the "underdatabase" - what it is and why IMO it's critical

[nino33]

THE UNDERDATABASE - WHAT IT IS AND WHY IT"S NEEDED


I believe strongly that for optimized realism all EHM1 rosters need a user created "underdatabase"

The "underdatabase" is all the Players and Staff the game creates at startup.
If you don't add players/staff, the game adds 8,000 players and 4,000 staff
If you add players/staff, the game adds 23,000 players and 11,000 staff
(with a retro database you can select "Disable" extra players/staff)

It's all these extra players and staff that the game creates that often damages the efforts of researchers to make a quality database, because left on it's own the game does NOT produce realistic extra players/staff (and does NOT "fill in" blank Attributes realistically for Staff)



Here's some examples from a very recent test I did with TBL 8.2...

Regarding Scouts.....there are 115 NHL Scouts whose Judging Ability/Potential together averages 18+
and there are 59 free agent scouts whose Judging Ability/Potential averages 18+
So when a user starts a game, and they have (on average) less than 4 scouts on their NHL team that has an 18+ average rating, there are 59 potential fake free agents who they can go after! Among the non-NHL and non-Free Agent Scouts (so the rest of the world's leagues) there are 56 out of 1272 scouts that have an 18+ average rating - of the 55 NHL Scouts that average of 19-20, none of them stand out to me (I'd think maybe an appropriate number of such NHL scouts might be a handful or two in the entire league IMO)


The NHL GMs current have a 15.3 average rating (of Judging Ability + Judging Potential)
Only 3 average 19-20 (compared to 55 Scouts) and only 3 average 18 (compared to 60 Scouts)
14 NHL GMs have an average rating below 15
This very likely connects to poor GM trading/team management


The hidden Staff Attribute "Physical" apparently indicates how much a Coach likes to use physical players/tactics.....With the 1998 DB it caught my attention because Scotty Bowman had an 18, which is double or more what he should've been IMO (the Red Wings were not a physical team, Bowman NEVER promoted physical tactics and always an offensive/skill team with a few good grinders (from 1976-1998).....I look at the 8.2 Head Coaches as they appear ingame (I added no players/staff, then used the EHM Assistant to export everything, so no variables/everything's an actual value).....15 of the NHL Head Coaches ended up with a Physical rating of 14-17, and only 3 were below 10 (the 30 Head Coaches average 13 for Physical.....for me 13-15 would be the highest I'd go in 99% of cases) - and with this I wonder if this is affecting the (excess) number of power plays that are occurring in EHM


Free Roles (higher number = fewer defensive responsibilities) doesn't work out well either...Lindy Ruff with a 15 is #1 and John Tortorella is #4 with a 12
Meanwhile guys like Darryl Sutter with a 6 and Ken Hitchcock and Mike Babcock (each with a 4) are all apparently pretty much unconcerned with defensive play! HaHa


The hidden coaching Att Attacking (higher = attacking/offensive style) has similar undersireable results.....The top Attacking coaches (15+) include John Tortorella, Todd McLellan, Bob Hartley and Mike Babcock.....While Patrick Roy is a 5 and Michel Therrien is a 4.....Tampa Bay's Jon Cooper has an 8




It only took me 2-3 days to get the 98 DB to it's "incomplete form" that still has the noticeable improvements/realism.......as Beukeboom said

This DB is the best thing since sliced bread, its not just names and Dates but a nearly perfectly balanced new world that gives EHm a whole new Level of fun and realism. This DB is so good it actually makes the game in itself work better by doing the little things like level of talent with players and especially staff on a perfectly balanced and competitve level . Gone are the times when you had a perfect staff memebr in every category, while still theres enough around to not starve ut have to compete. Opposed to recent complaints the scoring especially for the timeline also seems perfect. Everything just seems so balanced and real it gav eme back joy for the game again, and when i comare it to my save with the modern DB its so much more fun because the Players seem more unique in their skillset. Ist not perfect but darn close to it.

And this was only with a few days to get what little I could done!



Here's how I did it
1. created a blank database, selected every "add players/staff" that was available
2. used the EHM Assistant to export this entire database into spreadsheets
3. deleted from the blank database all players/staff that were "not needed" because they're populated by real players/staff from the TBL database (tedious at the end, but easy at first.....delete all NHL, all AHL, all WHL, all KHL, etc)
4. then you have your "underdatabase" & editing is done to this underdatabase (like ensuring no fake free agents are world class talent)
6. of course, with a modern database you'd keep Riz's high PA players age 9-14 at startup, BUT you can review them to ensure they're realistic/appropriate/acceptable (no more PAs in the 190s with a CA under 10)


With the 98 DB I inputted the bare minimum Atts (like just Judging for Scouts/GMs or just Coaching G/D/F for Assistant Coaches), and then started a game with that DB and used the Assistant to spit out what the game "filled in" - and then I used that to "review and make adjustments" (I wish I had spent 2-3 weeks on it instead of 2-3 days)



IMO a lot of issues with poor quality staff and staff realism are database related.
The game seems to do much better if it's not allowed to be as random/nonsensical as it is when left fill in so many critical Attributes.
IMO there is no doubt that a decent "underdatabase" can go along way to increasing realism and increasing enjoyment when playing

[Beukeboom]

