Question on random PA

[nino33]

Never used the Assistant (or is the Assistant another name for the DB Editor?)

There are multiple "DB Editors" so I'm not sure what you're referring to; the EHM Assistant is not a DB Editor, but rather a Saved Game Editor - I've never used it to edit anything, but as the EHM Assistant allows the exporting of the CA/PA and all Attributes into an excel spreadsheet, I find it extremely helpful for testing/compiling data (I'm surprised you're not aware of it) http://www.ehmtheblueline.com/forums/vi ... &start=525

And on the notes of prospects always developing, when you take care of your own prospects with min/maxed training schedules I have a 100% success rate in ~2100 hours of getting prospects to their full PA.

It looks like I didn't word that well/sorry for the misunderstanding; the poster I'm referring to wasn't talking about a player developing/reaching their PA, but rather that PA was always at the top of the range (there wasn't variability...so for example a player with a -14 would always get a PA close to 160)

goalies with less than 170 PA are essentially useless in the NHL as anything but backups

I don't think this is true myself...

By that standard (170+ PA) there's only twenty such goalies possible in the entire database...ten with a set PA, six with a -15 and four with a -9
To me this actually makes sense, as IMO few goaltenders IRL nowadays are anything special/significantly above their peers (the difference between a Good goalie and a Top goalie is minimal at best...there simply aren't that many Top goalies); I've thought before that I'd like to see a consistency rating that went year-to-year as the best goalies IRL nowadays are rarely consistently the best year-after-year, but as of right now we don't have that option in EHM

There is evidence that people can win the Cup without such a goalie (TBL Challenges, the "How's it going in EHM 1? What team are you?" thread, etc); I don't think there needs to be more 170+ PA goalies myself...



In reading your large post here and your other large posts today I do think that you write as if you have knowledge/evidence of the game's coding and Riz's reasoning for things that I haven't seen/heard myself (I don't agree with everything you post; I'm not looking to debate the details as I just don't have the time, and we've had the discussion before)

[Shindigs]

Never used the Assistant (or is the Assistant another name for the DB Editor?)

There are multiple "DB Editors" so I'm not sure what you're referring to; the EHM Assistant is not a DB Editor, but rather a Saved Game Editor - I've never used it to edit anything, but as the EHM Assistant allows the exporting of the CA/PA and all Attributes into an excel spreadsheet, I find it extremely helpful for testing/compiling data (I'm surprised you're not aware of it) http://www.ehmtheblueline.com/forums/vi ... &start=525

I always just used Archibalduk's. Didn't bother looking for other ones.

And on the notes of prospects always developing, when you take care of your own prospects with min/maxed training schedules I have a 100% success rate in ~2100 hours of getting prospects to their full PA.

It looks like I didn't word that well/sorry for the misunderstanding; the poster I'm referring to wasn't talking about a player developing/reaching their PA, but rather that PA was always at the top of the range (there wasn't variability...so for example a player with a -14 would always get a PA close to 160)

Oh, I see. Well there seems to be a slight tilt towards higher results, but it's clearly a fair degree of randomization. Need much larger sample sizes to know for sure I guess. Did you look at the PA spread for players in -8 and -9 when you did that huge 20 year test?

goalies with less than 170 PA are essentially useless in the NHL as anything but backups

I don't think this is true myself...

By that standard (170+ PA) there's only twenty such goalies possible in the entire database...ten with a set PA, six with a -15 and four with a -9
To me this actually makes sense, as IMO few goaltenders IRL nowadays are anything special/significantly above their peers (the difference between a Good goalie and a Top goalie is minimal at best...there simply aren't that many Top goalies); I've thought before that I'd like to see a consistency rating that went year-to-year as the best goalies IRL nowadays are rarely consistently the best year-after-year, but as of right now we don't have that option in EHM

There is evidence that people can win the Cup without such a goalie (TBL Challenges, the "How's it going in EHM 1? What team are you?" thread, etc); I don't think there needs to be more 170+ PA goalies myself...

