EHM:EA Attribute testing - Staff

[nino33]

I did some significant testing in July, and I've been posting some results from Test 1 (that went from 2014 to 2030).
There's posted data and related discussion of CA, PA and Attributes for the NHL in this thread http://www.ehmtheblueline.com/forums/vi ... 10&t=14804
And in this thread there's posted data and related discussion of CA, PA and Attributes for Major Junior (WHL/OHL/QMJHL) http://www.ehmtheblueline.com/forums/vi ... 10&t=14811

And now in this thread, a look at the Staff (NHL) numbers...


NHL Staff
702 total Staff, average CA 108.4, range 35-193

• 22 Chairman, average CA 106.6, range 58-162 (12 over 100, 5 over 130)
  21 President, average CA 101.7, range 55-160 (9 over 100, 4 over 130)
  39 Director Of Personnel, average CA 108.2, range 35-179 (24 over 100, 9 over 130)
  6 Managing Director, average CA 84.0, range 60-113 (2 over 100, 0 over 130)
  1 General Manager/Head Coach (136 CA/136 PA)
  29 General Manager, average CA 139.9, range 60-174 (28 over 100, 21 over 130)
  30 Assistant GM, average CA 124.4, range 61-170 (25 over 100, 12 over 130)
  29 Head Coach, average CA 123.3, range 64-173 (25 over 100, 10 over 130)
  126 Assistant Coach, average CA 112.6, range 48-182 (86 over 100, 30 over 130)
  382 Scout, average CA 104.8, range 35-193 (227 over 100, 77 over 130)
  77 Physio, average CA 99.1, range 55-120 (42 over 100)

A look at the individual Attributes of some Staff...
mean = average
median = middle value in numerical list
mode = value/number that appears most often


Head Coach
Adaptability - mean 11.4, median 12.0, mode 12
Ambition - mean 12.6, median 12.0, mode 12
Determination - mean 12.3, median 12.0, mode 11
Loyalty - mean 11.4, median 12.0, mode 12
Pressure - mean 11.6, median 12.0, mode 12 & 13
Professionalism - mean 13.2, median 13.0, mode 11 & 13
Sportsmanship - mean 12.9, median 13.0, mode 14
Temperament - mean 13.2, median 13.0, mode 12, 14 & 15

Attacking - mean 9.6, median 10.0, mode 10
Directness - mean 9.9, median 9.0, mode 8 & 12
Free Roles - mean 8.7, median 8.0, mode 6 & 10
Line Matching - mean 9.6, median 11.0, mode 15
Penalty Kill - mean 11.1, median 8.5, mode 11
Power Play - mean 10.5, median 10.0, mode 10
Physical - mean 11.5, median 11.0, mode 9

Tactical Knowledge - mean 11.4, median 12.0, mode 15
Coaching Defensemen - mean 11.6, median 12.0, mode 12
Coaching Forwards - mean 11.2, median 11.0, mode 11
Coaching Goaltenders - mean 10.1, median 10.0, mode 6 & 11
Coaching Technique - mean 11.3, median 11.0, mode 8 & 13
Discipline - mean 12.4, median 13.0, mode 13

Judging Player Ability - mean 12.7, median 13.0, mode 13
Judging Player Potential - mean 12.6, median 12.0, mode 10 & 11
Man Management - mean 11.8, median 12.0, mode 11
Motivating - mean 12.2, median 13.0, mode 11 & 13
Working With Youngsters - mean 13.2, median 14.0, mode 14
Physiotherapy - mean 3.0, median 2.0, mode 1

Business - mean 4.2, median 2.0, mode 1
Interference - mean 3.8, median 2.0, mode 1
Patience - mean 4.0, median 2.0, mode 1
Resources - mean 4.5, median 2.0, mode 1