Very intersting read , and as i was mentioned here i also want to say that in the 98 database with staff the Overall quality still seems very balanced after years...but there is another problem...while i can easily add ex top coaches as assistants to get a very unrealisitc top coaching staff, the AI seems to be left with the pants rest. To be concrete: i have Keenan , Bowman etc with all coache sint the very high Greens, while other Teams mostly even have had coaches with Attribute in orange and red acros the board. This kind of destroys the overall balance and challenge a lot.
With players the DB seems great still and competiveness a lot greater , after years you of course get an an advantage , especially because you have an easy time knowing the good players in the draft by their real name..but this isnt a DB issue. I realized there are a lot less top players in the game and market for them is a lot more competitve than in the modern DB where after some time you had top players you couldnt fit on 4th line anymore because of depth of talent.
Woudl be really awesoem if those inner workings of the udnerdatabase could be used even more for a realism , in 98 DB it already seems to work very good for beginnings and the game feels so much better , you fell so much more in a realstic environment and dont have the Feeling the game "screws up" the overall balance with way too much talent over time. This is especially interesting for me as all those real palyers in the DB would hava ehad the potential to even add to the overall talent a lot too, in the modenr DB you always had the Impression every rela life Player is just addiotnal talent to the one created by the game. In 98 i didnt experience this so much and it Looks like theres really a lot of potential in digging deeper into those innner workings.

[nino33]

Very intersting read , and as i was mentioned here i also want to say that in the 98 database with staff the Overall quality still seems very balanced after years...but there is another problem...while i can easily add ex top coaches as assistants to get a very unrealisitc top coaching staff, the AI seems to be left with the pants rest. To be concrete: i have Keenan , Bowman etc with all coache sint the very high Greens, while other Teams mostly even have had coaches with Attribute in orange and red acros the board. This kind of destroys the overall balance and challenge a lot.

Yeah, it was a "quick & dirty" done over a couple/few days instead of a few weeks (with LOTS being left "as is" due to lack of time), so I think a lot more can still be done to improve things with staff (which I think would likely significantly address the problem of poor AI Staff)

[marksbros6]

So Nino; when you're talking about the 'under-database' you're referring to the players/staff that the game creates on start-up?

And if I'm reading right the proposed solution is to export these players/staff from a save-game and edit them to be 'realistic' and import them into the main database.

You're also talking about existing coaches attributes being 'off' (which is close to my heart with the UK non-playing staff anyway) would this not be a seperate issue?

I can see how this could give us all a more 'stable' database with fewer 'outlier' players (like the 14 year-olds who regularly turn up on the top-line of my EPIHL games) and the scouts/staff I usually recruit to replace the 'rubbish' British ones I start with.

[nino33]

So Nino; when you're talking about the 'under-database' you're referring to the players/staff that the game creates on start-up?

And if I'm reading right the proposed solution is to export these players/staff from a save-game and edit them to be 'realistic' and import them into the main database.

Yes

You're also talking about existing coaches attributes being 'off' (which is close to my heart with the UK non-playing staff anyway) would this not be a separate issue?

I suppose you could say that, but to me they're not; they're both an issue of "needs editing" to me
And because the fake staff are often better than the IRL staff, the issues are intertwined (and there are examples of both overrating and underrating among both the fake and the IRL, so again it's one editing task from one editing view...at least it is for me)

As I tend to see the database as a whole, and work on and think about and edit all aspects of a database, to me I see it as a whole


I just took a look now with 8.2...
- the starting (Full) database has 43000 players and 11000 Staff
- if you choose to add nothing the game says it's adding 8000 players and 4000 staff
- if you choose Add Key Staff the number of Staff rises from 4000 to 11000
- if you choose Add Players to Playable Teams the game adds 0 Players (with a Full Database)
- if you choose Add Players to National Team Pools the game adds 1000 players
- if you choose Add Extra Junior Players the game adds 14000 players
Seems to me "by the data shown" that TBL 8.2 is OK for Players, but not OK for Staff (or depth/Junior Players)


The other aspect is the many unrated Attributes that Staff who are in the database have, and the poor (random?) and unrealistic job the game does at filling these in.....from the Patience Attributes (is that why there's so many firings?), to the Physical Attributes (is that why there's to many penalties in the NHL?), to the Free Roles//Attacking Attributes (John Tortorella is not an "all offensive, forget the defense" style coach as he was rated!)...there's LOTS that I think is potentially improvable/fixable via database improvement

But it takes an enormous amount of time to test/validate and then retest; it takes some effort just to understand things...
And there's barely enough researchers to get just the list of players names and basic ratings in for the playable leagues! let alone do more!
And few people but me seem to have the time/passion for the issue that I do (I'd say no one else does, because no one else is crazy like me HaHa)

So it's not a quick and easy thing to address, not at all...but it can be done! I'm doing it for the 1974 DB now


P.S. I confirmed with Riz that the NHL Team Budgets are to low in the TBL Rosters (pretty much all NHL Budgets should be well above the Salary Cap, but only 6 of 30 are above the Cap at all); I suspect this is contributing significantly to the financial/contract issues being reported

[Primis]

And because the fake staff are often better than the IRL staff, the issues are intertwined (and there are examples of both overrating and underrating among both the fake and the IRL, so again it's one editing task from one editing view...at least it is for me)

...SNIP...

The other aspect is the many unrated Attributes that Staff who are in the database have, and the poor (random?) and unrealistic job the game does at filling these in.....from the Patience Attributes (is that why there's so many firings?), to the Physical Attributes (is that why there's to many penalties in the NHL?), to the Free Roles//Attacking Attributes (John Tortorella is not an "all offensive, forget the defense" style coach as he was rated!)...there's LOTS that I think is potentially improvable/fixable via database improvement

I'll just speak from my own experience. I've been playing EHM since the freeware days. I admittedly barely played NHL:EHM because it was too "unfinished" for me, and the interface was just bad. I played some 2005 though and warmed up to it, and then a lot of 2007, and now a lot of 1. To me there's been one, single problem with databases since the freeware days even, and it's this: too much parity. The good players are not good enough compared to the other players, and the lower-end ones are often overrated still. The same can be said for the Non-Playing Staff IMO.