I'm not saying you can't win with a bad goalie, you can win despite having a bad goalie. It's just that goalies that you'd expect to be in the .900 to .910 range IRL often end up with results in the .880 to .890 range in EHM. I did a look into goalies to improve my scouting of them a while back. And it is theoretically possible to get a Vezina challening goalie with 158 CA, but the odds of that happening are so slim. A goalie with 170+ doesn't need to be "perfect" because he has enough total points to get most attributes through brute force, rather than perfect attribute distribution. But you can easily win the cup a with sub .900 goalie if you use custom tactics, when you use AI coaching 100% it gets very dicey though. So I should have clarified my statement related to not watching the 2D engine and letting the AI control all tactical decisions. But when the already drafted goalie prospects in the (now old) 15-16 db have a shot at becoming a franchise goalie, and the draftable ones don't. That means any team that starts the game with no good goalie prospects have a very frustrating situation on their hands, unless you just steal their goalie once they dump them due to poor AI. But that just means you won't ever draft goalies and only trade for them, which isn't as rewarding as having a "homegrown" goalie. Also you can get goalies with 170+ PA that won't cut it as a starter either, there are so many stars that have to align to create a good goalie in EHM and even one attribute being a few points short can really hurt that goalie's ability to put up consistent results for you.

I track lifetime shooting% against all goalies in the NHL in my current save, at first I was just trial and erroring to find the best shot targeting for each goalie. But after a while I started to be able to read goalies fairly well just looking at their attributes. Juuse Saros is a good example, he's got well over 170 PA in my save (by eye, haven't looked him up) but we still have him at sub .900 against us. The reason is that his reflexes are 1-2 points below the bare minimum for a proper starter, his agility is also a few points off and his mentals are also a bit lacking. But 1-2 agi and reflexes under the minimums means he will be bad at either Low or 5-hole, guaranteed. And he is, we use Low on him every game and always beat Nashville as a result. The AI probably doesn't use shot targetting, maybe it does. But it would surprise me. So they will just shoot wherever on your goalie, and the more holes he has from those tiny 1-2 missing attributes, the more possible holes that those shots will "accidentally" target, and the key to reliable success isn't a goalie that has .930 on the season but is super inconsistent. A super-consistent .905 that sits in that ballpark every single game is actually better. Because as long as your goalie doesn't single-handedly lose you the game, your players should be able to win the game (about 90% of the time). Even goalies with 17+ Consistency aren't that reliable though, so having a 170+ PA goalie gives you margin for error, when he pulls out a Rating 6-7 game, you can still win. When that 160 PA goalie has a 6-7 Rating game you just get so shafted you simply cannot overcome the deficit. And I loathe losing to being goalied, hence why I find lower PA goalies pointless. I'd never waste a 1st rounder on getting a goalie that can't get more than 160 PA, when I could use it on a 160 PA Dman/forward. The opportunity cost simply isn't ever worth it.

To me a 141 PA winger can be on your first line and score 80 points a season. But a 141 PA goalie can't be good no matter what you do, he can be good enough for your team. But not good. There is just no margin for error at all when it comes to goalie attributes in this game. Like Saros against us, his PA says "good goalie", but being a measly 1-2 agi and reflexes off means he's not even good enough to be an IRL starter against us. This makes granularity when balancing goalies really really hard, and it makes it so the perfect storm required to create a newgen/regen goalie means you could go an outrageous amount of time without a single starter quality goalie hitting the NHL. Even high 1st round goalies will very rarely cut it as backups when compared to the current goalies in the NHL at the start of a save.

In reading your large post here and your other large posts today I do think that you write as if you have knowledge/evidence of the game's coding and Riz's reasoning for things that I just don't have (I don't agree with everything you post; I'm not looking to debate the details as I just don't have the time, and we've had the discussion before)

That's perfectly fine. You don't have to agree. We do approach the game from very different angles with different goals (yours being making the db as good as it can be, mine being to win as much as I possibly can). And for me a big part of that is making assumptions based on what I see and then having those assumptions as working theories until evidence that directly oppose them arise, at which point I need to revise my theories (this happens a lot with trying to pair scout snippets to actual attribute values). But trying to become good at a game through trial and error with an open mindset is in my findings a heck of a lot more enjoyable and efficient than trying to 100% scientifically prove all aspects (since a lot of those aspects will never apply to any decisions I need to make in-game) before implementing them to my playstyle. The game is unrealistic enough that trying to play it like real hockey only leads to me getting annoyed, so I play it as a game instead. Just trying to set silly records in my saves. Because I find that somewhat fun, despite all the flaws this game has. Which funnily enough are a lot of the same issues the AI in EAs NHL games have as well based on the "Be a Pro" series I'm watching online.

