Question on random PA

[pantsukki]

Many players have their PA set as a random value; what if the game randomizes it to be lower than their CA at startup? Or is it even possible?

[Manimal]

I don't think that can happen but I guess a player would get worse if he has a lower PA.
We never give players lower PA when we research them. If they are old we set PA same as CA at worst.

[Shindigs]

It seems like this can happen with newgens though, I'll rifle through my screenshots and show you what happens when a newgen starts with more CA than their PA allows:


So there, that's what happens if a player has higher CA than PA at launch/creation.

[CJ]

Maybe this should be reported to Riz? If this is the case. Should run it in EHM Assistant to confirm it though to be sure.

[landry2119]

That's something that has bothered me for a while now (I don't post much, but I come here almost everyday for the past 8 years at least... anyways). I'll take Nathan Beaulieu as an example. In the current roster, I think his CA is 132. His PA is -14. Since, at -14, his top PA is 160 and he is already at 132, shouldn't his PA be at -8, considering he has already attained the floor PA of -8 and the top PA (160) is the same for both -8 and -14? I don't know if what I'm saying is clear... There are a few examples in this very specific situation.

[nino33]

Should run it in EHM Assistant to confirm it though to be sure.

I agree


EDIT - I just checked some old EHM Assistant exports I have, one from July after one season, another from August 2025 and another from August 2045 (three different saves) and in none of them was there a player that showed as having a PA below his CA

[Edgars]

It has nothing to do with PA being below CA, because, I don't think it's even possible. The reason behind this is that if a player plays in row reputation league and/or club, while having a high CA (relative to league/club reputation), then his CA may drop significantly (at least that's my theory).
I noticed it after creating and testing Latvian league. I have seen players that started with CA between 50 and 60, but ending up with CA between 10 and 20, in extreme scenarios, in mid-season (December) and maybe even sooner, because I didn't really followed their CA development day by day or month by month.

[nino33]

So there, that's what happens if a player has higher CA than PA at launch/creation.

If you have a save that the EHM Assistant shows the PA being lower than the CA, it'd be helpful if you could report the issue and upload the save on the SI EHM Forum

[nino33]

That's something that has bothered me for a while now (I don't post much, but I come here almost everyday for the past 8 years at least... anyways). I'll take Nathan Beaulieu as an example. In the current roster, I think his CA is 132. His PA is -14. Since, at -14, his top PA is 160 and he is already at 132, shouldn't his PA be at -8, considering he has already attained the floor PA of -8 and the top PA (160) is the same for both -8 and -14? I don't know if what I'm saying is clear... There are a few examples in this very specific situation.

I think this connects to a lack of volunteer researchers (and thus limited available researcher time)...it's never been thought that a PA could generate below a CA, so it seemingly made no difference whether the PA was -14 or -8 (so one less thing for the overworked/overwhelmed researchers to do)

[nino33]

It has nothing to do with PA being below CA, because, I don't think it's even possible. The reason behind this is that if a player plays in row reputation league and/or club, while having a high CA (relative to league/club reputation), then his CA may drop significantly (at least that's my theory).

Based on the Attributes shown, it doesn't even appear to be much of a drop (the difference in visible average Attribute value is 0.381; the top screenshot shows a 156 total, the bottom 148)

[Edgars]

It has nothing to do with PA being below CA, because, I don't think it's even possible. The reason behind this is that if a player plays in row reputation league and/or club, while having a high CA (relative to league/club reputation), then his CA may drop significantly (at least that's my theory).

Based on the Attributes shown, it doesn't even appear to be much of a drop (the difference in visible average Attribute value is 0.381; the top screenshot shows a 156 total, the bottom 148)

Are we looking at the same picture, because I see that every single attribute have droped by 2 or 3 which is a lot (of course if 1-20 scale is used not 1-100), except those that never change anyway (like aggressiveness, determination, flair)

[Manimal]

That's something that has bothered me for a while now (I don't post much, but I come here almost everyday for the past 8 years at least... anyways). I'll take Nathan Beaulieu as an example. In the current roster, I think his CA is 132. His PA is -14. Since, at -14, his top PA is 160 and he is already at 132, shouldn't his PA be at -8, considering he has already attained the floor PA of -8 and the top PA (160) is the same for both -8 and -14? I don't know if what I'm saying is clear... There are a few examples in this very specific situation.