Bear with me, this is a bit of a long explanation, and there is a point to covering all of this.

When Jan Smetana (pretty sure it was him) did the original freeware EHM rosters, that parity was not present. Lemieux, Gretzky, etc (that's how far back this was: those guys were still active playing at the tail-end of their careers and therefore in the DB's) were rated as actual Superstars. This meant their attributes exceeded 100 on the normal 1-100 scale of the freeware (Superstars could have attributes up to 125). At some point, the league's top players (even the Ovechkins and Crosbys) were then brought out of Superstar range down into the 90's max instead, and at that same time the 4th liners and fringe AHLers were brought up from 60's Overall generally into the 70's Overall (a huge mistake IMO). And despite calls over the years for that to be reverted, it never did, and even the newest Hokay freeware rosters are in this mold of parity, and it's why people have often struggled to get 50- or 60-goal scorers in the freeware even when they stack teams. The game-generated players in the various freeware versions had really major problems though (1.15 did best, except for it lacked generating top-end talent.,1.16 did worse all-around, and 1.18 had the Enforcer Bug that got player ceilings out of whack and tried turning everyone into Enforcers even though some other generating issues were fixed). To me, in the freeware the Staff impact was minimal anyways in the NHL (it actually seemed greater in the Minors)

Move ahead now to the newest EHM game DB's. Somehow.... IMO, this sense of keeping that parity carried at least partially over. The best players never really are head-and-shoulders the best like they should be. On up through 2007, this GREATLY affected the "regens" IMO, and thus the playing talent in later years. (Obviously the regen situation is now different in EHM 1).

While a lot of attention on the DB has been directed towards tweaking players, I think the game-gen staff now seems so "good" because their effect in-game is too great on too many players, not just the top-end ones. I could make an argument that bringing back that "talent gap" in the modern rosters would go a long ways towards straightening out how the coaches and GM's work. Scouts and physios are a different topic entirely (although a genuine argument is to be made that given modern scouting, the gap between the best and next tier of scouts is minimal, and everyone has access to far more scouting data and tools).

There's just been a huge amount of Attribute Creep over the years and even then across game versions. People rarely complain to nerf a player, but they come out in droves to complain to boost one. And it adds up over time. I think it's also maybe just human nature that it's easier to boost someone than it is to reduce someone.

Anyhow, I get the impression that some of all this is why the Staff seems to have so many "good" options in the DB. There's a bit of chicken-or-the-egg at work here. Their impact should be minimal anyways, and I think the player side of DB is making it seem more impactful.

I'm seeing that even with good Staff with the new retro DB, their impact is blunted a bit by a player DB that has a much greater "gap" and disparity in it talent-wise. So the players part of the DB being toned down makes the coaches more realistic IMO, even with the good coaches that are more like the modern good ones. So I could make an argument that the player side of the DB needs to be tweaked across the board first, and then see if the coaches/staff portion are all that out of line (though I suspect they could use a small tweak down still). Ultimately, I think the player attributes are what drive things, and the coach/staff ones are complementary.

But it takes an enormous amount of time to test/validate and then retest; it takes some effort just to understand things...
And there's barely enough researchers to get just the list of players names and basic ratings in for the playable leagues! let alone do more!
And few people but me seem to have the time/passion for the issue that I do (I'd say no one else does, because no one else is crazy like me HaHa)

This is a big reason why I try to stay out of ratings/attributes talk mostly. It's a ton of work,I can't do it myself, and I don't want to upset people who are doing the work.


I think a minor tweak maybe on Riz's end, plus then bringing the game DB to have greater attribute disparity on the player side... between the two you'd solve a lot of discrepancies. That's tougher though because I think EHM is now hard-coded more with these greater-parity DB's in mind even.

Or actually maybe even better: maybe a slider setting or two should be added to the game to address this, and give some flexibility to the user. So you could make the game itself generate more/better Staff, or less, depending on the DB at the time. This would also then give say retro or fictional DB's even more flexibility, as you could get the game to generate Staff to better match the DB you start the game with.

[A9L3E]

An intriguing text, Primis. Regarding your suggestion about a slider, (based on what little experience I have about programming) I think 'twould be easier for Riz to make a configuration file where the people generation can be adjusted rather than a slider in preferences, similar to the other database configuration files. This would also have the advantage that the casual player does not need to bother finding out what the settings mean, as they would be already optimised by the roster researchers.

[nino33]

To me there's been one, single problem with databases since the freeware days even, and it's this: too much parity. The good players are not good enough compared to the other players, and the lower-end ones are often overrated still.

Ever try the 1974 database?
Because what you're saying here is exactly what I've been saying since I discovered EHM in 2010

I made substantial changes to the 1974 database to address this, and "the community" did not like it (in fact, IIRC, I was asked to edit/improve players before the database was allowed to be part of a Challenge)

I'm working on the 1974 database again! and this time I'm going with what I think is best, and am not concerned that the community may find it "jarring"

And I'm going to "fix" the changes I made to better suit the community (and fix the changes to try and deal with things in EHM07 that are no longer a problem in EHM1 + I need to review/correct Player Roles too, as the game gave 190 CA 16 Off Role 16 Def Role Bobby Clarke the role of Grinder!

I found overall the community didn't want a realistic database, they wanted superstars at every position, they wanted the high ratings (and when EHM1 first came out, and players didn't develop "immediately" at 18-19 many people loudly complained!)

The problems last year with extreme offensive or extreme defensive players was connected to bugs with Offensive/Defensive Role (Off/Def Role is something I've long advocated for and still do think is critical, but everyone else last year said it didn't matter/leave it blank - with my work on Major Junior for the 1998 database every player had Offensive/Defensive Role and Player Role filled in).....some preferred to focus on Attributes and let the game determine Player Role (and as the Bobby Clarke example shows, that's not a good idea either!)