Assistant Coach
Adaptability - mean 13.0, median 13.0, mode 13 & 14
Ambition - mean 12.7, median 13.0, mode 14
Determination - mean 13.6, median 14.0, mode 14
Loyalty - mean 13.1, median 13.0, mode 16
Pressure - mean 12.2, median 13.0, mode 15
Professionalism - mean 13.4, median 14.0, mode 16
Sportsmanship - mean 12.5, median 13.0, mode 13
Temperament - mean 13.3, median 13.0, mode 12

Attacking - mean 8.9, median 9.0, mode 12
Directness - mean 9.3, median 10.0, mode 12
Free Roles - mean 8.1, median 8.5, mode 9
Line Matching - mean 9.6, median 10.0, mode 14
Penalty Kill - mean 10.1, median 10.0, mode 9
Power Play - mean 9.5, median 10.0, mode 11
Physical - mean 11.8, median 12.0, mode 8

Tactical Knowledge - mean 10.2, median 11.0, mode 10
Coaching Defensemen - mean 11.2, median 11.0, mode 11
Coaching Forwards - mean 11.1, median 14.0, mode 14
Coaching Goaltenders - mean 9.7, median 8.0, mode 6
Coaching Technique - mean 11.1, median 11.0, mode 9 & 11
Discipline - mean 12.1, median 12.0, mode 12

Judging Player Ability - mean 12.7, median 13.0, mode 13
Judging Player Potential - mean 11.8, median 15.0, mode 13
Man Management - mean 11.2, median 11.0, mode 11
Motivating - mean 12.6, median 12.0, mode 10 & 12
Working With Youngsters - mean 11.7, median 12.0, mode 12
Physiotherapy - mean 2.4, median 2.0, mode 1

Business - mean 2.5, median 2.0, mode 1
Interference - mean 2.4, median 2.0, mode 1
Patience - mean 2.5, median 2.0, mode 1
Resources - mean 2.4, median 2.0, mode 1

[visualdarkness]

Generally the Head Coaches are good but some teams seems to hire every mediocre scout I or some other manager fires as their new head coach, dragging the average down bad. I think we can all agree than sub 5 in tactical knowledge should never be HCs in NHL and rarely ones under 10.

[Alessandro]

On the other hand we should check how many staffers actually have attributes filled in the DB, otherwise it's all pretty much random

[nino33]

On the other hand we should check how many staffers actually have attributes filled in the DB, otherwise it's all pretty much random

That would require many, many more days of work.

You'd have to review all 702 staff and compare with the database (determine who's a regen), then determine who had values and who had a "0" and then review all the database Attributes against the Attributes at start-up (because the game changes DB entries at start-up, or at least it does with players), and then look at the actual growth/development/results 16 years later.....I'm back at work now as of today, and my EHM editing/testing time is now severely reduced

The reason I got curious is I noticed krownroyal83 posted mike_B's TBL post about this issue over at SI, noting NHL Coaches have all coaching ratings below 8 (and my testing did not support this claim)

[Alessandro]

On the other hand we should check how many staffers actually have attributes filled in the DB, otherwise it's all pretty much random

That would require many, many more days of work.

yes, I know, the problem is that the players DB is on a much better shape, than the staff one...

[nino33]

On the other hand we should check how many staffers actually have attributes filled in the DB, otherwise it's all pretty much random

That would require many, many more days of work.

yes, I know, the problem is that the players DB is on a much better shape, than the staff one...

Even if it was true (and I don't think it is) I'm not sure how the lack of ratings in the original database would result in "random" results 16 years later

Looking at the testing results it doesn't look random at all to me...every Attribute average (except the Chairman/Physio Attributes) falls between 8-13, which is clear evidence it's not random

FYI - initially the "reason" for this test was the anecdotal evidence that all Coaches had coaching ratings below 9; the testing showed clearly this was not true



If you look at all NHL Head/Assistant Coaches in the 7.4C database
88% have a Coaching G Attribute
85% have a Coaching D Attribute
87% have a Coaching F Attribute
7% of all Head/Assistant Coaches (12 of 159) have a "0" for all three Attributes

the problem is that the players DB is on a much better shape, than the staff one...