[nino33]

Never used the Assistant (or is the Assistant another name for the DB Editor?)

There are multiple "DB Editors" so I'm not sure what you're referring to; the EHM Assistant is not a DB Editor, but rather a Saved Game Editor - I've never used it to edit anything, but as the EHM Assistant allows the exporting of the CA/PA and all Attributes into an excel spreadsheet, I find it extremely helpful for testing/compiling data (I'm surprised you're not aware of it) http://www.ehmtheblueline.com/forums/vi ... &start=525

I always just used Archibalduk's. Didn't bother looking for other ones.

The advantage has been the exportability of all the data into excel, which allows for much easier data compilation/review

And on the notes of prospects always developing, when you take care of your own prospects with min/maxed training schedules I have a 100% success rate in ~2100 hours of getting prospects to their full PA.

It looks like I didn't word that well/sorry for the misunderstanding; the poster I'm referring to wasn't talking about a player developing/reaching their PA, but rather that PA was always at the top of the range (there wasn't variability...so for example a player with a -14 would always get a PA close to 160)

Oh, I see. Well there seems to be a slight tilt towards higher results, but it's clearly a fair degree of randomization. Need much larger sample sizes to know for sure I guess. Did you look at the PA spread for players in -8 and -9 when you did that huge 20 year test?

Not sure how you see a slight tilt to higher results...all eight players average slightly below the the middle of the range, and none of them surpassed even the mid point of the range more than half the time (so I'd say definitely not a tilt to higher results)

No, I didn't look at the PA spread in previous testing

Juuse Saros is a good example, he's got well over 170 PA in my save (by eye, haven't looked him up)

Did you edit him? His PA is -8 (130-160)

[nino33]

We do approach the game from very different angles with different goals (yours being making the db as good as it can be, mine being to win as much as I possibly can). And for me a big part of that is making assumptions based on what I see and then having those assumptions as working theories until evidence that directly oppose them arise, at which point I need to revise my theories (this happens a lot with trying to pair scout snippets to actual attribute values). But trying to become good at a game through trial and error with an open mindset is in my findings a heck of a lot more enjoyable and efficient than trying to 100% scientifically prove all aspects (since a lot of those aspects will never apply to any decisions I need to make in-game) before implementing them to my playstyle. The game is unrealistic enough that trying to play it like real hockey only leads to me getting annoyed, so I play it as a game instead. Just trying to set silly records in my saves. Because I find that somewhat fun, despite all the flaws this game has. Which funnily enough are a lot of the same issues the AI in EAs NHL games have as well based on the "Be a Pro" series I'm watching online.

I wouldn't say my goal is to make the database as good as possible...I'd say my goal is to provide correct, validated information/feedback to both other users and to Riz (with as little opinion, conjecture, and/or hyperbole as possible...preferably none); my goal is a better game, a more realistic game (and I know I'm not alone when I say when I play the game I try to play it realistically like real hockey, and doing so doesn't annoy me)

[Shindigs]

Shindigs wrote:
We do approach the game from very different angles with different goals (yours being making the db as good as it can be, mine being to win as much as I possibly can). And for me a big part of that is making assumptions based on what I see and then having those assumptions as working theories until evidence that directly oppose them arise, at which point I need to revise my theories (this happens a lot with trying to pair scout snippets to actual attribute values). But trying to become good at a game through trial and error with an open mindset is in my findings a heck of a lot more enjoyable and efficient than trying to 100% scientifically prove all aspects (since a lot of those aspects will never apply to any decisions I need to make in-game) before implementing them to my playstyle. The game is unrealistic enough that trying to play it like real hockey only leads to me getting annoyed, so I play it as a game instead. Just trying to set silly records in my saves. Because I find that somewhat fun, despite all the flaws this game has. Which funnily enough are a lot of the same issues the AI in EAs NHL games have as well based on the "Be a Pro" series I'm watching online.
I wouldn't say my goal is to make the database as good as possible...I'd say my goal is to provide correct, validated information/feedback to both other users and to Riz (with as little opinion, conjecture, and/or hyperbole as possible...preferably none); my goal is a better game, a more realistic game (and I know I'm not alone when I say when I play the game I try to play it realistically like real hockey, and doing so doesn't annoy me)

