RNG (a long dead horse I'm sure)

Share your thoughts and ideas about Mordheim: City of the Damned
Magister Gir

RNG (a long dead horse I'm sure)

Postby Magister Gir » 15 January 2016, 09:26

So this isn't going to solely be based around RNG as I do think the game can improve on other things and not simply the RNG.


TO start it off I do believe this is a great game, it's addictive and challenging. Though I do believe it is a bit misleading about exploration, even though the maps are always an enjoyment to see. I love the lose of body parts, the stat growth, spell system and all of that fun stuff and can only praise it and it's what constantly draws me back to this game. That and my love for Warhammer.

The off setting part is the RNG. I kept reading that a system update on it was implemented and was fixed but I just didn't see it and feels that the enemies (AI) were heavily favored in it. I haven't played in 2 weeks and decided it was time, since the comments were saying RNG was fixed/updated. I don't know if I just had extremely horrid luck or not because I played two campaigns, one skaven and one merc's. Both of my leaders were instantly killed. One turn killed. My Skaven leader, and I think this is a bug so I'll be posting shortly after this, was ambushed and hit rather hard. THe opponent was the Cult leader and hit for 42 damage (big ouch). Thinking well lets get some damage in! Missed a 80% hit and was counter attacked for another big 40 damage. Not wanting to take the risk again, since two more hits I was dead, I put him in a 50% dodge. His turn then, dodge fail, hit for another 40, casts a spell, fails, swings again, hits leader is dead. That's when I quit I started a new.

Merc's turn, had an easier time with them before, got pretty far so should be fairly good. Started off in a bunch this time so no need to group up. Started moving them forward together, leader is a bit head but not much. Enemy champion spots me and jumps entering an ambush stance right next to my leader. Nothing else to do I charge him since my guys turns were right up next after an unknown. Charge misses from 73%, get hit, hard. 48 damage from a two hander. Swing again and hit for 22 damage on a 82% he doesn't swing thankfully. Enter a 44% parry. Enemy unknown turned out to be their mace/shield wielding leader who charged me. CHarge hits, parry failed, a quick thought of "well damage- oh god what!?" I get hit for 48 damage + stunned. There was a bug that didn't show the weapons of destruction buff. Now stunned it just took one more swing dealing 45 damage and leader was dead. At that I just left the game too frustrated to try another campaign.

I personally don't know what's going in RNG but I keep hearing that it's better, fixed, and not AI favored but I don't see it happening. It constantly feels like I'm at an disadvantage again NORMAL AI opponents, especially this time around. I went through patch notes and couldn't see anything actually being said about RNG so I'm assuming it was never really updated and still look forward to a possible update and improvement on the RNG because the harsh punishments for something that I personally cannot control at all just seem hardly fair. The steam forum/reviews just bash anyone that try and say something about RNG so thought maybe here could get some insight on possible touches being done to the RNG. I didn't see a post done on RNG so this is why I'm making one. The advice I can give, though it's from someone who doesn't know how the coding/difficulty of it goes, is that show the actual % of what the enemy's dodge/parry/natural stats do to your chance of hitting. I always feel like I'm lied to when I'm shown my hit % so maybe a clearer statement on how it works? would be greatly appreciated.

The stats buff, especially early game, feel like the game is setting up the player for failure and unneeded hardships for the term "hardcore" to stick. I think you don't need to do this except for brutal and deadly. I think keeping the AI at a fair playing field would be just fine, especially for new players to this type of game/genre. The hardcore is already there no less with the punishments, wyrdstone shipments and more, why add in something else that's not needed? Again I don't know coding but maybe instead of bumping up the stats on AI, especially on normal, have them go for objectives as well. It would add in a bit more and not just have the enemies surge for you, which gets incredibly repetitive and dull. As for the scaling hard, brutal, and deadly try a feature of adding in more bots for hard, Brutal has the same number but stats boosts, while Deadly is more Ai with stats boosts. This actually adds more strategy, intense battles and not just a simple repetition of what you do for any other mode. Just thoughts on what can happen and hope they're considered.

The difficulty ratings seem to be a bit wonky. By this I mean that since there's no clear description of the how difficult the normal, hard, brutal, and deadly missions are it's hard to tell if the normal is actually more difficult than the deadly. An example is that sometimes a brutal mission that I went on would actually be more of a normal while a normal mission was suddenly my entire warbands downfall (Mercs vs Skaven, Skaven swept me under a rug with just two rats) while the brutal mission I was on just before against Cult was actually a challenge but very doable. I think a possible % chance of succession is needed or a screen that shows how tough the enemies will be before deciding to go into it (of course not who they will be however).