EDIT - My testing when EHM:EA was released showed many many player Attributes were changed from the database values (because the mechanics of EHM:EA are significantly different than EHM07), so I'm not sure what you're basing your view on

[visualdarkness]

Can you exlain why the AI got such a fetish for player fired scouts as headcoaches and other guys with no tactical knowledge at all?

All of them are or have been HC in the NHL and generallt there are some teams that hire these type of guys repeatedly and fire each one within a year.







I am the Sharks here and the staff are the ones show above. What is so special with my scouts?

[Manimal]

Can you exlain why the AI got such a fetish for player fired scouts as headcoaches and other guys with no tactical knowledge at all?

The AI probably looks at CA/PA and reputation before attributes

[visualdarkness]

Can you exlain why the AI got such a fetish for player fired scouts as headcoaches and other guys with no tactical knowledge at all?

The AI probably looks at CA/PA and reputation before attributes

Which is just plain broken then. The really low skilled scouts must have a high PA then or the CA system needs fixing too. Can the AI "see" the PA of staff?

Edit: Also why the scouts? Why not go after all the great coaches on the market instead?

[Alessandro]

EDIT - My testing when EHM:EA was released showed many many player Attributes were changed from the database values (because the mechanics of EHM:EA are significantly different than EHM07), so I'm not sure what you're basing your view on

Have a look at the DB yourself. Don't focus on the NHL alone. A lot of players have attributes in, also in the other leagues. A lot of staff members, instead, have all zeros.

[nino33]

Can you exlain why the AI got such a fetish for player fired scouts as headcoaches and other guys with no tactical knowledge at all?

Maybe Tactical Knowledge isn't a "Head Coach only Attribute? The AI knows someone else can do the job

Can you exlain why the AI got such a fetish for player fired scouts as headcoaches and other guys with no tactical knowledge at all?

The AI probably looks at CA/PA and reputation before attributes

Which is just plain broken then. The really low skilled scouts must have a high PA then or the CA system needs fixing too. Can the AI "see" the PA of staff?

Edit: Also why the scouts? Why not go after all the great coaches on the market instead?

Maybe the AI puts the most emphasis on Head Coach only Attributes...so not Coaching G/D/F or Tactics (because Assistant Coaches can do this).....and maybe the ingame management (lines, etc) is based on Judging Ability/Potential?

The Man Management is (I think) a "Head Coach only" Attribute, and it's not bad to very good for every one of your examples above
Working with Youngsters looks pretty OK too in the examples above


FYI - I looked into this issue a little bit earlier today on the SI Forums (including looking at Reputation, CA/PA and more) http://community.sigames.com/showthread ... ad-coaches Maybe most interesting is it seems the game changes Staff Attributes much more dramatically than player Attributes!

[visualdarkness]

The thing is that there are way more "Toby OBrian" and worse in there than the skilled scouts, I just stopped the research as I found "enough" scouts. The high level of the scouts posted says more about my level of hiring staff than the AI, the AI doesnt seem to care how skilled the scout I fired is.

I think you give the AI way more slack in this issue than it should get. I mean, even the AI seems to realize how dumb the hirings are as they fire them after a couple of months, but just hire new ones from the same section.

[nino33]

EDIT - My testing when EHM:EA was released showed many many player Attributes were changed from the database values (because the mechanics of EHM:EA are significantly different than EHM07), so I'm not sure what you're basing your view on

Have a look at the DB yourself. Don't focus on the NHL alone. A lot of players have attributes in, also in the other leagues. A lot of staff members, instead, have all zeros.

So? What difference does it make that a lot of non players/staff have "0" for many Attributes?

Have you not noticed the concerns brought forth by others/the reason this thread exists is because of discussion about NHL Staff?

You're talking about stuff that I'm not disputing (many players/staff in lower leagues have "0" for Attributes)
We're not talking about lower players/leagues, and we're not talking about lower League Staff who had "0" for most/all Attributes

[CeeBee]

maybe assistant coaches and scouts will sign for a lot less money than head coaches so with the financial system being kinda screwed up maybe the AI is going the cheap route.

