Scoring, PIM and Fighting Levels - Feedback Thread

[archibalduk]

Version 6.0 of the TBL Rosters introduces a fairly significant change to the way we rate players in the database: We now have reduced scoring, penalty and fighting levels.

As most of us are aware, the scoring and penalty levels in EHM 2007 are too high. A few years ago, Bob McGoo released a Realism Patch. This much improved the scoring and penalty levels in EHM. Over the summer, we have been trying out a refined version of Bob McGoo's patch to see if we could achieve similar results but without having to make quite so many extensive changes.

The result is that we think we have achieved a similarly realistic scoring and penalty (and fighting) levels. We have done this by using a 3 - 10 range for Consistency of forwards whilst using the usual full range for goalies and defencemen (although in reality most goalies and defencemen fall within the 10 - 18 range). This has helped reduce scoring levels. Similarly, setting every player in the DB to a 20 Temperament rating has reduced the penalty levels in the game. However, fighting levels still remain realistic.

As this is a new change, it needs much more testing. Most testing has concentrated solely on the NHL and in relation to points, PIM and fighting stats. From our experience of the Realism Patch, the reduced levels also benefit most other North American and European leagues - so hopefully this will be a positive change for all leagues. In the event that these changes are more detrimental than beneficial, we can easily restore the original Consistency and Temperament attributes because we have a backup of them.

We really need everybody's feedback on how they find the scoring, penalty and fighting levels in all leagues (including the NHL).

[CJ]

As I have this in the 98 DB also, I'm gathering data of different season to see how "realistic" the scoring/penalties are. It will take some time though.

[Chopper]

I definitely find the scoring more realistic with 6.0. I see a lot more 2-1, 3-2, 1-0 games than I used to. Individual players aren't racking up goals and points like they used to. I feel like I'm seeing fewer power plays but haven't paid close enough attention - the next time I fire it up I will keep track.

In general though, it feels like NHL teams playing NHL games – good stuff.

[RomaGoth]

I am not sure where to post this, I looked at the other feedback threads but those seem to be aimed at specific players/positions.

I simmed a full season twice now using the updated 6.1 rosters (I did not make any changes to it) and the numbers look somewhat similar as prior versions. There were 12 players with 100 points or more and the top 50 players nearly averaged a point per game. For some reason, Patrick Marleau is a nearly 100 point player again (he had 96 points in 81 games). Duncan Keith had 84 points to lead all d-men.

This appears to be a work in progress, and I would love to see more realistic numbers across the board. Great work guys!

[RomaGoth]

Looks like in 6.1 the scoring is still a little off. Zdeno Chara, at age 37, set a career high with 77 points. Duncan Keith had 84 points to lead all d-men. There were also 12 players with over 100 points and another 8 players with over 90 points.

[batdad]

Good to see defensemen getting some points LOL.

I am not sure we will ever get scoring toned down to exactly what it should be, and there will be variances from game to game and user to user, depending on how the user plays and how the random things are generated. However....I know there has been some improvement as I have seen it in my games and other users reporting some improvement as well.

On track...on the right track...

[jesterx7769]

It seems like defense is all about their forwards and offense. The game gives defenseman a lot more assists than they would have in real life it seems, even if overall goals per game has been turned down.

[batdad]

Yeah sometimes it does give more assists, but for the most part...the top level dmen have a hard time reaching the point levels they should have. Read Karlsson etc.

[philou21]

Funny enough I like seeing numbers like this because it reminds me more of what hockey used to be than what it is today.

[archibalduk]

I simmed a full season twice now using the updated 6.1 rosters (I did not make any changes to it) and the numbers look somewhat similar as prior versions. There were 12 players with 100 points or more and the top 50 players nearly averaged a point per game. For some reason, Patrick Marleau is a nearly 100 point player again (he had 96 points in 81 games). Duncan Keith had 84 points to lead all d-men.

I'm really surprised by that. I wasn't getting anything near that high with my v6.0 tests. I was of course getting variation between sims, but not the extent of having 12 players with 100+ points. I don't think I ever saw more than around 4 or 5 players at 100+ and usually it was closer to 3 players. I haven't tested v6.1, but it shouldn't be much different because we haven't touched Consistency. I'll run some sims this weekend if I can.

Can you go into the Team Stats screen at the end of season 1 and post the how many goals the top scoring team scored and how many the lowest scoring team scored? I'd be interested to hear what range you're getting. For example, in one sim I did with v6.0 the highest scoring team scored 282 and the lowest scoring team scored 168 - and they were the sort of figures I was getting with my sims.

[philou21]

Oops maybe I shouldn't have raised the consistency of some players then in my DB. But I mostly did it to some I think will be really good in the futur. Anyway, its not that bad.

[archibalduk]

Oops maybe I shouldn't have raised the consistency of some players then in my DB. But I mostly did it to some I think will be really good in the futur. Anyway, its not that bad.

Yeah this is the problem with trying to fix the scoring levels by bending the rules for Consistency. It's going to cause some confusion! It's frustrating because all it would take is a quick patch from SI to fix the levels completely!