A suggestion on the punishments. I think it might be better to lessen the punishment depending on the difficulty, I can't tell you how frustrating and discouraging it is to love 4 men and 2 mortally injured men at the end of a mission due to a failure on a normal. 2 of the dead actually being alive at the end of the mission no less! (route test failure) This severely ruined any chance of getting back on my feet as I was simply a level 4 warband but I was already ahving difficulty trying to level up fresh recruits and being forced to start all over wasn't going to happen since it's based on warband level, not character levels. Maybe a stack punishment change can be implemented so that failing a normal doesn't feel like it's going to force you to restart a new campaign. I know about the seasoned warriors that you can buy but they're so pricey and even guides/forums have said that it's better to start anew then purchase them (I wouldn't have enough gold for more soldiers anyway).

The last thing is climbing/jumpdown damage. Later on it's not so bad, unless constant failure, but early on it's huge and discourages any form of using it unless you have a 90-95% chance of making it. The damage isn't very reliable either as it can jump from a simple 5 to a sudden 12 from the same height and back to back. Maybe a scaling damage system can be implemented so it's so random, the higher the more damage. I saw this recommended in another forum so I hope it's considered.

After all that this is simply suggestions/complaints, no denying the complaints. It can be listened to or not, I'm going to keep on trying, playing, and hopefully succeeding but I do hear these as well from friends who aren't as driven with Warhammer as I am which is unfortunate. I look forward to future patches and hope more players come into the beautifully done game!

Final thing: the captcha is a huge pain.

Wulfrik
Posts: 2
Joined: 23 January 2016, 06:24

Re: RNG (a long dead horse I'm sure)

Postby Wulfrik » 23 January 2016, 06:34

I gotta be honest, the RNG you described is nothing to what I experience constantly, and drives me absolutely f**king nuts. On a regular basis, and i've been keeping track of this game to game, I will miss half or sometimes over half of my attacks that reguire an 85%+ chance to hit. This further extends to my mercenaries who are often rank 10 and have a 95% hit chance, who still manage to miss a good third of the time minimum. I look at the roll charts, and the rolls I get will be without exaggeration 97, 92, 99, 100, 73, 92. This is not a statistical average whatsoever, it's garbage...

Keep in mind that this is just attempts to hit, not even factoring in their dodge chances (which they will often pass on a 30% chance while I regularly fail 90% parries). I am also aware of confirmation bias, and I assure you that i've specifically been avoiding that from happening. Then on to magic, where i've recently been using sisters. I will suffer divine retribution every single spell, occasionally getting lucky and not having a negative buff. Then the enemy sister proceeds to use 3 back to back comets, hits every single one, and doesn't suffer a miscast once.

I really wish I could love this game, and I hate myself for constantly being dragged back to it, as i've always been a massive warhammer fan. The game is awesome, and this alone single handedly ruins 90% of the enjoyment for me. I can even look past the enemies AI regularly running into walls and spamming ambush right in front of me. Even if this was implemented to scale the difficulty and make the game challenging it is such a cop-out. I would have 0 problem with a good RNG system and the enemies just being overpowered, with more hp or damage. As it is, I will likely keep obsessively playing, but I cannot give this game a positive review when this happens.

User avatar
Drildgen
Posts: 564
Joined: 07 November 2014, 15:32

Re: RNG (a long dead horse I'm sure)

Postby Drildgen » 26 January 2016, 14:59

[...] the rolls I get will be without exaggeration 97, 92, 99, 100, 73, 92. This is not a statistical average whatsoever, it's garbage...
As per the laws of probability and mathematics, you have just as much chance to get this series of rolls (97, 92, 99, 100, 73, 92) on a D100 than getting six consecutive 1's or six consecutive 100's... Understanding that is what makes the difference between fake and true RNG. Mordheim uses a true RNG system, meaning that every time you perform an attack, you make the same dice roll, with the same outcome probabilities. Because you just failed a roll, doesn't guarantee your next roll will be a success, since you have the exact same outcome probability as your first roll. This is exactly how it worked on the tabletop game, because you used real dices to perform your checks.

You also need to consider that an attack resulting with the enemy dodging IS a successful attack. The target just happens to have succeeded his/her dice roll on his/her dodge check.

I hope this clarifies the situation and the doubts concerning the RNG system. It is here to stay and we're not planning to modify it in any way, shape or form.

Best regards,
Good hunting in the streets of Mordheim!

