BfW is a game that relies on save-scumming too much

General feedback and discussion of the game.

Moderators: Forum Moderators, Developers

Post Reply
User avatar
Xalzar
Posts: 268
Joined: April 4th, 2009, 10:03 pm
Location: New Saurgrath

Re: BfW is a game that relies on save-scumming too much

Post by Xalzar » April 9th, 2019, 10:52 pm

I don't want to monopolize the topic so I'll take a little pause after this.
Aldarisvet wrote:
April 9th, 2019, 8:18 pm
Xalzar wrote:
April 9th, 2019, 8:10 pm
0 damage (miss) 20 %
1 damage 10%
2 damage 10%
3 damage (half) 20%
4 damage 10%
5 damage 10%
6 damage (full) 20%
This is not a plain distribution. Why 20% at the middle and the ends of outcomes and 10% in others? Must be 100%/6 everywhere. The problem is correcting this accounting a defence level.
I know it's not a normal distribution. I was talking about you proposal:
Aldarisvet wrote:
April 9th, 2019, 7:43 pm
But instead I got an idea that we not necessarily have to use a normal distribution. But it must be possible to create a plain distribution where a probability to make a zero damage by a dragonguard would be equal to probability of doing half damage. Dices creates a normal distribution but we have no need to stick with it.
...and I tried to see if I understood it providing an example.
The graphic helps, you want to have the black line so all results would be equally probable. Intriguing. And how would it work with defences? I'm still trying to figure out some details.
Aldarisvet wrote:
April 9th, 2019, 8:18 pm
Xalzar wrote:
April 9th, 2019, 7:30 pm
the duality damage/no damage remains, but every single strike rolls more than one dice.
Than it would change nothing.
Yes of course, that's the plan. :whistle: Also no, it does change something... :eng:
Aldarisvet wrote:
April 9th, 2019, 8:18 pm
There is no "single dice" actually. For now every strike is generated as a random number between 0 and 1 and the result is compared with a defence level.
"A" random number. Singular. One, which with a 80% cth could be 0 five times in a row. That's what I'm trying to change: I'm proposing multiple dice per strike.
It is known that the more dice you roll, the better the results are likely to be as expected. If I want a 6, I know I always have a 1/6 chance. But in 10 throws I can never get it, while in 100 I will have it exactly nearly 1/6 times.
It works this way in Wesnoth too: during a turn results vary immensely, while at the end of the match (if not too short) generally the RNG is balanced.
But turns are not the same: a critical turn can be ruined by strikes of bad luck and seriously skewed bad rolls. So to reduce the intra-turn variation the rolls need to be more.

For example, I'm aiming at the curve at having only the curve on the right when I have 70% of cth (assuming the left values are misses and the right ones hits, with the mean being the threshold)...
Image
...and not possibly anyone of the three in the same situation (skewed distributions).
An easier example would be that when I have 50% cth I want the middle curve, and not the others. If I collect all the results from all the fights at 50% cth during the match sure, I will have it; but I want it every turn, when I need it.

The only question is how much time does the game need for additional attack calculations, since I'm no expert about performance.

Konrad2
Moderator
Posts: 1739
Joined: November 24th, 2010, 6:30 pm

Re: BfW is a game that relies on save-scumming too much

Post by Konrad2 » April 10th, 2019, 6:52 am

Xalzar wrote:
April 9th, 2019, 10:52 pm
It is known that the more dice you roll, the better the results are likely to be as expected. If I want a 6, I know I always have a 1/6 chance. But in 10 throws I can never get it, while in 100 I will have it exactly nearly 1/6 times.
I'd like to make clear that, while your first two sentences are correct, your last sentence is is wrong. 100 times is not big enough for your certainty. You have to add at least a few zeroes before you can be so sure about it.