[jesterx7769]

With all of the talented utility programs that have been made and updating is there no way to patch the game/do anything with the game code or mechanics itself? I take it there you have been unable to get in touch with/work with some of the original devs from back in the day?

[batdad]

Yeah unfortunately with the hard coding archi and others have played around with as much as they can....and the game tends to become unstable the more you do. So...we have to have workarounds to get the game into a more realistic situation. SI still holds the rights to the game, they are not releasing their code stuff...and the gang (Riz, Graeme et al) that worked on the game for them...are either busy working other games or are on to new things (Jeff R with OOTP,s FHM).

So formally there is nothing that can be done. Informally the boys have tried to work on things, and as I said it has not always been effective. The CSD patch is one example of something that has been successful for many.

Anyway...let us get this back on topic. The coding of the game does not allow for us to work within it to change scoring, fighting and PIM levels. SO we work by adjusting things like consistency, aggression, etc in order to do so.


Also...it is all done basically by trial and error. Try a theory, see if it works.

[RomaGoth]

I simmed a full season twice now using the updated 6.1 rosters (I did not make any changes to it) and the numbers look somewhat similar as prior versions. There were 12 players with 100 points or more and the top 50 players nearly averaged a point per game. For some reason, Patrick Marleau is a nearly 100 point player again (he had 96 points in 81 games). Duncan Keith had 84 points to lead all d-men.

I'm really surprised by that. I wasn't getting anything near that high with my v6.0 tests. I was of course getting variation between sims, but not the extent of having 12 players with 100+ points. I don't think I ever saw more than around 4 or 5 players at 100+ and usually it was closer to 3 players. I haven't tested v6.1, but it shouldn't be much different because we haven't touched Consistency. I'll run some sims this weekend if I can.

Can you go into the Team Stats screen at the end of season 1 and post the how many goals the top scoring team scored and how many the lowest scoring team scored? I'd be interested to hear what range you're getting. For example, in one sim I did with v6.0 the highest scoring team scored 282 and the lowest scoring team scored 168 - and they were the sort of figures I was getting with my sims.

I ran a second sim this morning, and this time I made some changes to offensive attributes and consistency to many of the players I saw with 100+ points and d-men with more points than they probably should have. I found that lowering the consistency in these players really does help, especially with d-men. For example, in my first sim Zdeno Chara had 77 points (about a point per game). In my second sim, he only had 39 points (about a point every 2 games). I toned down his consistency and some offensive attributes (passing, stickhandling, vision, getting open, deking, shot) to more realistic levels.

I will post the team totals here in a few minutes, but from what I saw they actually looked pretty realistic.

[RomaGoth]

Ok, here is what I have from 2 sims using the 6.1 rosters:

Sim #1

2013-14 Goals For

Chicago 279 (Most)
Florida 176 (Fewest)

2013-14 Goals Against

N.Y. Rangers 160 (Fewest)
N.Y. Islanders 285 (Most)

Sim #2

2013-14 Goals For

Chicago 281 (Most)
Florida 162 (Fewest)

2013-14 Goals Against

L.A. Kings 175 (Fewest)
Toronto 283 (Most)

So far, I have come to the conclusion that the Florida Panthers can't score.

[RomaGoth]

Sim #3

2013-14 Goals For

Chicago 279 (Most)
Florida 187 (Fewest)

2013-14 Goals Against

N.Y. Rangers 165 (Fewest)
Edmonton 286 (Most)

Florida still can't score, Chicago leads the league in goals for...again. Seems pretty consistent as far as team goals for/against is concerned. Unfortunately, I had 14 players with 100+ points this time around. Some of it can be attributed to linemates (Hossa/Kane/Toews in Chicago). Consistency still is an issue and I believe that lowering it for all players except goalies may reflect a more accurate NHL.

[RomaGoth]

Ok, last one.

Sim #4

2013-14 Goals For

Chicago 279 (Most)
Florida 161 (Fewest)

2013-14 Goals Against

N.Y. Rangers 161 (Fewest)
N.Y. Islanders 297 (Most)

There is a consistent theme here with goals for and goals against. Chicago always leads in goals for and the Rangers seem to lead in goals against. The team statistics as a whole seem to fairly consistent and for the most part, realistic. 279 goals in an 82 game season is an average of 3.40 goals per game, which is rather high, but I can live with it, especially since it is several dozen fewer than the original database.

I again had 12 players with over 100 points, and there seems to be no way around this. The disturbing part of this is that Ryan Suter appears to be god-like, with 107 points in this last sim (he was averaging around a point per game in the other sims). Not sure why this is except it clearly has something to do with playing with Zach Parise and Dany Heatley and Mikku Koivu. Not sure what to do about it, since I already adjusted his ratings to lower numbers across the board.

I appreciate any thoughts others have on this, including any results from simming and if they are similar to mine.

Hope this helps the dev team here at TBL, let me know if you need anything and keep up the great database work!