-Drildgen

brasky

Re: RNG (a long dead horse I'm sure)

Postby brasky » 27 February 2016, 05:00

Actually Drildgen, the number choices you used have a significantly lower chance of occurring than what was stated. To roll not just a single 1, but consecutive ones, does NOT have the same chance as rolling say a 13, then a 35. Also, It does not take a mathematician to simply keep track of their rolls over a period of several battles. According to the rolls I made over 5 battles that I kept track on, I had characters who rolled over 70% of their attacks of an 85. None of my characters had an average below 70%. This is not RNG. This is the game deciding before you even roll that you will miss. And it is easy to keep track of hits vs dodge/parries, since the game tells you if it was a miss or a dodge/parry. Remember an alcoholic that denies he has a problem is still an alcoholic. Denying that the supposed RNG in this game is anything but fixed, doesn't not help the game.

brasky

Re: RNG (a long dead horse I'm sure)

Postby brasky » 27 February 2016, 05:39

Hate to keep beating a dead horse, but after reading the most recent patch notes, I noticed the RNG algorithm was changed. If it was a PURE rng system as previously stated, then what could you change to make it more pure? I like that less that a month after saying it didn't have problem they have to put up in the patch notes that it has been tweaked. You guys have a real chance at a great game here, the extra stats and such given to the enemy is enough of a difficulty curve. You don't need to cheat numbers behind the scenes. Sadly at this point, there is NO strategy to the game, even with the log and all available information available to you, you can not make an educated or thought out decision. I understand that making a good A.I. is hard, hell if it was easy we would have terminators kicking the shit outta ISIS right now. I look forward to the next patch and hope that instead of simply admitting that their is a problem with the rng with a "small tweak to the algorithm in hope it’ll yield better results"(nothing like admitting you don't know if it will fix it only HOPE it will.) It will say "changed RNG to no longer unfairly favor the computer.

User avatar
MasterN
Posts: 82
Joined: 16 January 2016, 10:49

Re: RNG (a long dead horse I'm sure)

Postby MasterN » 01 March 2016, 23:07

I'm not going to quote the entire post....
Hate to keep beating a dead horse, but after reading the most recent patch notes, I noticed the RNG algorithm was changed. If it was a PURE rng system as previously stated, then what could you change to make it more pure?
There is an issue in generating random numbers on PCs leading to the next best usable thing: Pseudorandom number generation (https://en.wikipedia.org/wiki/Pseudoran ... _generator).

If time in initial development wasn't enough to make it as random as physically/algorithmically possible, but only close to that, then an improvement on the algorithm may change the statistical probability a bit towards true randomness. I don't really feel any difference though.


For all people having trouble with the rng I suggest a read of the following:
https://en.wikipedia.org/wiki/Confirmation_bias

User avatar
Drildgen
Posts: 564
Joined: 07 November 2014, 15:32

Re: RNG (a long dead horse I'm sure)

Postby Drildgen » 02 March 2016, 14:49

Excellent post MasterN ;)
Good hunting in the streets of Mordheim!

-Drildgen

User avatar
Morjax
Posts: 178
Joined: 19 September 2014, 16:22
Location: USA - Central Time (GMT -0600)
Contact:

Re: RNG (a long dead horse I'm sure)

Postby Morjax » 02 March 2016, 17:46

It may or may not add to the discussion, but here's a post about the update that occurred a couple months back:
Katarak [Developer]

ok here's my random talk.

for many years now the Mersinne-Twister algorithm has been the standard for RNG. It's a solid RNG, but in the case of videogames where the amount of randoms requested are comparatively low, It is not well Equidistributed. It also suffers from "bad states" where the "randomness" of the numbers will be skewed. It also has a warmup behaviour (where the initial X random numbers aren't really random).

SFMT is the usual algorithm used which corrects a little bit of these issues. https://www.wikiwand.com/en/Mersenne_Twister#/SFMT

Furthermore, WELL, based on MT and developed by some of the same engineers, presents better equidistribution, and removes (or atleast diminishes) warmup. https://www.wikiwand.com/en/Well_equidi ... iod_linear

WELL is what we used in the Mordheim.

Despite all that, there have been many reports internally, and externally that our RNG doesn't provide adequate randomness, is skewed to one side of the percentile spectrum etc. People saying that they have 75% chance to succeed their climb and fail 6 times in a row. All these situations are real. How often they happen, I don't know, but there is enough noise about them for them to be tackled somehow.