User avatar
sergey
Posts: 300
Joined: January 9th, 2015, 9:25 pm

Re: BfW is a game that relies on save-scumming too much

Post by sergey » April 10th, 2019, 7:40 am

Tad_Carlucci wrote: At present, if we have a 70% chance to hit, we "roll" a single "die" by generating a random number in the range zero (inclusive) to 1 (exclusive). If the result is less than 0.7, we score it a hit, otherwise we score a miss.
This is very trivial and important fact to understand how the hit/miss code works. Just wanted to emphasize that.
Xalzar wrote: I'm proposing multiple dice per strike.
I will translate your proposal into the random sequence terms.

Considering explanation by Ted_Carlucci, suppose we have a random sequence N1, N2, ..., Ni, ...
You are proposing to replace it with sequence
M1 = (N1 + N2 + ... + Nk) / K
M2 = (Nk+1 + Nk+2 + ... + N2*k) / K
and so on

As a result, you will make the sequence less "random" and more "average". That is what Tad_Carlucci said earlier:
Tad_Carlucci wrote: It sounds like you're suggesting we generate several random numbers and combine them.

This won't work.

We would need to rebalance everything. To keep the game the same "dificulty", in fact, we'd have to adjust the chance to hit downward, probably quite a bit. I've not done the math, but I'd guess to maintain a 70% actual chance to hit, we'd have to use something more like 52% or 53%. And, then, the number we're using is not "chance to hit" any more. It's the number dividing the area under a Bell curve at a 70%/30% split. That means every UMC author would need to solve an Integral equation rather than simply state "70% chance to hit". (We could make their job easier by pre-calculating the values, if we use a fixed number of samples.)
In the other words, since you will shift the sequence closer to average values, you will get 70% cth attack to hit more than 70% times and 30% cth attack to hit less than 30% times. If value K is very high, your sequence will be something like 0.49, 0.58, 0.46, 0.51, ... With that sequence, 70% cth attack will always hit and 30% attack will always miss. If you take not a very big K, anyway you are making 70% cth attack to hit more times and 30% cth attack to hit less times.
Xalzar wrote: every single strike rolls more than one dice. This way probability is conserved but more dice* throws garantee better the expected results (e.g. an attack with 80% cth will hit more likely 80% of the time instead of missing every single time).
That is wrong statement. This way probability is changed, but mean value is conserved. It doesn't guarantee better the expected result, it guarantees average result more often. With your proposal, a red mage (for example) will have a lower probability to miss all 4 strikes. But that is not because your algorithm is better, it is because the 70% cth attack will have higher probability to hit.

The problem is not related to the random generator quality. You can't make it better by tweaking it. The problem is related to the fact that random generator plays a significant role in this game. But wait, is it a problem or a design decision? :D
Last edited by sergey on April 10th, 2019, 4:13 pm, edited 1 time in total.
Author of SP scenario Dragon Fight and SP campaign Captured by a Nightmare

User avatar
Aldarisvet
Translator
Posts: 734
Joined: February 23rd, 2015, 2:39 pm
Location: Moscow, Russia

Re: BfW is a game that relies on save-scumming too much

Post by Aldarisvet » April 10th, 2019, 11:06 am

Xalzar wrote:
April 9th, 2019, 10:52 pm
That's what I'm trying to change: I'm proposing multiple dice per strike.
It is known that the more dice you roll, the better the results are likely to be as expected. If I want a 6, I know I always have a 1/6 chance. But in 10 throws I can never get it, while in 100 I will have it exactly nearly 1/6 times.
I still cannot understand what you propose.
If there is still miss/hit system (full damage/no damage at all system) for a single strike you can find no better system than current. A random number between 0 and 1 is already random enough to compare it with defence level to determine if there it is a hit or a miss. It is better than million dices because even million dices create fixed number of outcomes and there is no limit of outcomes between 0 and 1. If you take rational numbers.

______________________________________________________________

I suggest to abandon miss/hit system (full damage/no damage at all system).
Yes, I got that normal distribution (using multiple random 0 to 1 numbers compared to the defence level) is not good for that. Because it would create extremely low probability of a very low damage and a damage close to full.