Not necessarily.
With -8 and -14 the top is the same but with a -14 it is less likely that he draws the top rating. He be at 132-160 but the researcher might have wanted to have him weighted toward the lower end but with a slight possibility of the higher end

[nino33]

It has nothing to do with PA being below CA, because, I don't think it's even possible. The reason behind this is that if a player plays in row reputation league and/or club, while having a high CA (relative to league/club reputation), then his CA may drop significantly (at least that's my theory).

Based on the Attributes shown, it doesn't even appear to be much of a drop (the difference in visible average Attribute value is 0.381; the top screenshot shows a 156 total, the bottom 148)

Are we looking at the same picture, because I see that every single attribute have droped by 2 or 3 which is a lot (of course if 1-20 scale is used not 1-100), except those that never change anyway (like aggressiveness, determination, flair)

Oops not sure what my problem was, but I miscounted for sure!

Top shows 208, bottom 148 (assuming I didn't miscount again!); based on visible Attributes that's a CA drop from approximately 100 down to approximately 60


I don't suspect it has anything to do with League/Club Reputation myself (but maybe it does, I don't know).....players CAs do go down, and I suspect it occurs at certain times & the player in the example given was randomly selected to have his CA go down at that time (or one of the times, I'm not sure how often it happens each year) - if Shindigs still has the save from the first screenshot it should be pretty easy to use the EHM Assistant to check the player's CA/PA and see if the PA is actually lower than the CA

[nino33]

With -8 and -14 the top is the same but with a -14 it is less likely that he draws the top rating. He be at 132-160 but the researcher might have wanted to have him weighted toward the lower end but with a slight possibility of the higher end

I did not know this/I've never heard of this before; I thought Animal31 did some testing years ago that showed the likelihood of any specific value was evenly distributed over the potential range (but maybe my memory's as bad as my counting today! HaHa)

For a -14 player to be more likely to draw a lower rating, wouldn't that mean the game would have to potentially give the player a PA lower than their CA sometimes? Seems to me if they both have a 132 CA a -14 or a -8 player both have the same potential of getting a value between 132 and 160 as their "set" PA.....

[Manimal]

With -8 and -14 the top is the same but with a -14 it is less likely that he draws the top rating. He be at 132-160 but the researcher might have wanted to have him weighted toward the lower end but with a slight possibility of the higher end

I did not know this/I've never heard of this before; I thought Animal31 did some testing years ago that showed the likelihood of any specific value was evenly distributed over the potential range (but maybe my memory's as bad as my counting today! HaHa)

For a -14 player to be more likely to draw a lower rating, wouldn't that mean the game would have to potentially give the player a PA lower than their CA sometimes? Seems to me if they both have a 132 CA a -14 or a -8 player both have the same potential of getting a value between 132 and 160 as their "set" PA.....

I've never tested the theory, just assumed it was so.
You make a good point against my theory!

[landry2119]

The way I understand it (and I runned a few tests to confirm it), and I agree with the logic, is that a player can never have a PA under his actual CA. So, to keep Nathan Beaulieu as an example (CA 132, PA -14), if he is getting a random PA anywhere between 90 and 132, it's gonna be set at 132. With this system, he has a 40% chance to see his PA attribute raised from 132. If he had a PA of -8, this chance would be at 94%. I agree with the logic of keeping the chances of player to see his PA raised on the lower end, as Manimal said.

The only downside, and it is a big one in my opinion, is if said player has a random PA set at 132, his CA will be dropped by a few numbers. It means that in the database, he is seen as a useful player in the NHL, but in the game, you can see his CA drop and not be able to use him anymore.

[nino33]

I believe I learned something new about EHM editing today!

Manimal and landry, I believe you are right!

I just did two tests of 10 saves each looking at Nathan Beaulieu. For both tests I selected the NHL only, and the only two Options I selected were Disable Extra Players and Disable Extra Staff (to make the process a little quicker). I used the TBL 9.1 Rosters, where Nathan Beaulieu has a 134 CA and a -14 PA

For the first 10 tests I would just quit and return to the startup screen, the next 10 tests I quit the game completely and restarted it (just in case there was a difference). Based on Beaulieu's 134 CA there's a 63% chance of Beaulieu having a PA of 134 (the 90-134 range is 44, 135-160 range is 26.....44 of 70 equals 62.9%). So, it would be expected that 6 out of 10 saves would see Beaulieu's PA be 134 - and that's exactly what the results were!