Discussions have bounced around between Kes and I about the best approach to correct this and normally talks go to "cheating" the random. This is something we did when working on Dawn Of Heroes (DS) where you might have a very powerful skill at 25% chance to succeed. You only had the opportunity to use the skill about 4 times in a match. If you failed all attempts then you felt cheated. In order to counteract this we cheated the randoms for every fail. You had 25% chance. If you failed that would go up to 50% chance. and once you succeeded it would drop back down to 25%.

This to us, however, is totally last resort. It will change a lot the behaviour of the game. Change it's charm, and totally change the difficulty curve. Many people state that the AI cheats with its random, but I can guarantee it does not. The random rolls work exactly the same for all warriors in the game. Yours, Deamons, Bosses, All AI, Injuries, Market rotation, etc.It is totally unbiased, we just call the WELL algorithm and get the value.

Rencently, a new RNG algorithm has surfaced. So new in fact it doesn't even have a wikipedia page. http://www.pcg-random.org/ I took the time to read the docs, and it seemed interesting. Some interesting points are better equidistribution and faster computation. The computation time is most probably negligable, but I still feel in my heart that it will reduce loading times where we do over 5000 random checks sometimes.

So we just plugged it in, removed WELL, put in PCG. That's it. We didn't add any cheats, we didn't add skewing code. We just replaced the algorithm. PCG should provide more balanced results.

I hope this clears up everyone's concerns about the RNG changes.

Thanks, Katarak
(my emphasis added)

As a tangential little tidbit, there was a previous fire emblem game where the Japan release had a "true" RNG, somewhat like what I believe XCOM 2 to have, where a 95% chance literally has a 5% chance to fail. The US/EU release of that same fire emblem game, however, had a padded RNG where the more times you failed, the more likely you were to pass (so the feeling of losing streaks was diminished).

I believe Mordheim uses a "true" RNG, and does not pad the numbers (except, for example, the recent addition of "known path" after failing a climbing test). This means that a 95% chance to hit is just that. There is a 0.05 chance that you'll miss, and a 0.05*0.05 chance that you'll miss twice and a 0.05*0.05*0.05 chance that you'll miss three times. The data sets that would typically be used to test how well-distributed an RNG algorithm is I'm sure is very large, and likely larger than the number of rolls that a typical player will experience. losing streaks suck, but they happen... as do winning streaks.

P.S. With "true" RNGs you get lovely thingslike this little gem from XCOM 2 ;)

brasky

Re: RNG (a long dead horse I'm sure)

Postby brasky » 04 March 2016, 02:34

If what Katarak said was posted as a response to all the people complaining(both here and on steam) then this wouldn't seem like such an issue with people. Its one thing to have a developer come out and admit that there are things like "warmup" behavior and the randoms are not well Equidistributed. It is quite another to have mods and random forum people try to tell you that it is just a bias on our part. I, and most others, can deal with rng not being perfect, however its much more frustrating coming here to complain about a real issue, as recognized by developers, and then be told by forum monkeys and mods that were just bias or wrong. Just because YOU have not dealt with or were unable to recongnize those flaws in the rng, doesn't mean they didn't/don't exist. And to MasterN might I suggest reading the following.


Katarak [Developer]

ok here's my random talk.

for many years now the Mersinne-Twister algorithm has been the standard for RNG. It's a solid RNG, but in the case of videogames where the amount of randoms requested are comparatively low, It is not well Equidistributed. It also suffers from "bad states" where the "randomness" of the numbers will be skewed. It also has a warmup behaviour (where the initial X random numbers aren't really random).

SFMT is the usual algorithm used which corrects a little bit of these issues. https://www.wikiwand.com/en/Mersenne_Twister#/SFMT

Furthermore, WELL, based on MT and developed by some of the same engineers, presents better equidistribution, and removes (or atleast diminishes) warmup. https://www.wikiwand.com/en/Well_equidi ... iod_linear

WELL is what we used in the Mordheim.

Despite all that, there have been many reports internally, and externally that our RNG doesn't provide adequate randomness, is skewed to one side of the percentile spectrum etc. People saying that they have 75% chance to succeed their climb and fail 6 times in a row. All these situations are real. How often they happen, I don't know, but there is enough noise about them for them to be tackled somehow.

Discussions have bounced around between Kes and I about the best approach to correct this and normally talks go to "cheating" the random. This is something we did when working on Dawn Of Heroes (DS) where you might have a very powerful skill at 25% chance to succeed. You only had the opportunity to use the skill about 4 times in a match. If you failed all attempts then you felt cheated. In order to counteract this we cheated the randoms for every fail. You had 25% chance. If you failed that would go up to 50% chance. and once you succeeded it would drop back down to 25%.