Needed a "plain" distribution. Let us take a Dragonguard as attacker because it has only a single stike.
So something like that for 50% terrain defence level of defending unit (for 5 outcomes, can be more or less outcomes):

0 damage - 20% probability
10 damage - 20% probability
20 damage - 20% probability
30 damage - 20% probability
40 damage - 20% probability

How to get this in math? Just take a random number between 0 and 1. Then if it is between 0 and 0.2 - then it is a 0 damage, between 0 and 0.4 - 10 damage etc, between 0.8 and 1 - 40 damage.

But maths go much more complex given a defence level and a poison/slows problem (such simple distribution as above would work only for 50% terrain defence rate of the attacked unit and a poison problem simply cannot be solved here, needed even number of outcomes for it and in this case you do not know if 20 damage is a poison case or not).
If devs are interested in adandoning miss/hit system then there is a sense to discuss it further else I see no reason to waste my time on it anymore. I am not going to go into deep maths for nothing.

And finally about that.
Caladbolg wrote:
April 9th, 2019, 4:01 pm
Where are you pulling that 99% statistic from? You continue hammering on the point that the average player extensively uses save-loading to avoid bad rng, but I see no reason to believe that is the case. And I don't think your 11-year-old son is a good representative of the average player, nor would the two of you make up a sufficiently large sample to draw conclusions from even if he were.
I have my own campaign and seen enough of replays to judge how people play my campaign. The less experienced the player is, the more often he relies on reloads. So my view is less unfounded than an opposite which is simply founded on nothing.
Bah, "we urge you let your lvl3 veteran die because of strings of unluck since fair playing demands that". Do you really think it would work??? (sounds like an urge to not have a sex before a marriage, really). Moreover only on this thread devs realized that in no place in the game they explained that reloads are bad and decided to put something about it in the tutorial. While a game exists for decades! In no place in the game it is explained that it is normal when your mage misses three times in a row.
Last edited by Aldarisvet on April 10th, 2019, 1:21 pm, edited 1 time in total.
facebook.com/wesnothian/ - everyday something new about Wesnoth
My campaign:A Whim of Fate, also see Zombies:Introduction single map campaign
Art thread:Mostly frankenstains

Tad_Carlucci
Developer
Posts: 386
Joined: April 24th, 2016, 4:18 pm

Re: BfW is a game that relies on save-scumming too much

Post by Tad_Carlucci » April 10th, 2019, 12:45 pm

Thankfully, this entire thread is just talk and has no chance of ever making it into the game.

If you're interested in it, you can always make a mod. But the model proposed it far too complex. It would be much easier to simply eliminate the randomness, making it so all attaacks always deal full damage. That will model, well enough, the actual effect of this proposal. It won't solve the basic problem, however, because "what is sauce for the goose is sauce for the gander." The player will love always dealing full damage, but come the oppoonents turn the player is not gonna like being hammered. Since reloading won't change anything, the player will quickly see "the only winning move is not to play the game."

Perhaps the title of the thread should be changed to "Players, we don't need to stinkin' players!" so future readers will understand the true effect of this proposal before they waste time reading this thread.
I forked real life and now I'm getting merge conflicts.

User avatar
Aldarisvet
Translator
Posts: 734
Joined: February 23rd, 2015, 2:39 pm
Location: Moscow, Russia

Re: BfW is a game that relies on save-scumming too much

Post by Aldarisvet » April 10th, 2019, 1:37 pm

Tad_Carlucci wrote:
April 10th, 2019, 12:45 pm
It would be much easier to simply eliminate the randomness, making it so all attaacks always deal full damage.
You mean an average damage not full. So Dragonguard will always make a damage which is counted as its attack multiplied on terrain defence rate. For example against a shaman in the forest it would be 0,3*40=12. Against a peasant in the water it would be 0.8*40=32.