CA - PA (first 10 tests)
134 - 134
134 - 134
134 - 134
134 - 134
134 - 149
134 - 134
134 - 134
134 - 138
134 - 140
134 - 137

CA - PA (second 10 tests)
134 - 157
134 - 134
134 - 134
134 - 134
134 - 134
134 - 135
134 - 140
134 - 134
134 - 145
134 - 134

I had honestly thought that the game would disregard the range below the CA (in this case, the 90-134), ands use just the range from CA to PA...but I was wrong!

[Manimal]

Nice testing, Nino!
I guess knowing this for certain will change the way we look at young players and prospects.

[landry2119]

Glad to see that the CA drop is rare, only 5% of the time in your tests. I've seen Beaulieau's CA drop from 134 to 118 in one of my saves. Let's say I then had a problem to solve on the defense.

[nino33]

Glad to see that the CA drop is rare, only 5% of the time in your tests. I've seen Beaulieau's CA drop from 134 to 118 in one of my saves. Let's say I then had a problem to solve on the defense.

I don't understand what you're saying...while I never simulated forward (all my results just show day 1 of a newly started game), the CA did not "drop" at all, not even once

CA can drop for players as you simulate forward in a game, but there's no evidence that CA drops at all upon startup; my understanding is CAs will drop primarily due to age and/or due to playing at lower levels than appropriate ("stunting a player's growth"), as well as due to injury, not playing and/or due to randomness (a player "not developing")

[landry2119]

Oh! Maybe you mistyped the first entry as it is CA 124. Glad to see there is no drop at all then. I was just unlucky with him!

[nino33]

Oh! Maybe you mistyped the first entry as it is CA 124. Glad to see there is no drop at all then. I was just unlucky with him!

Oops yep, typing error (now fixed)

[nino33]

The testing I did got me curious...I recall a poster on another forum last year saying they'd played EHM for many years and they believed that players in EHM essentially "always developed" so there wasn't significant variability - so I ran another twenty tests like I did above, and this time I looked at 4 players with a CA near 130 and a -8 PA & 4 players with a 90 CA and a -14 PA

I looked at Kyle Connor (128 CA -8 PA), Christian Dvorak (130 CA -8 PA), Mikhail Grigorenko (132 CA -8 PA) and Derrick Pouliot (128 CA -8 PA) + Brock Boeser (90 CA -14 PA), Vince Dunn (90 CA -14 PA), Zachary Fucale (90 CA -14 PA) and Dmitri Sokolov (90 CA -14 PA).....the CAs were unchanging at startup, here's the results of the PAs

• Kyle Connor - 144, 133, 138, 131, 143, 139, 142, 134, 160, 152, 158, 153, 144, 141, 142, 139, 154, 142, 148, 155 (range 131-160, average 144.6)
• Christian Dvorak - 146, 153, 140, 138, 139, 137, 137, 135, 144, 148, 137, 133, 160, 133, 139, 152, 138, 156, 145, 142 (range 133-160, average 142.6)
• Mikhail Grigorenko - 153, 136, 149, 137, 146, 132, 141, 133, 132, 142, 132, 132, 145, 146, 148, 160, 138, 134, 154, 152 (range 132-160, average 142.1)
• Derrick Pouliot - 140, 134, 146, 132, 156, 144, 142, 152, 153, 131, 133, 144, 148, 130, 150, 159, 152, 142, 158, 137 (range 130-159, average 144.2)

• Brock Boeser - 130, 134, 137, 153, 94, 94, 108, 157, 115, 98, 159, 141, 134, 116, 153, 148, 97, 99, 106, 93 (range 93-159, average 123.3, 10/20 times 130+)
• Vince Dunn - 105, 144, 131, 91, 153, 135, 111, 117, 100, 147, 128, 118, 98, 122, 127, 104, 146, 138, 95, 95 (range 91-153, average 120.3, 7/20 times 130+)
• Zachary Fucale - 109, 98, 99, 144, 125, 105, 108, 143, 107, 127, 108, 151, 101, 93, 115, 98, 141, 150, 160, 135 (range 93-160, average 120.9, 7/20 times 130+)
• Dmitri Sokolov - 149, 93, 139, 116, 105, 91, 127, 108, 154, 145, 121, 152, 159, 101, 145, 100, 124, 152, 111, 113 (range 91-159, average 125.3, 8/20 times 130+)

[CeeBee]

I started a Canuck save a couple of weeks ago and Boeser had a PA of 111. I used the assistant to fix him rather than start over since IMO he's a lot more than a 111 IRL.