[philou21]

279 goals in an 82 game season is an average of 3.40 goals per game, which is rather high, but I can live with it

I don't know where's the problem there honestly, it's perfect IMO. I know playing EHM isn't like watching hockey but I don't want the game to end up too realistic so the scores are as boring as the real NHL. 12 players with 100 is maybe high but 7-8 would be okay for me.

[archibalduk]

Hope this helps the dev team here at TBL, let me know if you need anything and keep up the great database work!

Very useful indeed! Thank you very much!

The team statistics as a whole seem to fairly consistent and for the most part, realistic. 279 goals in an 82 game season is an average of 3.40 goals per game, which is rather high, but I can live with it, especially since it is several dozen fewer than the original database.

IMO 3.40 is about right for the top-scoring team in the league. Compare this with the top GF in real life for the past 5 seasons (including the current season so far):

2013/14 (Chicago): 177 GF in 49 GP = 3.61 gpg
2012/13 (Pittsburgh): 165 GF in 48 GP = 3.44 gpg
2011/12 (Pittsburgh): 282 GF in 82 GP = 3.44 gpg
2010/11 (Vancouver): 262 GF in 82 GP = 3.20 gpg
2009/10 (Washington): 318 GF in 82 GP = 3.88 gpg

The average gpg from those five seasons is 3.51 gp (1204 GF / 343 GP).

Looking at your sims, the lowest scoring teams (i.e. Florida!) averaged 2.09 gpg (686 GF / 328 GP). Compared with the lowest GF in real life:

2013/14 (Buffalo): 83 GF in 46 GP = 1.80 gpg
2012/13 (Nashville): 111 GF in 48 GP = 2.31 gpg
2011/12 (Minnesota): 177 GF in 82 GP = 2.16 gpg
2010/11 (New Jersey): 174 GF in 82 GP = 2.12 gpg
2009/10 (Calgary): 204 GF in 82 GP = 2.49 gpg

The average gpg from those five seasons is 2.20 gp (749 GF / 340 GP).

I think the scoring is pretty much spot on based on the above.

From what you're saying, it seems the top scorers are scoring too much and, presumably, the lowest scorers aren't scoring enough. I wonder if this relates to the fact that the players are rated top-heavy right now (i.e. too many players rated highly and not enough towards the average/poor end). I'll be interested to see how things are once Nino, CJ and Manimal have evened out the CAs a little (although I'm not sure it'll have that much effect as it's the attributes and not the CA that ultimately control a player's performance).

Alternatively, maybe the range of Consistency could be narrowed for the forwards. I.e. rather than using a range of 3-10, maybe using something like 5-10 instead.

I ran a second sim this morning, and this time I made some changes to offensive attributes and consistency to many of the players I saw with 100+ points and d-men with more points than they probably should have. I found that lowering the consistency in these players really does help, especially with d-men. For example, in my first sim Zdeno Chara had 77 points (about a point per game). In my second sim, he only had 39 points (about a point every 2 games). I toned down his consistency and some offensive attributes (passing, stickhandling, vision, getting open, deking, shot) to more realistic levels.

Interesting... I would have thought reducing the Consistency of d-men would result in higher scoring levels. I guess not!! It almost sounds like Consistency really only affects offence (and goalies) rather than defence - which isn't all that surprising given how offence is given greater value in the game than defence. I wonder if we should be using something like a 3-10 range for all skaters.

One final thought - Something we have to keep in mind is how reducing the scoring levels affects leagues other than the NHL. I remember from bobmcgoo's testing with his patch that it was pretty beneficial for various Euro leagues such as the SHL and KHL. When I was testing the UK, I found scoring levels were now much lower than real life.

[philou21]

So after my first season with the 6.1 I ended up with 7 players over 100 points. If Landeskog wasn't injured for me he would've reached 100 pts, finished at 99 lol 17 players did 90+ pts. Ovechkin finished first with 128 pts and 71 goals!

Despite the boost I give to Subban and Karlsson in my DB they didn't do that well in terms of pts. Subban finished with 43 and Karlsson with a small 24 pts. Leader in pts was Bouwmeester with 58 and Alzner (!) with 48.

I don't know if I raise Karlsson weight, like alot, if he's going to be better. Putting him around 200 pounds should help I guess even if it's not close to reality.

[Fredlig]

Is there any easy way to change the consistency attributes back to where they were before? For personal use.

[archibalduk]

Is there any easy way to change the consistency attributes back to where they were before? For personal use.

Yes! Download this file: http://sdrv.ms/1kYSvGx

Then load your DB in the Updater, click on the Import button and then select the appropriate xml file (see note below) - and then save the DB.

Note: I've included in the file above two xml files. It should be obvious from the filename, but one has the original Consistency ratings and the other has the original Temperament ratings. So you can choose to restore both sets of ratings or just one of them.

[archibalduk]

Given the improvements with EHM:EA, various bugs such as the high scoring levels and the overemphasis on bigger players are effectively resolved. Thus from TBL Rosters v7.3 onwards, we have restored the full range of attributes (e.g. we will no longer restrict Consistency for forwards to 1 - 10).