I suggested it from the very beginning. For sure it would be a better system than what it is now ;)
...
Heh, nope, this will ruin slow and poison. So no, it cannot be the only one outcome as a result of a single strike.
Tad_Carlucci wrote:
April 10th, 2019, 12:45 pm
Perhaps the title of the thread should be changed to "Players, we don't need to stinkin' players!" so future readers will understand the true effect of this proposal before they waste time reading this thread.
With years I get the growing feeling that this game exists not for players but for artists, programmers and add-on makers so they all can practice or satisfy their thirst for creativity ;)

Also, just for fun, I've made a search in Google for "Wesnoth RNG". I never did that before and just have got that idea. Kinda google results is full of "stinkin' players", some quotations from first lists:

"the horribly RNG playstyle just puts me off too much"
"Good game but one of the frustratingly rng games I've ever played"
"Another potentially great ... ...with a completely broken RNG"
"there are a bunch of mods that deal with RNG BS in a number of diffirent ways"
Last edited by Aldarisvet on April 10th, 2019, 2:35 pm, edited 1 time in total.
facebook.com/wesnothian/ - everyday something new about Wesnoth
My campaign:A Whim of Fate, also see Zombies:Introduction single map campaign
Art thread:Mostly frankenstains

otzenpunk
Posts: 32
Joined: February 11th, 2018, 5:32 pm
Location: Hamburg / Germany

Re: BfW is a game that relies on save-scumming too much

Post by otzenpunk » April 10th, 2019, 2:26 pm

I don't get what's the point. Some people don't like randomness. (At least whenever it turns against them.) That's fair. But I don't see reducing randomness in this way as a possible solution. The reason those people get annoyed sometimes is obviously, that they neglect the chance of events to happen, if the probability is low enough, maybe lower than 10%, 5%, whatever. That their beloved mage misses 3 times in a row, stays exposed, gets killed, and the scenario is lost, should simply not happen, just because it's unlikely by the rules.

But what would happen, if the distribution of the results is narrowed like proposed, for example by splitting every attack on 10 "dice"? My guess is, the same people would quickly adapt to the new "normal" battle outcomes. The same mage is now obligated to inflict at least half of the maximum damage with every attack, not just 1/3, because the chance to hit less than 15 out of 30 times with a 70% chance is probably ridiculously low. But of course it will happen from time to time, and the mage will be killed, and the player will rage about the RNG. What's next then? Splitting every attack on 100 dice rolls?

As far as I know, there are already several "no random" add-ons, who eliminate randomness completely, and I guess, that's what you people basically want. I for myself will happily stick to the traditional rules. (And I am a player, not an artist, programmer or add-on maker. ;) )

User avatar
Aldarisvet
Translator
Posts: 734
Joined: February 23rd, 2015, 2:39 pm
Location: Moscow, Russia

Re: BfW is a game that relies on save-scumming too much

Post by Aldarisvet » April 10th, 2019, 2:58 pm

otzenpunk wrote:
April 10th, 2019, 2:26 pm
I don't get what's the point. Some people don't like randomness. (At least whenever it turns against them.) That's fair. But I don't see reducing randomness in this way as a possible solution. The reason those people get annoyed sometimes is obviously, that they neglect the chance of events to happen, if the probability is low enough, maybe lower than 10%, 5%, whatever. That their beloved mage misses 3 times in a row, stays exposed, gets killed, and the scenario is lost, should simply not happen, just because it's unlikely by the rules.
We already passed this. People do not complain on RNG in general, they complain on Wesnoth RNG and mention that RNG is other games is not broken compared to Wesnoth RNG. You can study google results for it. So it is incorrect to mark people who do not like Wesnoth RNG as some silly people who always want better results for themselves and worse for the opponent.

Obviously people do not like excessive deviations. It is wrong logic that if deviations would be reduced then players would "get used of it" with time and will start complaining again.
facebook.com/wesnothian/ - everyday something new about Wesnoth
My campaign:A Whim of Fate, also see Zombies:Introduction single map campaign
Art thread:Mostly frankenstains

otzenpunk
Posts: 32
Joined: February 11th, 2018, 5:32 pm
Location: Hamburg / Germany

Re: BfW is a game that relies on save-scumming too much

Post by otzenpunk » April 10th, 2019, 4:07 pm

Aldarisvet wrote:
April 10th, 2019, 2:58 pm
We already passed this. People do not complain on RNG in general, they complain on Wesnoth RNG and mention that RNG is other games is not broken compared to Wesnoth RNG. You can study google results for it. So it is incorrect to mark people who do not like Wesnoth RNG as some silly people who always want better results for themselves and worse for the opponent.
But that's what it is. For some reason, I've never seen anyone writing something like "Oh, no, my badly wounded footpad just held a forest tile against four grunts. That shouldn't have happened. The RNG is definitely broken."