I love EHM, and I sure appreciate all the work Riz has put in
I've been a primary defender of him and his lack of time, as well as the lack of TBL researchers!
But I found the end result was I was attacked and poorly treated by many IMO...so no more defending people for me!
It's easier for me not to care/not to get stressed if I'm not trying to convince anyone of anything "for the better of the community" - now I'll work on my own projects, and provide the testing/facts I've always done, and don't feel the (for me natural) need to defend people because "I'm part of the team"



Maybe things (including the community) are different now because of FHM? I mean in a good way, in that since FHM is seemingly trying their best to produce a REALISTIC game based on facts/numbers, maybe the "fun game experience" of EHM isn't good enough anymore (and so maybe people are more inclined to accept some of the changes that have happened and are still needed, even if some in the old guard are against it because they think they know better). I find a lot of people don't want to actually discuss/debate/learn and understand EHM, they just want to repeat their view, regardless of facts/testing/etc

Maybe the key is to realize there's many thousands of people playing, and the loud few are not representative of the masses?



A poor player in the NHL was supposed to be around 100, an average player around 120
The TBL databases have the average close to 140, and that's with them being reduced in recent years (the excess of talent was worse before Manimal came along). The bottom 4th line fringe players have CAs in the high 120s

Given the distance between average and top has been closed substantially, it makes sense that "parity" is the result (and in the NHL that might be kinda correct! HaHa while the top stars are maybe not being well represented, IMO a great deal of NHLers are replaceable with another similar player with little if any impact.....and some believe modern players are "better at everything" and so they think they all deserve 15+ in their Attributes - there are no more true superstars now per capita then there was 40 years ago IMO)

I could make an argument that bringing back that "talent gap" in the modern rosters would go a long ways towards straightening out how the coaches and GM's work.

IMO I don't think you could, because the Staff Attributes are in horrible shape (whether human entered or game generated)

The examples above are just a few of many
Taking a look at the TBL Staff that have a CA now.....there are 234 NHL Staff who have a CA between 34-100 (IMO 120 is the starting point for NHL Staff!), there are 431 Scouts worldwide at start-up that have better Scouting Potential/Ability average than the NHL General Managers average!.....and this is without including all the "superstaff" the game creates at startup!.....I could give many, many more examples

The overall community desire for fact/testing based knowledge, analysis and change is limited & the time to do so for those that believe in it is limited...in my years of working on the TBL Rosters Staff received very little attention (though I did push for it and did what I could)

And the examples about the hidden coaching Attributes are also examples of significant problems that wouldn't be fixed at all by anything being done to the players

Or actually maybe even better: maybe a slider setting or two should be added to the game to address this, and give some flexibility to the user. So you could make the game itself generate more/better Staff, or less, depending on the DB at the time. This would also then give say retro or fictional DB's even more flexibility, as you could get the game to generate Staff to better match the DB you start the game with.

The game is awful at generating Staff now (close to random), so I think that would need to be improved first before adding sliders; making what is simpler and doesn't work now more complicated seems unlikely to result in improvement to me

There's just been a huge amount of Attribute Creep over the years and even then across game versions.

Actually I think it's gone down over the years, but was so ridiculously high in the early years (with all the superstar regens created at startup) it might be hard to notice HaHa

The handful of researchers that I spoke regularly with, and the handful that do almost all the work, are well aware of these issues.....but it takes an enormous amount of time to fix (and there aren't enough people to do the work), and you're almost certainly going to receive significant criticism/flack (I do regularly just because I don't give anecdotal evidence a lot of credence, and prefer actual testing/facts, and people take offense at being asked questions/asked to clarify....and after testing/facts has shown someone to be wrong multiple times it's hard for me to keep listening to their views)


P.S. And there's of course stuff I can't say that would IMO help even more in explaining things.....but I will say this, Scouts should have much lower CAs than GMs, Head Coaches or Coaches (IMO in almost all cases a 140 is "the top end" for the best Scouts, a handful or two at best) - but when the game creates superstar Scouts better than the IRL Scouts there's a problem! thus the need IMO of the underdatabase, because the human efforts to create a "perfect" database will be ruined when the game adds players/staff

[Primis]

I was in the midst of replying to this last night and a storm hit and knocked out our power, and a live wire was on fire in the middle of the street in front of our house. So I'm going to give this another try...

Ever try the 1974 database?
Because what you're saying here is exactly what I've been saying since I discovered EHM in 2010

Yes, and I noticed it has more of the talent gap. So does the 1998 DB. A couple of the other partially-completed retro DB's seem to be the same. Even Chip's 1980 Retro mod for the freeware EHM had that gap included. People apparently realize the game needs it obviously, when it comes to retro DB's of all eras. So I still can't wrap my mind around why people then resist it with the current rosters so much. *shrug*

I found overall the community didn't want a realistic database, they wanted superstars at every position, they wanted the high ratings (and when EHM1 first came out, and players didn't develop "immediately" at 18-19 many people loudly complained!)

Well, there *still* are big problems with player development IMHO. I don't expect players to fully develop in 3 years like they sometimes could in the freeware, but I personally am REALLY tired of having a guy that's basically ready to be out of the AHL, bring him to the NHL and give him the appropriate ice time, and have him not develop at all for a couple years for what seems to be no discernible reason. The player development still seems to be broken as heck to me, at times, and needs to be one of the A #1 things that is fixed/worked on.

