A mod to make Wesnoth's RNG a little less frustrating

General feedback and discussion of the game.

Moderator: Forum Moderators

enclave
Posts: 936
Joined: December 15th, 2007, 8:52 am

Re: A mod to make Wesnoth's RNG a little less frustrating

Post by enclave »

gnombat wrote: May 12th, 2018, 11:33 pm In fact, I don't think there have ever been any preferences that actually change the rules of the game (as opposed to merely modifying the user interface) - this would be the first?
Yeah that's quite true :) The largest effect on gamestate is currently made by "accelerated speed" which would make game faster for everyone if host change it on map with AI computer players.. the planning mode probably takes #2 in that list... which doesn't really change any game state at all :) While both definitely don't change the rules of the game ;)

But
Who said that it should be added into preferences... it could be a tick when you host a game... in this case there are a lot of things that change the gamestate from that host menu (eras, modifications, timer, income.. etc etc etc etc etc etc)
Pentarctagon wrote: May 12th, 2018, 10:44 pm I assume a UI checkbox will be added for this at some point? Otherwise there probably aren't many people who will provide feedback.
So I guess gnombat is right, it shouldn't be a UI checkbox.. it must be a host game menu..
gnombat
Posts: 682
Joined: June 10th, 2010, 8:49 pm

Re: A mod to make Wesnoth's RNG a little less frustrating

Post by gnombat »

enclave wrote: May 12th, 2018, 11:54 pm But
Who said that it should be added into preferences... it could be a tick when you host a game... in this case there are a lot of things that change the gamestate from that host menu (eras, modifications, timer, income.. etc etc etc etc etc etc)
Pentarctagon wrote: May 12th, 2018, 10:44 pm I assume a UI checkbox will be added for this at some point? Otherwise there probably aren't many people who will provide feedback.
So I guess gnombat is right, it shouldn't be a UI checkbox.. it must be a host game menu..
The current implementation sounds like a preference ("set use_prng=yes in your preferences file") but maybe that's only intended for testing the feature. It would make more sense if it were one of the settings for hosting a game.
Computer_Player
Multiplayer Moderator
Posts: 181
Joined: March 16th, 2008, 6:39 am

Re: A mod to make Wesnoth's RNG a little less frustrating

Post by Computer_Player »

gnombat wrote: May 12th, 2018, 11:33 pm
Computer_Player wrote: May 12th, 2018, 10:39 pm The current mods that mess with RNG already destroy balance (since the entire game wasn't balanced for different RNG schemes, save for some rare eras).
enclave wrote: May 12th, 2018, 11:24 pm Same thing will happen if I use my "different luck" modification with my Q civ add-on... it will be a mess... so why worry... most add-ons and modifications already incompatible with each other, but nobody makes a panic of it..
The current mods that modify the RNG are all add-ons, though. I think there's a difference between an add-on that modifies the game and a core preference that changes the game's behavior. In fact, I don't think there have ever been any preferences that actually change the rules of the game (as opposed to merely modifying the user interface) - this would be the first?
The main difference really is that mods don't affect campaigns. The easiest way to implement this actually is to choose a mod, and graft it into the main game such that it affects campaigns as well. OR, "mainline" a RNG mod, and make it so that you can play campaigns from the scenario player menu (since that menu already has mod capability)
User avatar
Ravana
Forum Moderator
Posts: 2949
Joined: January 29th, 2012, 12:49 am
Location: Estonia
Contact:

Re: A mod to make Wesnoth's RNG a little less frustrating

Post by Ravana »

I hope use of this mod if this is actually released will be be possible to detect with Lua without having to rely on statistical methods like [do_command]. Similarly, access to this randomness generator might be useful for other scenario events.
The main difference really is that mods don't affect campaigns.
No such limitation anymore.
User avatar
Celtic_Minstrel
Developer
Posts: 2166
Joined: August 3rd, 2012, 11:26 pm
Location: Canada
Contact:

Re: A mod to make Wesnoth's RNG a little less frustrating

Post by Celtic_Minstrel »

Can we please stop referring to this as "experimental PRNG"? Wesnoth's existing system is also a PRNG (Mersenne twister), but this phrasing implies that being PRNG is what makes this proposal differ form the default RNG.

Without understanding what exactly this does it's hard to propose a good alternative name, but... even something like "Experimental alternative RNG" would be better than "Experimental PRNG". I imagine Dave (as the person who implemented it) could probably come up with something better.