I've played Civilization on the Amiga back in the nineties, and the same "problem" existed back then. Because there was randomness involved, it happened, that an inferiour enemy unit just resisted several attacks, and your sophisticated plans were ruined. And then you either accepted it, or you reloaded, or you just trashed the game and played something else. Despite this obviously flawed RNG^^, the Civilization series were some of the most successful computer games at the time.

Of course, you find google results for your point, if you search for "wesnoth rng", because people only write about a game's rng, if they're mad about it. But I can imagine, what kind of people probably write this amongst others. Just watch some Youtube Let's Plays, where people repeatedly attack orcs with elfs at night on the plains in HttT-Blackwater Port, and then blame the game and the RNG, because they lose regardless how often they saveload. Play any commercial online games today, and people will continuously blame in chat everything for their poor performance but themselves. If there's no RNG involved, then their opponents are obviously cheating, etc.

Look instead at the rating of the game on Steam or Google Play Store. They are overwhelmingly positive, so it doesn't seem to be a problem for most of the players. Of course, there are also users giving bad reviews like
Battle for Wesnoth is a really cool game that I can't recommend. The RNG feels broken. And no, I don't mean the fact that there IS RNG somehow ruins the game, because it doesn't. The reason it feels broken is because the percentages shown never seem to match up with what actually happens. Seriously, 50% chance to hit feels more like 30%, 40% feels like 20%, etc. Oh, and your chance to hit will rarely go above those for some reason. Occasionally you might get the chance to attack with a 60% or 70% chance to hit, but it doesn't really matter because those still feel like <50% anyway. Only play this if you plan on doing away with the RNG entirely via add-ons or if you particularly enjoy pulling your hair out.
Well, we all know that this user's feelings just don't reflect reality.

Maybe this is just a thing because of the transparency and openness of Wesnoth itself. If you don't know exactly, how battle results are resolved, it's maybe psychologically easier to accept the result, than if you exactly know how unlikely what just happened was.

Whatever… there are "no random" add-ons. Good for people who don't like randomness.
They most likely won't replace the existing default mechanism. Good for everyone else.

User avatar
Aldarisvet
Translator
Posts: 734
Joined: February 23rd, 2015, 2:39 pm
Location: Moscow, Russia

Re: BfW is a game that relies on save-scumming too much

Post by Aldarisvet » April 10th, 2019, 4:25 pm

otzenpunk wrote:
April 10th, 2019, 4:07 pm
I've played Civilization on the Amiga back in the nineties, and the same "problem" existed back then. Because there was randomness involved, it happened, that an inferiour enemy unit just resisted several attacks, and your sophisticated plans were ruined. And then you either accepted it, or you reloaded, or you just trashed the game and played something else. Despite this obviously flawed RNG^^, the Civilization series were some of the most successful computer games at the time.
Well, you happen to meet a person who spent several years playing a multiplayer in Civ4 (I also played the game from civ1).
I assure you, no one there ever cried about RNG. [censored] could happens there too of course. I've seen a scout killed a horse archer defending on plains, but this happened only 2 times in my experience. There is totally different RNG system there and the probability of a weak unit to win a strong unit is very very low. Just google "civilization rng" and try to find if there persons who do not like Civ's RNG.

So there is a RNG and there is a RNG.
otzenpunk wrote:
April 10th, 2019, 4:07 pm
Look instead at the rating of the game on Steam or Google Play Store. They are overwhelmingly positive, so it doesn't seem to be a problem for most of the players.
It is positive despite of RNG. Or, to say accurate, the RNG is not a problem for most players because they always can reload an excessive unluck. And they gladly use this possibility. They do it, almost everyone do, I myself did it a lot when I just started playing Wesnoth. Let us stop denying obvious.