And I'm going to be honest: I'm completely sick of reading through the Practice Thread and seeing complex schedules and spreadsheets from people trying to get their guys to develop. If you have to do that kind of manual practice, the game is broken IMO. I personally have no interest in getting into practice micromanagement. If I wanted to deal with that sort of thing, I'd go play EVE Online or something. One of my least favorite parts of EHM freeware online leagues back in the day was needing to do spreadsheets to scout the drafts and player development.

So there are some legitimate complaints there about development.

Maybe things (including the community) are different now because of FHM? I mean in a good way, in that since FHM is seemingly trying their best to produce a REALISTIC game based on facts/numbers, maybe the "fun game experience" of EHM isn't good enough anymore (and so maybe people are more inclined to accept some of the changes that have happened and are still needed, even if some in the old guard are against it because they think they know better). I find a lot of people don't want to actually discuss/debate/learn and understand EHM, they just want to repeat their view, regardless of facts/testing/etc

Maybe the key is to realize there's many thousands of people playing, and the loud few are not representative of the masses?

This is absolutely true now that the game is on Steam. I know some people that play EHM and that never dip their toes into the communities like TBL here. I actually know a guy from Brazil even that's been playing it since 2005 I think. The people that come here say to TBL represent a small percentage of those playing and using the DB's and add-ons. Most won't ever register here to post, even if they come to read things and lurk a bit once or twice.

A poor player in the NHL was supposed to be around 100, an average player around 120
The TBL databases have the average close to 140, and that's with them being reduced in recent years (the excess of talent was worse before Manimal came along). The bottom 4th line fringe players have CAs in the high 120s

Given the distance between average and top has been closed substantially, it makes sense that "parity" is the result (and in the NHL that might be kinda correct! HaHa while the top stars are maybe not being well represented, IMO a great deal of NHLers are replaceable with another similar player with little if any impact.....and some believe modern players are "better at everything" and so they think they all deserve 15+ in their Attributes - there are no more true superstars now per capita then there was 40 years ago IMO)

I would say the Bottom six forwards and bottom 2 or 3 d-men on most teams are all replaceable parts, many of them easily replaceable with AHL guys to be honest.

The danger that comes from deciding a "1st liner gets a CA of x, and a 2nd liner gets y" is that you get teams with weird ideas of line/unit creation. Detroit insists Justin Abdelkader and Darren Helm are Top Six players, when they're both more 3rd liners talent-wise. Or you'd get DET's insistence that Ericsson is a top pairing d-man (that after 2 seasons I think they're finally off of) when he's actually a 3rd pairing guy and nothing more. Abdelkader playing on DET's "top" line doesn't mean he should get a top liner's CA. So then what criteria do you use to decide what line a guy "belongs" on?

There's just been a huge amount of Attribute Creep over the years and even then across game versions.

Actually I think it's gone down over the years, but was so ridiculously high in the early years (with all the superstar regens created at startup) it might be hard to notice HaHa

Well that's the thing. I think in some ways it's gotten better, but I think it's only continued with regard to many of the lower-end NHL players. I don't care that lower-end NHLers now are better than they were 20 years ago, so are the top-end players IMHO. So the gap shouldn't shrink. I want attributes scaled to accurately represent the talent gap in players, not how they compare to players of yesteryear.


I think A9L3E's suggestion of a config file or string in a pre-existing config file is a solid one. It's something then the DB creator has control over, and if the player really doesn't like it, it could still be changed. As you say, nino, there's still work on Riz/SI's end as well in terms of generating, and having it make more sense. There's no doubt. But if you make it modifiable, then you don't have to have the "under"base quite so perfect, it just needs to be a bit better than it currently is. And I'm all for "power to the user". I hate, hate, HATE hard-coded things when to comes to this genre of games. EHM still has too many hard-coded things, even though it's improved from 2007.

I wonder if Riz could "out-source" the staff generation issue somehow.

[m0fownz0r]

P.S. I confirmed with Riz that the NHL Team Budgets are to low in the TBL Rosters (pretty much all NHL Budgets should be well above the Salary Cap, but only 6 of 30 are above the Cap at all); I suspect this is contributing significantly to the financial/contract issues being reported

Could this possibly be edited for the 98/99 DB? Is there I way I can sort this out on my own?

[marksbros6]

The handful of researchers that I spoke regularly with, and the handful that do almost all the work, are well aware of these issues.....but it takes an enormous amount of time to fix (and there aren't enough people to do the work), and you're almost certainly going to receive significant criticism/flack (I do regularly just because I don't give anecdotal evidence a lot of credence, and prefer actual testing/facts, and people take offense at being asked questions/asked to clarify....and after testing/facts has shown someone to be wrong multiple times it's hard for me to keep listening to their views)


P.S. And there's of course stuff I can't say that would IMO help even more in explaining things.....but I will say this, Scouts should have much lower CAs than GMs, Head Coaches or Coaches (IMO in almost all cases a 140 is "the top end" for the best Scouts, a handful or two at best) - but when the game creates superstar Scouts better than the IRL Scouts there's a problem! thus the need IMO of the underdatabase, because the human efforts to create a "perfect" database will be ruined when the game adds players/staff

Some really good points nino. Staff attributes are close to my heart as you know with the British staff not being worth employing. Like I've said before I just usually just employ foreign coaches (Maybe even a few of them are 'newgens' though they don't need 'super' attributes to be better than the British staff!)!

I'm still looking into the British staff myself and working out how to make them 'useful' and give them personality. I've not spent as much time playing NHL, and when I do I usually coach myself and haven't noticed coaches 'playing' their teams differently to in real-life (but then I'm not as familiar with the NHL coaches beyond the Rangers' set-up).