[nino33]

I think you give the AI way more slack in this issue than it should get.

Sorry, not sure what you're talking about
I actually believe that the CA range for Staff should be the same as the League (so essentially no NHL Staff below a 120 CA) & I believe there should be a "League minimum" for many/most Attributes (so such low values wouldn't exist if I had my way)


The reason I posted/started this thread is to respond to the post here at TBL and at SI that after playing awhile ALL coaching Attributes for ALL Head Coaches in the NHL are under 9.....so I looked at a save I had and saw that the average was above 8. Then I responded to Alessandro's view that the cause of the NHL "hiring issue" connects to Staff having "0" for Attributes (I don't see that myself)

[Alessandro]

EDIT - My testing when EHM:EA was released showed many many player Attributes were changed from the database values (because the mechanics of EHM:EA are significantly different than EHM07), so I'm not sure what you're basing your view on

Have a look at the DB yourself. Don't focus on the NHL alone. A lot of players have attributes in, also in the other leagues. A lot of staff members, instead, have all zeros.

So? What difference does it make that a lot of non players/staff have "0" for many Attributes?

Well, I think (just think, not stating that it is a fact), that if players/staff don't have attributes set, it's much more likely that we get strange results.

[nino33]

EDIT - My testing when EHM:EA was released showed many many player Attributes were changed from the database values (because the mechanics of EHM:EA are significantly different than EHM07), so I'm not sure what you're basing your view on

Have a look at the DB yourself. Don't focus on the NHL alone. A lot of players have attributes in, also in the other leagues. A lot of staff members, instead, have all zeros.

So? What difference does it make that a lot of non players/staff have "0" for many Attributes?

Well, I think (just think, not stating that it is a fact), that if players/staff don't have attributes set, it's much more likely that we get strange results.

I agree with that thinking, but nothing in my research showed the Staff being talked about had "0" for Attributes.....look at the SI thread and you'll see these aren't Staff with "0" as Attributes (and at first glance it doesn't look like Reputation plays much of a part if any either)

[archibalduk]

I've snipped the latter set of posts in this thread as it had gotten too heated - sorry if it doesn't read quite right or appears out of context.

Let's have a fresh start with this thread.

[visualdarkness]

This is probably the worst coach attribute-wise. His coaching record in the OHL is pretty good though, but I still wonder how much this handicaps the AI teams.

[Asher413]

IMO- we're scratching the surface on a big issue. The AI doesn't seem to do a good job hiring coaches (at least, from what I think coaches should be), no argument from me there. Things I don't have time to look at for a while but potentially could be related: Has anyone looked at the AI choices for GM? It's quite possible the coach hiring is working well because it's terrible GM's making the hires (or low patience/loyalty/judgement... something). I also feel like the AI puts too much pressure on immediate results on GMs. I can't say I've done serious research on this, but my anedoctal thoughts are that more GMs by far are fired in one year of EHM NHL than maybe 3? years in the RL NHL. Another potential factor or flaw could be around penalties for not being loyal. Does the game have the complexity that staff don't trust a GM who's gone through 3 coaches in 15 months and then they really are hiring the best staff willing to work for them?

Maybe in a while (weeks) I can look at my longer saves and see what data is there around GM turnover and coach turnover, as well as the 'quality' of the coaches and the GMs. As has been said, maybe the AI is looking for simply a head coach who doesn't even coach in practice, so the attributes that we as a community think make for a strong coach don't have as large of an impact as we believe?

As has been stated, we may also be better off going 20+ years into the future to eliminate the chance that it's the databases that are causing the problem, if there's a lot of newgens it may be a clearer picture of the games ultimate behavior (and maybe be potential feedback for the database folks)

One last random thought- I didn't pay this much attention to 2007 in a long time- did that game also have this issue?

Just spitballin', maybe some of these ideas have merit, or maybe they all are rubbish, but just things to think about (and hopefully if life slows down for me, to look into).

[visualdarkness]

The GM got good atts as far as I can see but I dont know the hidden ones.