This to us, however, is totally last resort. It will change a lot the behaviour of the game. Change it's charm, and totally change the difficulty curve. Many people state that the AI cheats with its random, but I can guarantee it does not. The random rolls work exactly the same for all warriors in the game. Yours, Deamons, Bosses, All AI, Injuries, Market rotation, etc.It is totally unbiased, we just call the WELL algorithm and get the value.

Rencently, a new RNG algorithm has surfaced. So new in fact it doesn't even have a wikipedia page. http://www.pcg-random.org/ I took the time to read the docs, and it seemed interesting. Some interesting points are better equidistribution and faster computation. The computation time is most probably negligable, but I still feel in my heart that it will reduce loading times where we do over 5000 random checks sometimes.

So we just plugged it in, removed WELL, put in PCG. That's it. We didn't add any cheats, we didn't add skewing code. We just replaced the algorithm. PCG should provide more balanced results.

I hope this clears up everyone's concerns about the RNG changes.

Thanks, Katarak


Conformation bias ehh? Seems the people who actually made the game say the problems reported do exist, and they can only be mitigated to occurring as infrequently as possible. Meaning some of us are just naturally gonna get the raw end of the deal. Just because it didn't happen to you, doesn't mean it didn't happen. Maybe you should read the comfirmation bias page again?

User avatar
MasterN
Posts: 82
Joined: 16 January 2016, 10:49

Re: RNG (a long dead horse I'm sure)

Postby MasterN » 04 March 2016, 10:03

First, I don't really give a toss about steam... So all comments refer to this forum.
Its one thing to have a developer come out and admit that there are things like "warmup" behavior and the randoms are not well Equidistributed. It is quite another to have mods and random forum people try to tell you that it is just a bias on our part.
Drildgen, as part of the dev team, is, for all intents and purposes, representing Rogue Factor. This forum is solely moderated by him (as far as I can tell).
The randomness of people in a community is fairly irrelevant. What's being said is what matters. (helps if people are registered memers ;) )

I, and most others, can deal with rng not being perfect, however its much more frustrating coming here to complain about a real issue, as recognized by developers, and then be told by forum monkeys and mods that were just bias or wrong. Just because YOU have not dealt with or were unable to recongnize those flaws in the rng, doesn't mean they didn't/don't exist. And to MasterN might I suggest reading the following.
I read it the first time buddy...

Let me tell you why I consider most complainers biased:
- I have not seen a single post where somebody complains about a situation where the rng "screwed up" in his/her favor.
Having a screwed distribution means, that the player is gaining equal benefit/detriment from it (statistically speaking).
- There's just the term "feels" around. Where's your data? The log is continuously written into a .txt file. Make a statistic and prove your complains, then no one can argue. (I'll mention braskys input shortly)
- many people do not complain about the rng per se, but that they (again->) "feel" cheated

According to the rolls I made over 5 battles that I kept track on, I had characters who rolled over 70% of their attacks of an 85. None of my characters had an average below 70%.
a) 5 battles (in all likelihood) give you less than 1000 rolls. That's a low amount.
b) How did the AI roll? Important here, because if it also has an average above 70%, then that would be an entirely different issue with the rng.
c) what is the total distribution of rolls?

Sadly at this point, there is NO strategy to the game, even with the log and all available information available to you, you can not make an educated or thought out decision.
How the hell do you make that out?
That's all the information you need... How can any of this affect your decision making? Do you mean that you can't guess right 100% of the time?


Couple of things at this point:
If it can be proven, that there's a sort of cheating or biased rng, then that would need to be corrected. But no complainer so far brought forth any significant evidence to prove that. I've also been hit with the bad luck. But I didn't fail to notice the good luck, so I don't have a reason to complain.

I make you a proposition. I'll write a program that analyses the log, and upload it's source code here (so people can check whether it's flawed or not (the code and the rng)).
Then, people can check for themselves.

I planned (had the idea some time ago, but little time to actually do it) to analyse the general and AI/player distribution , AI/Player biasing on CtH, CHC, dodge/parry, stuns and all alone checks (climb happens too rarely to be of any meaning).
I haven't done it so far, bacause I even lack the time to play atm...

update:
AWWW MAN.... the logs are rubbish (... to parse for the desired information...)
update2:
finally managed to get the rolls out of the text. my most recent log only has 107 rolls, so no real data yet. now to figure out how to assign the rolls to the faction. (c probably wasn't the best choice...)


Return to “Feedbacks”

Who is online

Users browsing this forum: No registered users and 1 guest

cron