'Filling out' the 'underdatabase' is certainly a worthy approach. Of course you know as well as anyone the difficulties actually keeping the database up-to-date, so it would probably have to be a long-term project if at all.

You've looked at it though; how many players/staff would have to be generated? Could it all be youngsters? A mix of ages? A mix of nationalities? Does position matter at all? Does the number of active leagues have an effect?

If we could say "We need 2000 more American players with a CA/PA of *so much*" and "We need 900 more British players with a CA/PA..." etc then it would make it a bit more manageable. Even better we could look at using real players who are UFA or youth players to broaden the game world (I've been looking at youth rosters for the UK very recently, most of these kids will never go 'pro' even at our level, but if they retire early they'd just keep the re-gen pool topped-up I believe? Don't know whether this would satisfy the needs of the game not to generate it's own players though).

[nino33]

Yes, and I noticed it has more of the talent gap. So does the 1998 DB. A couple of the other partially-completed retro DB's seem to be the same. Even Chip's 1980 Retro mod for the freeware EHM had that gap included. People apparently realize the game needs it obviously, when it comes to retro DB's of all eras. So I still can't wrap my mind around why people then resist it with the current rosters so much. *shrug*

I think it's because the modern player's talents are massively overrated in the minds of many (the rule/style changes and defensive focus means the modern player does very little that resembles the hockey I saw growing up, but they're in position and they dump it in/dump it out and they cycle & IT'S BORING! HaHa Hockey's dying in Canada, less and less participation and less and less watching on TV (and with the cost of playing + the variety of options to watch, I don't see it changing); the "rising Cap" is another joke, as year after year it goes up 2-3% and the players give 15% of their salary back at season's end (hockey's NOT growing/getting better), but I find the vast majority don't understand this at all and think the rising Cap is proof that hockey's growing

Well, there *still* are big problems with player development IMHO.....So there are some legitimate complaints there about development.

Not disagreeing with you! But being me, I need "proof"
I need a hypothesis that can be tested and initial conclusions drawn from, and then more of the same is done/repeated and refined until reasonable conclusions can be drawn...

Alessandro brought this up to me months ago, and I want to look into it, but I'm only one person and I've only got so much time (and I don't believe there's a second person doing what I'm doing, it's just me...and there's so much that needs looking into)

I would say the Bottom six forwards and bottom 2 or 3 d-men on most teams are all replaceable parts, many of them easily replaceable with AHL guys to be honest.

Totally agree!

The danger that comes from deciding a "1st liner gets a CA of x, and a 2nd liner gets y" is that you get teams with weird ideas of line/unit creation.

When I edit I don't say 30 teams x Left Wing = 30 first line left wingers, because I think there's to many teams and the talent is so diluted that many teams have first lines of non-first liners.....but again, many disagree with me because the modern hockey player is so great and the modern game is so great and needs to expand because there's so many great players

When I first discovered EHM I did this and encountered fierce resistance; there are still some that argue things that they're just wrong on, they don't care about testing/facts/etc because "they know" & it has a lot to do with my "defending" or "sensitivity" - I think my hundreds/thousands of hours of testing matters, and I think anecdotal evidence matters little as "proof" of anything, especially if there's anecdotal evidence saying the opposite (which there almost always is HaHa), and most definitely when there's actual testing evidence showing they're wrong


The spreadsheets I see are usually showing people trying to EXPLOIT the game, that's completely different than using spreadsheets to keep track of things IMO


NOTE - Since I'm commenting on research/the NHL I want my opinion to be known/clear that CJ (who does the NHL) is a spectacularly good researcher! And Manimal, Alessandro and Named also.....but those 4 people account for an enormous amount of the research done, and it's simply not enough (there's others too! Swedish/UK groups, and a few that do smaller nations/tasks, but for the hockey world that the vast majority play there's only a few researchers to cover it all)

I think A9L3E's suggestion of a config file or string in a pre-existing config file is a solid one. It's something then the DB creator has control over, and if the player really doesn't like it, it could still be changed. As you say, nino, there's still work on Riz/SI's end as well in terms of generating, and having it make more sense. There's no doubt. But if you make it modifiable, then you don't have to have the "under"base quite so perfect, it just needs to be a bit better than it currently is. And I'm all for "power to the user". I hate, hate, HATE hard-coded things when to comes to this genre of games. EHM still has too many hard-coded things, even though it's improved from 2007.

I wonder if Riz could "out-source" the staff generation issue somehow.

Here's where people get offended by my "defending Riz/SI" because maybe they didn't understand what I was really getting at, so this time I'll be blunt - I don't believe there have been any signs from Riz/SI that such significant efforts are going to happen from their end...we get small, incremental improvements, and that's it!

Big issues like this (and others) need to start with testing, and I suspect Riz frequently doesn't even have the the time for that.
Don't get me wrong, I believe things WILL get addressed, but not quickly (I'd like it if it were different, but I see no evidence thus far...I'd love SI to give Riz a year off from FM, but it seems that's not going to happen (in fact, it seems at least in part the Riz efforts we do get are done on his own personal time!)