[Shindigs]

I'm going to try and find the save again, but regardless of it I still have it it will probably be a year or two further along. Never used the Assistant (or is the Assistant another name for the DB Editor?) so don't know if that's an issue.

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. There is a single exception to this, if a player has say, a max PA of 146 (a well rolled Nikita Feoktistov in my case in the 15/16 db) and you play in Swe-2 which is a quite low rep league, especially in my custom version of the db where I tuned down all rep in Europe to limit the use of NHL free agents in lower euro leagues. Since a player with that high PA isn't intended to ever be in Swe-2, he simply won't grow past around ~125-135 or so PA. I can't recall the exact value, but it's in that area. Once I got promoted to Swe-1 he instantly grew to his max potential in the off-season. Normally this scenario will never happen, but Feoktistov is an ooold prospect(-14 in the db) who is still left as a ghost player in the db with very low reputation, and he's in the belarusian league. So he will never go to the NHL due to starting at age 25, thus dodging the drafts. And his low reputation means he's perfectly happy to sign a 3+2 year deal in Swe-2.

But barring those extreme cases I have never had a prospect not develop fully, there is however a very interesting thing that happens tied to playoffs (but only in ECHL, AHL, NHL and CHL that I know of) where if a player has a performance in the 6.8-7.0 range he has a very large likelihood of gaining 1-3 Important Matches, as well as 1-2 Temperament and 1-2 Pressure. However since these attributes count towards a player's CA it has to "free up" CA if the player is already at a cap (either age PA cap, or his real PA cap). This leads to players looking like they "lost" attributes after a playoff (it favors taking skating, con and teamwork first) when they were actually just lowered to make room for the added Temp, Pressure and Imp. Matches gain. Since Temp is pretty much the best attribute in the game, this is a fair trade. But it means that if you go to the playoff a lot some of your player's skating will suffer because of it. Which isn't very logical, but it's what happens. Players with less than 6.8 also have a chance to gain the above mentioned attributes, but it's about half as likely. And player above a 7.0 rating have a very low chance to gain them. Which means that eventually the gained attributes will make the player so good in the playoffs that he will get close to no chance to gain any further gain in the attributes. This means you don't have to prioritize Important Matches and Temperament as much when you scout drafts. But Temperament is still just amazing, so I personally still rate it very highly.

Edit1: Found the save, It's indeed about a year along after the generation of that player. download link

Edit2: I'm surprised the -14 wasn't skewed more towards the top honestly. Have you run a test on -15? Because goalies with less than 170 PA are essentially useless in the NHL as anything but backups, having goalie prospects unable to go above 170 (-8 or -14) makes drafting goalies a pretty horrible experience. I've personally swapped all the "hotter" goalie prospects to -15, since that way they at least have a chance to be useful, but not the guarantee that -9 gives, since you just never freaking know with goalies. The thing I'm noticing is that a very odd amount of both the goalies that already had -15 (like Juuse Saros) in the db, as well as the ones I set to it end up with very good PA a somewhat uncanny amount of the time. I don't think I've seen Saros be sub 170 in like 7ish NHL saves. And with a -15 that's pretty sick, since you expect an average of 150ish. being 20 or more over for ~7 saves only really makes sense if it has some sort of bellcurve-esque distribution with a slight offset upwards. On the flip side I don't think I've ever seen him above 180, so he seems to end up in about a 10 PA range 7 times in a row (which also points towards a bellcurve distribution). Which could just be really uncanny RNG on my end. But it just seems like -15 favors the top end a lot. But almost never the very high end of it.

Most human ability follow a somewhat bellcurve-esque distribution in real life if I'm not mistaken, so it would make perfect sense if the PA ranges did too.

I just really wish there was a 130-180 PA option for goalies and mid 1st round skaters, since -8 often feels on the low end for them in a good draft year, and -9 makes them way too reliably good, which hurts replayability.

The gap between -7 and -8 also feels way too big, really wish there was a 120-150 in there for later 1st rounders and early 2nd rounders. -14 sort of fills that gap, but not really.