As for the proposal itself, I guess I'm pretty neutral about it? It'll be interesting to see how it works out, at least.
Author of The Black Cross of Aleron campaign and Default++ era.
Former maintainer of Steelhive.
User avatar
Iris
Site Administrator
Posts: 6797
Joined: November 14th, 2006, 5:54 pm
Location: Chile
Contact:

Re: A mod to make Wesnoth's RNG a little less frustrating

Post by Iris »

enclave wrote: May 12th, 2018, 11:24 pm Those add-on authors can then state the recommended RNG theme to use with their mod. I feel this would be my last concern if a mod is compatible/balanced for specific type of RNG or not.. despite I made an add-on (Q CIV) which would only work allright with default RNG I don't see a problem in it.
It’s easy to say that when none of your add-ons are incredibly long single-player campaigns.

I’ll just quote myself from Discord:
shadowm wrote:I'd like to think that players should have a choice with regards to how to play my campaigns, but I can't help but feel that any kind of semi-officially-sanctioned changes devalue the amount of effort I put into testing and tweaking scenarios.

And yes, that also extends to the CTH mods (as in actual WML/Lua mods, not engine changes) that people have developed. I did end up having to deny people the ability to play my campaigns with one of them but that was really for technical reasons. (Namely one of the events it injects interferes with several situations where I need to override CTHs to 0% or 100%.)

A casual observer wouldn't realize that the amount of time I spent testing and tweaking IftU's final boss scenario is on the order of days on and off.
And most of that was precisely to account for the random nature of battle outcomes.
Author of the unofficial UtBS sequels Invasion from the Unknown and After the Storm.
User avatar
Pentarctagon
Project Manager
Posts: 5526
Joined: March 22nd, 2009, 10:50 pm
Location: Earth (occasionally)

Re: A mod to make Wesnoth's RNG a little less frustrating

Post by Pentarctagon »

Maybe I'm not understanding something about this new experimental mode or general balancing, but it doesn't seem like this should have any affect on balance? As I understand it, this mode just makes the series of hits and misses stay around the expected values for damage dealt and damage received over a shorter time scale compared to the current pure Law of Large Numbers approach. In which case, where does balance come into play?
99 little bugs in the code, 99 little bugs
take one down, patch it around
-2,147,483,648 little bugs in the code
User avatar
Iris
Site Administrator
Posts: 6797
Joined: November 14th, 2006, 5:54 pm
Location: Chile
Contact:

Re: A mod to make Wesnoth's RNG a little less frustrating

Post by Iris »

I don’t really know. What I said could be based on incorrect information from gfgtdf (who is a developer and probably looked at the code, though), especially considering that Dave (Denivarius) didn’t really address what I said above at the time, which was a follow-up to this conversation:
Screenshot_20180512_234456.png
Author of the unofficial UtBS sequels Invasion from the Unknown and After the Storm.
enclave
Posts: 936
Joined: December 15th, 2007, 8:52 am

Re: A mod to make Wesnoth's RNG a little less frustrating

Post by enclave »

shadowm wrote: May 13th, 2018, 3:45 am I don’t really know. What I said could be based on incorrect information from gfgtdf (who is a developer and probably looked at the code, though), especially considering that Dave (Denivarius) didn’t really address what I said above at the time, which was a follow-up to this conversation:

Screenshot_20180512_234456.png
what I done to my different luck mod is I cancelled different luck on all units with special abilities (poison, slow), they stayed unnaffected..

this could be done here too probably if its really making assasin 100% poison with new system.
or at least make it softer for poison slow units.
The truth is that it sounds quite realistic that if assassin has 3 hits and marksman it is really expected to do at least one hit, no? you will still not know which of them will be hit 1st or last hit.. so it's in a way unpredictable.. but I guess maybe the chance should be limited to 85-90% then? somehow...
shadowm wrote:And yes, that also extends to the CTH mods (as in actual WML/Lua mods, not engine changes) that people have developed. I did end up having to deny people the ability to play my campaigns with one of them but that was really for technical reasons. (Namely one of the events it injects interferes with several situations where I need to override CTHs to 0% or 100%.)
in this case maybe content creators could tick something to make different RNG themes disabled in host menu if they want.. but i think recommendation for RNG theme could be enough... maybe if its strongly recommended then highlight text in red...
Last edited by enclave on May 13th, 2018, 4:17 am, edited 1 time in total.
User avatar
Celtic_Minstrel
Developer
Posts: 2166
Joined: August 3rd, 2012, 11:26 pm
Location: Canada
Contact:

Re: A mod to make Wesnoth's RNG a little less frustrating

Post by Celtic_Minstrel »

I think the most logical way under this new system would be to divorce poison/slow chance from hit chance. For example, before the attack, roll once to see if poison/slow takes effect, then ignore poison/slow for the actual hits.