What would really help is if the people that say they can do whole seasons in 20 minutes would do the testing [takes me 20 hours.....and now that my laptop's died again (as of yesterday), and for reasons unknown EHM hasn't worked on my wife's computer, I'm down to one computer].....another thing that would help is if a tools or tools were made that spit out the data in helpful ways as it takes me days and days to compile (after taking two weeks to complete the testing), though all praise to Garfild as without his EHM Assistant almost all of my efforts wouldn't be possible

I felt in my short time in the lead role I got an enormous amount of information/clarification from Riz by asking testing based questions, with facts already determined and essentially Riz just needed to clarify/confirm.....I don't see any indication that Riz has the time to tackle the significant issues on his own in a timely fashion; that's why I'm trying to help as much as I do, because I think what I actually do (providing testing results/facts) actually does help a lot...and "FYI to all" that's why I can be in the opinion of some (many?) overly sensitive, but I'm not sure of that and since I have no one to compare myself to I don't really agree (as no one is doing what I'm doing, and "putting themselves out there" like I do, and my efforts help the community as a whole significantly, I think my sensitivity is justified)

And in conclusion, regarding "I hate, hate, HATE hard-coded things when to comes to this genre of games. EHM still has too many hard-coded things, even though it's improved from 2007" - my understanding is it probably has WAY LESS hardcoded things then you think, we just don't have an editor (yet!) that allows us to see this

Part of the "under the hood" changes I was referencing since EHM:EA came out was the Player Role and Key/Essential/Non-Essential/Irrelevant Attribute system (a HUGE change/improvement).....but part of it also is Riz working to make less and less and less hardcoded; I'm positive that right now there are things that are not hardcoded that we don't realize yet simply because we don't yet have an editor that can show us "hey! it's not hardcoded anymore!"

My understanding is we'll be able to edit MUCH MORE in the future!
That's why I still work on retro projects that I'd never release until we can change the number of NHL teams (and because it takes about two years to create a retro database! so if I don't have some "irons in the fire," when the editing ability comes I won't be ready.....I won't be ready anyhow! HaHa but I'll at least have a good start, and be closer to being ready!)

[nino33]

P.S. I confirmed with Riz that the NHL Team Budgets are to low in the TBL Rosters (pretty much all NHL Budgets should be well above the Salary Cap, but only 6 of 30 are above the Cap at all); I suspect this is contributing significantly to the financial/contract issues being reported

Could this possibly be edited for the 98/99 DB? Is there I way I can sort this out on my own?

Yes, first you'd need to figure out what the Team Budgets should be (I'd think you'd need to look at multiple years using a source like capfriendly.com or something similar), and then you would make the edits using an EHM07 version of the database and the Pregame Editor (Clubs/Club Info tabs), and then you would import that edited database into EHM1

FYI to all - any editing of a player using the Pregame Editor erases the player's Player Role

[nino33]

'Filling out' the 'underdatabase' is certainly a worthy approach. Of course you know as well as anyone the difficulties actually keeping the database up-to-date, so it would probably have to be a long-term project if at all.

Actually from start-to-finish "filling out" (creating) the underdatabase doesn't take much time at all; it's the editing of the underdatabase that takes some time

People are noticing a difference with the 1998 DB and I suspect I spent 5-10 hours on it over 2-3 days.....from an editing perspective that's nothing!
I'm a little embarrassed about it's quality now that it's a subject of discussion, as there's LOTS I didn't get too

Ask CJ! If it was left to me I'd never be "done" HaHa

If you know what you're doing in excel, and you know what you're looking to do (you've thought through "appropriate/wanted" ratings) I think a fabulous job could be done in 2-3 weeks (instead of 2-3 days! HaHa)

You've looked at it though; how many players/staff would have to be generated? Could it all be youngsters? A mix of ages? A mix of nationalities? Does position matter at all? Does the number of active leagues have an effect?

EVERYTHING I do is always done the same way - all Leagues Enhanced, Full Database, Maximum Detail (for everything)
I have no knowledge on Quick Sim or smaller database sizes or limiting Leagues

So what you get in terms of numbers is exactly what the game produces (and my way, you get the maximum number of players/staff the game produces).

Since you're starting from a blank database, you delete the fake/created players and staff from the underdatabase in all cases that you have IRL players and staff

So for example, you'd delete 100% of the players/staff generated for NHL teams
If you are going through a League and 1 or 2 teams are "missing staff" (because the team/league websites don't list them) then you can use the fake/created staff; this would be helpful in Major Junior for example, as the WHL/OHL usually list all their Staff and the QMJHL does not

But it's the editing of CA/PA and key Attributes that takes the time, and makes the big difference/improvement

If we could say "We need 2000 more American players with a CA/PA of *so much*" and "We need 900 more British players with a CA/PA..." etc then it would make it a bit more manageable.

I think in terms of databases, and there's no tool that creates just the specifics you're referencing (you get a whole database or nothing HaHa)

Even better we could look at using real players who are UFA or youth players to broaden the game world (I've been looking at youth rosters for the UK very recently, most of these kids will never go 'pro' even at our level, but if they retire early they'd just keep the re-gen pool topped-up I believe? Don't know whether this would satisfy the needs of the game not to generate it's own players though).

Ideally I would assume any researcher working on a League/Nation would include all such players already (if not, then absolutely yes that's how it should be done); I know from my experience I lacked the time to get all the 12-16 year olds in that I wanted (and with all the tasks going on they were pretty low priority)

As you note, they are needed to keep "the re-gen pool topped-up"

With a yearly/modern database though, and the lack of researchers/researching time, I'm not sure there's much value in including very young/low talent IRL players (maybe fake players would be better, because then the same underdatabase could be used the following year? or at least the same underdatabase could be edited a bit and used? without having to worry about some IRL names within)

[m0fownz0r]

P.S. I confirmed with Riz that the NHL Team Budgets are to low in the TBL Rosters (pretty much all NHL Budgets should be well above the Salary Cap, but only 6 of 30 are above the Cap at all); I suspect this is contributing significantly to the financial/contract issues being reported

Could this possibly be edited for the 98/99 DB? Is there I way I can sort this out on my own?

Yes, first you'd need to figure out what the Team Budgets should be (I'd think you'd need to look at multiple years using a source like capfriendly.com or something similar), and then you would make the edits using an EHM07 version of the database and the Pregame Editor (Clubs/Club Info tabs), and then you would import that edited database into EHM1

FYI to all - any editing of a player using the Pregame Editor erases the player's Player Role

Can this be achieved with the EHM07 Save-game Editor as well?