Remove an excessive RNG volatility - and you will solve so many problems and lose nothing. Is not that so simple?
Last edited by Aldarisvet on April 10th, 2019, 4:56 pm, edited 3 times in total.
facebook.com/wesnothian/ - everyday something new about Wesnoth
My campaign:A Whim of Fate, also see Zombies:Introduction single map campaign
Art thread:Mostly frankenstains

User avatar
sergey
Posts: 300
Joined: January 9th, 2015, 9:25 pm

Re: BfW is a game that relies on save-scumming too much

Post by sergey » April 10th, 2019, 4:38 pm

Battle for Wesnoth is a really cool game that I can't recommend. ... 50% chance to hit feels more like 30%, 40% feels like 20%, etc. ... Only play this if you plan on doing away with the RNG entirely via add-ons or if you particularly enjoy pulling your hair out.
This quote made my day :D I prefer to pull my hair out until I feel 50% like 50%, 40% like 40%, and so on :lol:
Author of SP scenario Dragon Fight and SP campaign Captured by a Nightmare

otzenpunk
Posts: 32
Joined: February 11th, 2018, 5:32 pm
Location: Hamburg / Germany

Re: BfW is a game that relies on save-scumming too much

Post by otzenpunk » April 10th, 2019, 4:54 pm

Aldarisvet wrote:
April 10th, 2019, 4:25 pm
I assure you, no one there ever cried about RNG. […] Just google "civilization rng" and try to find if there persons who do not like Civ's RNG.
Well, maybe you should've done that first… :mrgreen:

https://forums.civfanatics.com/threads/ ... tor.65359/
https://forums.civfanatics.com/threads/ ... rng.96902/
https://rnggodsite.wordpress.com/
[…]

User avatar
Aldarisvet
Translator
Posts: 734
Joined: February 23rd, 2015, 2:39 pm
Location: Moscow, Russia

Re: BfW is a game that relies on save-scumming too much

Post by Aldarisvet » April 10th, 2019, 5:10 pm

otzenpunk wrote:
April 10th, 2019, 4:54 pm
Well, maybe you should've done that first… :mrgreen:
I did that first. Compared "Wesnoth RNG" results it is absolutely nothing. A couple of threads in civfanatics, heh, just find how many times this question was raised in these official forums. But the fun is that most Wesnoth RNG-angry posts google finds in other sites than this official. And also do not forget to compare popularity of Civilization and Wesnoth because Civ has much greater audience.
facebook.com/wesnothian/ - everyday something new about Wesnoth
My campaign:A Whim of Fate, also see Zombies:Introduction single map campaign
Art thread:Mostly frankenstains

User avatar
sergey
Posts: 300
Joined: January 9th, 2015, 9:25 pm

Re: BfW is a game that relies on save-scumming too much

Post by sergey » April 10th, 2019, 5:26 pm

Aldarisvet wrote: People do not complain on RNG in general, they complain on Wesnoth RNG and mention that RNG is other games is not broken compared to Wesnoth RNG.
Let's see how many people like / dislike the Wesnoth RNG Poll: do you like Wesnoth RNG?
Author of SP scenario Dragon Fight and SP campaign Captured by a Nightmare

User avatar
Aldarisvet
Translator
Posts: 734
Joined: February 23rd, 2015, 2:39 pm
Location: Moscow, Russia

Re: BfW is a game that relies on save-scumming too much

Post by Aldarisvet » April 10th, 2019, 9:25 pm

https://www.gog.com/forum/general_archi ... of_wesnoth

Some old short thread on Gog, 13 post, and 5 persons complain about bad 'dices' system of Wesnoth. They do not use 'RNG' word, I found that thread by googling 'f.k wesnoth' just for curiosity. Why I posting this? Because cocooning ourselves in this forum we can never see the outer world. I am quite surprised myself about the degree people do not like Wesnoth RNG. They write they like the game but not 'hit rolls'.
facebook.com/wesnothian/ - everyday something new about Wesnoth
My campaign:A Whim of Fate, also see Zombies:Introduction single map campaign
Art thread:Mostly frankenstains

Post Reply