 2045, and checked the database in 2035/2040/2045 in each simulation.

The following numbers are an average of the results from the three time periods checked on each of the five simulations rounded to the nearest even number.

Example- 5 / 7 / 4 / 6 / 9
Would mean that in the first simulation the average number from the three years of 2035/2040/2045 was 5. The second simulation the average number was 7 etc etc.
Please inquire if this isn't clear, I'm doing it this way as otherwise this post would be about 10x as long as it is, and it's already going to be quite long.


Now with all leagues disabled except the AHL/NHL:

Defencemen rated 3 Star or above: 73 / 57 / 68 / 71 / 62
Defencemen rated 4.5 Star or above: 9 / 6 / 11 / 8 / 9

Forwards rated 3 Star or above: 168 / 175 / 172 / 181 / 169
Forwards rated 4.5 Star or above: 15 / 13 / 14 / 17 / 12

Goalies rated 3 Star or above: 11 / 9 / 10 / 6 / 8
Goalies rated 4.5 Star or above: 1 / 0 / 2 / 1 / 0

Team Cap space
Teams with 10M or less in cap space: 3 / 3 / 7 / 4 / 3
Teams with 20M or less in cap space: 9 / 11 / 17 / 10 / 14

So, clearly we can see teams aren't spending anywhere even close to the rate they should. But that was a known bug, so I'll leave that aside.

Forwards and Defence look extremely good. Defence are consistently a tad higher then the starting database, but nothing notable. Forwards are consistently almost exactly the same as in the starting db. These looks terrific.

But goalies... clearly their not developing properly. Superstar goaltenders are extremely rare. You could go a decade and not see as many as there are in the starting database in one year
The number of 3 Star goaltenders are better, but we're still looking at about a third the number as it starts with.


Now, with all leagues enabled.
Again, I ran five separate simulations out to 2045, and checked the database in 2035/2040/2045 in each simulation.

Defencemen rated 3Star or above: 82 / 89 / 85 / 96 / 80
Defencemen rated 4.5 Star or above: 13 / 20 / 15 / 20 / 12

Forwards rated 3 Star or above: 245 / 312 / 261 / 232 / 249
Forwards rated 4.5 Star or above: 32 / 48 / 37 / 30 / 33

Goalies rated 3 Star or above: 24 / 27 / 25 / 23 / 24
Goalies rated 4.5 Star or above: 7 / 5 / 9 / 6 / 7

Team Cap space
Teams with 10M or less in cap space:16 / 5 / 7 / 4 / 4
Teams with 20M or less in cap space: 27 / 22 / 17 / 13 / 19

Good news... the goalies now look good! Consistently a tiny bit lower then at the beginning, but nothing to worry about here. Entirely plausible.
Bad news.... forwards now have major problems. Forwards develop way too well, elite forwards are far more common, ranging from 50% more to over double the number.
Defencemen have problems too, we're typically looking at around 50% more defencemen at a given quality.

As for spending, well... that's a known bug presumably fixed in the most recent update.

I also did 5 simulations to 245 with NHL/AHL/Canada/Sweden... I won't list those results, as they were very close to the results from that with all leagues enabled so they don't bring in any new information.

With the earlier beta versions I was seeing issues with player aging, but that no longer appears to be an issue at all with the release version in any sim, so I won't present the numbers here.


Basic synopsis:
Goaltenders- Develop extremely poorly with only NHL/AHL enabled. Develop properly with the full set of leagues enabled.
Defence- Looks okay with only NHL/AHL enabled. Develop too well with all leagues enabled.
Forwards- Looks fantastic with only NHL/AHL enabled. Develop far far too well with all leagues enabled.

Clearly the number of leagues enabled substantially impacts how well players develop, all positions develop far more talent with it enabled. Ideally this would be pretty similar.
Unfortunately, whether your running with all leagues disabled or all leagues enabled major issues start popping up long term as the league grows increasingly further out of balance. There isn't really any way to get a long term simulation that keeps a realistic level of talent.

[Rand]

Simulations were run on two different computers, pretty consistently since release but I can't image that would make any difference. (I'm thankful the game is largely single threaded so several cores are left free, so it doesn't have any real impact on actually using the computers whilst the simulations are running!)

If anyone wants more information/copies of the save files/exact numbers from each year in any simulation ask me. I have it all in an Excel spreadsheet


And now I'm going to go update to the latest version of FHM and start simulating again. Eventually I will have time to actually play the game and not just endlessly test it's long term development

Incidentally, all of this isn't meant to say the game sucks or it's unplayable. I like the game, so I want to see it in the best condition it can. Hence why I'm running these tests. Hopefully some of this proves useful to the development team in tracking down and resolving the issues with player development.

[Nino33]

Originally Posted by Rand

Incidentally, all of this isn't meant to say the game sucks or it's unplayable. I like the game, so I want to see it in the best condition it can. Hence why I'm running these tests. Hopefully some of this proves useful to the development team in tracking down and resolving the issues with player development.

I certainly appreciate your efforts! and I'm sure others do too


I've done the same kinda thing with EHM and I've found many things that "require improvement/addressing" (and I have the same motivation...I like the game/want to see it in the best condition it can be)

[Rand]

Originally Posted by Nino33

I certainly appreciate your efforts! and I'm sure others do too


I've done the same kinda thing with EHM and I've found many things that "require improvement/addressing" (and I have the same motivation...I like the game/want to see it in the best condition it can be)

I just thought it should be noted so nobody misunderstands.

I used to beta test the EHM for the first three versions before it got cancelled way back when, it's great to see some of the old EHM fanbase still around for FHM!
It was a long time without a hockey management game.

[Wan6Tent]

Do you know if this is fixed in thg update?

[pens66]

forget this... wrong thread.

[Rand]

Originally Posted by Wan6Tent

Do you know if this is fixed in thg update?

Patch notes mention fixes related to the financial side, but nothing with respect to player development.
I'm running sims again though, so I'll post results for the latest version when done.