It wouldn't extend well to other similar abilities though, such as stun.
Author of The Black Cross of Aleron campaign and Default++ era.
Former maintainer of Steelhive.
enclave
Posts: 936
Joined: December 15th, 2007, 8:52 am

Re: A mod to make Wesnoth's RNG a little less frustrating

Post by enclave »

Celtic_Minstrel wrote: May 13th, 2018, 4:22 am I think the most logical way under this new system would be to divorce poison/slow chance from hit chance. For example, before the attack, roll once to see if poison/slow takes effect, then ignore poison/slow for the actual hits.

It wouldn't extend well to other similar abilities though, such as stun.
I feel like I probably like this idea... sounds very nice.. not 100% sure, but idea sounds good..

Again thinking of it.. it will probably be incompatible with custom addons realying on it... might even be a nightmare (I fortunately don't have enything I would care about much in this regards)
User avatar
Crow_T
Posts: 851
Joined: February 24th, 2011, 4:20 am

Re: A mod to make Wesnoth's RNG a little less frustrating

Post by Crow_T »

Celtic_Minstrel wrote: May 13th, 2018, 4:22 am I think the most logical way under this new system would be to divorce poison/slow chance from hit chance. For example, before the attack, roll once to see if poison/slow takes effect, then ignore poison/slow for the actual hits.

It wouldn't extend well to other similar abilities though, such as stun.
Yep, this was done in the mod in my sig- poison/slow used the "classic" system still.

I'm curious if one could simulate a few thousand instances of both the classic and new-alt-exp PRNG and come up with a formula to adjust the turns in a scenario for the new system (or maybe the new code will show the way).
User avatar
Celtic_Minstrel
Developer
Posts: 2166
Joined: August 3rd, 2012, 11:26 pm
Location: Canada
Contact:

Re: A mod to make Wesnoth's RNG a little less frustrating

Post by Celtic_Minstrel »

Crow_T wrote: May 13th, 2018, 2:18 pm
Celtic_Minstrel wrote: May 13th, 2018, 4:22 am I think the most logical way under this new system would be to divorce poison/slow chance from hit chance. For example, before the attack, roll once to see if poison/slow takes effect, then ignore poison/slow for the actual hits.

It wouldn't extend well to other similar abilities though, such as stun.
Yep, this was done in the mod in my sig- poison/slow used the "classic" system still.
Uhhh, that isn't what I suggested at all? Unless you're using "classic" in a weird way.
Author of The Black Cross of Aleron campaign and Default++ era.
Former maintainer of Steelhive.
User avatar
octalot
General Code Maintainer
Posts: 783
Joined: July 17th, 2010, 7:40 pm
Location: Austria

Re: A mod to make Wesnoth's RNG a little less frustrating

Post by octalot »

Celtic_Minstrel wrote: May 13th, 2018, 4:22 am I think the most logical way under this new system would be to divorce poison/slow chance from hit chance. For example, before the attack, roll once to see if poison/slow takes effect, then ignore poison/slow for the actual hits.

It wouldn't extend well to other similar abilities though, such as stun.
I'm not sure that follows - if someone's complaining that assassins sometimes fail to poison the target, having a separate RNG roll for whether they poison the target seems to the opposite of the point of the mod. What if Dave just documents that assassins will deterministically always poison the target with this mod enabled?

I think that wouldn't extend to the slow ability though, as can be important to look at the odds of which attempt adds the slow, and how many unslowed counterattacks could hit.
User avatar
Celtic_Minstrel
Developer
Posts: 2166
Joined: August 3rd, 2012, 11:26 pm
Location: Canada
Contact:

Re: A mod to make Wesnoth's RNG a little less frustrating

Post by Celtic_Minstrel »

Ah, good point on slow... but still, I think poison always taking effect could have quite a significant effect on game balance...
Author of The Black Cross of Aleron campaign and Default++ era.
Former maintainer of Steelhive.
Post Reply