[nino33]

P.S. I confirmed with Riz that the NHL Team Budgets are to low in the TBL Rosters (pretty much all NHL Budgets should be well above the Salary Cap, but only 6 of 30 are above the Cap at all); I suspect this is contributing significantly to the financial/contract issues being reported

Could this possibly be edited for the 98/99 DB? Is there I way I can sort this out on my own?

Yes, first you'd need to figure out what the Team Budgets should be (I'd think you'd need to look at multiple years using a source like capfriendly.com or something similar), and then you would make the edits using an EHM07 version of the database and the Pregame Editor (Clubs/Club Info tabs), and then you would import that edited database into EHM1

FYI to all - any editing of a player using the Pregame Editor erases the player's Player Role

Can this be achieved with the EHM07 Save-game Editor as well?

I'm assuming you're talking about an EHM07 game - sorry, I have no idea, I never used the EHM07 SavedGame Editor
If you're talking about EHM1, then no AFAIK the EHM07 SavedGame Editor is not compatible with EHM1

[Shindigs]

As long as the NHL will at long period of saves be made up of newgens (and it will for about a decade and a half due to the 1.4 player aging "issue) whatever you do to help diversity by fixing player PA in the database won't really matter. Besides 16 offensive role is enough to make a 100 point scorer granted he's got the attribute spread a newgen with that value of a good player role will have. A quick player with as little as 11 wristshot can still be a 30-45 goal scorer in the NHL, meaning that you won't be able to get attributes low enough to create players who are bad enough to have a solid diversity. And even if you do they will be replaced by newgens and regens sooner rather than later. A player with PA in the 150s and 16 offensive role will be so good offensively that a 186 PA 20 offensive role player won't be enough of an upgrade (if any at all) to make up for the larger wage demands such a player would have. The only way to really make superstars "stand out" would be to move essentially any non-superstar player into the sub 140 PA range, but again newgens would undo that mighty fast. Literally any change you can think of that might fix this that isn't Riz changing the formulas for player generation and growth will not work long term. It's simply impossible to fix this without making sure that the newgens also follow whatever changes you make. The same things go for coaches too, since you get retiring players turning to coaches and causing the same issues all over again.

Also fun vaguely related fact, generated players reach peak PA at a younger age than all the premade prospects you deal with at the start of a save. A player with a sizeable enough difference between their CA and PA will take 6 years (split into three 2 year phases) to reach their PA. Premade players for the first draft are 17 or 18, second draft are 16 or 17 and generated players for the following drafts start as 14-16; premade thus can reach their cap at age 23-24 for the first draft, 22-23 for the second draft and generated between 20-22 for all following drafts. For Major Junior reasons the game favors putting high PA regens as 14 year olds meaning that when you draft someone like Erik Karlsson's regen he will be almost PA capped at 19 by the end of his first season. It completely devalues the need for planning for the future since you can get a finished top 4 dman or top 6 forward the same year you draft him essentially every year as long as you know how to scout efficiently. This behavior also means that if you're testing player growth you'll get a lot of incorrect info by looking at the first 6 years with premade prospects as they don't follow the same growth pattern as generated players do.

So if anyone's looking to compile information on growth patterns in prospects in a scientific manner you should certainly sim forward 6 seasons before you start recording to get rid of all the premade players whose growth will not fit "normal" patterns. The fact that I did not know this was a pretty large factor in why figuring out practice schedules was so maddening and took so long, because the first 6 years of any save there's some weird stuff going on with what I like to refer to as "ghost players". Guys who were prospects once that have since either stopped playing or just stopped being relevant so their db entry hasn't been updated or deleted, which means you have guys in their mid 20s who still have the attributes they had in their late teens, and the -14 PA value to go with it. Watching those players grow is so weird, because they are not permitted to gain physicals through PA gain at that age, so every single attribute point they gain go into technicals. But they still have their 6 year growth plan to go through, despite their age. Which means they won't finish growing at 20-22 like a normal generated player; They will finish growing at age 30+ instead, which is honestly borderline more realistic than the super express growth that's the norm in EHM. But it is also entirely unhelpful when trying to figure out how development works in this game.

[nino33]

As long as the NHL will at long period of saves be made up of newgens (and it will for about a decade and a half due to the 1.4 player aging "issue) whatever you do to help diversity by fixing player PA in the database won't really matter.

I have no idea what you're saying here, nor do I know what the 1.4 player aging issue is

they will be replaced by newgens and regens sooner rather than later.

With newgens in EHM, are you talking about those players that can be added by selecting the option to add additional players at startup? Because I never do this, and never edit in a way that considers it (and I'd never recommend using it myself, but "different strokes for different folks")

There's only regens, and based on the starting base the talent pool stays about the same (overall the "new" talent will be the same as the "old" talent); my understanding is essentially anything can change with a regen but not everything always does, and while PA may change (maybe not?) if it changes it doesn't seem to change significantly/noticeably (so lower PA players won't be replaced by significantly higher PA players, or vice versa)




With a retro database, and the EHM Assistant, I think it makes more sense to check your game for the newly created regens once each year, and severely reduce the PA (to below 100) of all those with an NHL quality PA that aren't historical/real players (that's what I intend on doing!)




The starting database determines the worldwide talent level, regens don't really change this
Adding fake players at startup (newgens) for sure changes things!...but I like the databases to be as accurate as reasonably possible with real players, based on real teams/leagues, so I'd never add fake players to a game

And I only add fake players to a database when I can totally control their Attributes...and thus the underdatabase