When you see the AI trade their future away time after time and letting their best goalie prospects go to the KHL while playing goalies not even fit for the AHL over them, as well as countless other issues. There is 0 immersion left for me at that point. Also I've played it "realistically" for so many hours at this point I'm done with that experience. I'm trying to go for the perfect season (100% wins) at this point since that's a more fun challenge, winning is automatic at this point in my save (President's and Cup) so going for that gives no sense of achievement what-so-ever.

I'd say my goal is to provide correct, validated information/feedback

And the point I'm making is that taking the approach to add nothing to your way of playing that hasn't been 100% proven, rather than testing ideas other players and yourself have and then keeping what works and dumping what doesn't work, is too slow and leaves you behind the curve in any competitive game. If you spent the time that proving something you already know 100% with statistics takes, on actually playing the game instead. You would improve much faster, that is my point. I know you have another mindset and other views, I'm just doing what I did to become top 20 in a game played by 11M at the time, and it works for me in EHM as well. So as much as your don't like the approach, it works in practice. Which is good enough for me.

Not sure how you see a slight tilt to higher results...all eight players average slightly below the the middle of the range, and none of them surpassed even the mid point of the range more than half the time (so I'd say definitely not a tilt to higher results)

No, I didn't look at the PA spread in previous testing

I wasn't referring to your numbers there, I was referring to what I've been seeing mostly in -15 players in the actual game while playing. My bad for not being clear enough. Regardless both in your numbers here and what I've seen in my NHL based saves the sample size is far too small. There are players I've used enough in Swe-2 to have some pretty good insight into, but even there the numbers for the longest time seemed to be very close. The player in question is the goalie Adam Werner who had a -6, for something like the first 20 times I used him he was never below 100 PA. Then he was in the 90-92 range 10 saves in a row after that. So it was probably just really weird RNG for me. Rather than it being any special distribution. But just because the small ranges(-7,-8,-9, etc.) don't seem to have it, that doesn't necessarily mean the big ranges(-13,-14,-15) don't. Regardless it doesn't really matter because you only have the ranges that you have, so you'd pick the same option in either case. Just curious if you knew which it was.

Did you edit him? His PA is -8 (130-160)

Apparently I did, been changing players as I go along for such a long time that I've apparently forgotten about some of them at this point. My bad, I think he's been -15 for all but my first NHL save waaay back though, so the strangeness of him always being in the top of the range still remains. But the sample size still means that just like with Werner it's highly possible that it's just a weird string of results.

[nino33]

I'd say my goal is to provide correct, validated information/feedback

And the point I'm making is that taking the approach to add nothing to your way of playing that hasn't been 100% proven, rather than testing ideas other players and yourself have and then keeping what works and dumping what doesn't work, is too slow and leaves you behind the curve in any competitive game. If you spent the time that proving something you already know 100% with statistics takes, on actually playing the game instead. You would improve much faster, that is my point. I know you have another mindset and other views, I'm just doing what I did to become top 20 in a game played by 11M at the time, and it works for me in EHM as well. So as much as your don't like the approach, it works in practice. Which is good enough for me.

I was answering a question, not talking about a way of playing; this thread isn't about a way of playing

I don't care that you were "top 20 in a game of 11 million" (I don't believe trying to exploit the game is the norm at all, especially amongst those that post on TBL); I think much closer to "competition" regarding EHM is the TBL Challenges or even online play


I don't want conflict, but if I was just discovering this game and reading about it (like I was in 2010) I wouldn't want there to be any confusion about someone posting who's goal is to exploit the game but sounding like they're explaining how the game actually works/how the developer coded the game/what data actually shows


I'm done with this now