[Request]Reduce RNG influence

[Gambit]

Don't forget that when someone comes up with an actual plan you can submit that to the task section on GNA and create a new forum thread for more specific feedback/discussion.

[Zarel]

I like those design goals, and I'm also fine with Sauron's mod. It seems like a variation on my minimum damage idea. I'd prefer to have a flat minimum damage, though.

[StevenAus]

How can I help? =) Btw: I really liked that

http://www.wikihow.com/Practice-Nonviol ... munication

link.

Best regards,
Steven.

PS: To all - have a great day! =)

[Gambit]

It was a great guide, and a good thing to keep in mind.

Anyway; Zachron you seem to have thought your karma system out quite well. You know exactly how you'd like it to work. It might be time for an actual proposal.

[Mabuse]

more strikes = less luck

[Daedal]

...RNG-algorithm apply to campaigns. Any ideas on that?

I don't play MP, but as I understand it there are going to be two main differences between MP and campaign play: units in campaign play carry over into future scenarios and this is a big part of campaign balancing, and the sides in campaign play are always fixed while MP play might give you unexpected matchups.

I suppose that if normal BfW is balanced for both MP and campaign play there's no reason that reduced luck BfW shouldn't be as well, but it won't be a non-trivial thing to accomplish and may require a different power distribution among units than we see now in BfW.

[Gambit]

Most new RNG systems won't be balanced for sure. Default is balanced for the current system.
It gets even harder when you consider that they'd like multiple algorithms for the fork. I'm not so sure if balance is an issue as long as you give a little disclaimer. Fabi or Yogi would have to elaborate more here.

[Yogibear]

...RNG-algorithm apply to campaigns. Any ideas on that?

I suppose that if normal BfW is balanced for both MP and campaign play there's no reason that reduced luck BfW shouldn't be as well, but it won't be a non-trivial thing to accomplish and may require a different power distribution among units than we see now in BfW.

That might be depending on the algorithm used, but in most cases it won't need any effort at all.

You might think of it this way: Without luck, everything is perfectly balanced. Factions in MP have equal strength and campaign scenarios are also fine. The more luck you experience, the greater an influence it has on balance - either overpowering or underpowering you. That is essentially the reason, why competitive mp games (like the ladder) are not played with 30% XP settings: That just throws the balance off too much.

Taking this point of view, every action to reduce the influence of the RNG should automatically enhance balance. There might be exceptions (like for example if an algorithm gives you opportunities to exploit it to your favour), but in general this is the case. So i am very confident that we are not having to deal with balance much (if any).

[chains]

Taking this point of view, every action to reduce the influence of the RNG should automatically enhance balance.

This might be minor, but removing RNG isn't directly enhancing balance, but rather making balance more dependent on faction, map, and unit balance. So, if one faction is better than another on a certain map for example, then the advantage would hold in all games.

I would really love a true stats project to come alive so we can see how this actually plays out in the community. Some factions I expect are really balance for pro players and really unbalanced for new players.

In Warcraft for example, shamans were really over powered in early releases of the game for pro players who executed the strategy well. But, the rest of the player base was in the dark.

[Daedal]

I'm not convinced that reducing the RNG influence will automatically result in a balanced game.

Consideration #1 - Shamans. Awesome en masse despite their apparently poor attack power due in large part to their slow attack. What happens if they start hitting more reliably? This isn't even an issue of unit doing more or less damage, it's just the slow effect hitting more that will change game balance.

Consideration #2 - Units with high # of hits or extremely low # of hits. The dwarvish thunderer, for example. I can't speak for these guys so much because I'm never able to use them very effectively (selective memory, heh), but I do recruit a few whenever I can just to gamble with. What happens if their single attack were to hit more reliably? On the other end, most of the staple melee units have 3-4 hits specifically because, I would assume, it helps negate some of the wonkiness that goes with having one or two shot units. If other units become more reliable, will these staple units need to be buffed so they remain staple units for story/flavor purposes?

It gets even harder when you consider that they'd like multiple algorithms for the fork. I'm not so sure if balance is an issue as long as you give a little disclaimer.

I agree that multiple algorithms will make this process more difficult unless Yogi is right after all and the game balances itself out. Only time will tell that one. Not sure I agree that balance is no issue if a disclaimer is given, though. What's the point of playing the game if it becomes unbalanced? You can just pick the better side and win. I am curious to see this reduced RNG fork, but I'd rather see fewer RNG variants that are all balanced rather than a whole lot of unbalanced variants.

[Yogibear]

I'm not convinced that reducing the RNG influence will automatically result in a balanced game.

Consideration #1 - Shamans. Awesome en masse despite their apparently poor attack power due in large part to their slow attack. What happens if they start hitting more reliably? This isn't even an issue of unit doing more or less damage, it's just the slow effect hitting more that will change game balance.

Consideration #2 - Units with high # of hits or extremely low # of hits. The dwarvish thunderer, for example. I can't speak for these guys so much because I'm never able to use them very effectively (selective memory, heh), but I do recruit a few whenever I can just to gamble with. What happens if their single attack were to hit more reliably? On the other end, most of the staple melee units have 3-4 hits specifically because, I would assume, it helps negate some of the wonkiness that goes with having one or two shot units. If other units become more reliable, will these staple units need to be buffed so they remain staple units for story/flavor purposes?

As i said, it depends on the algorithm. If you take sauron's mod for example, everything will remain the same. Even in deterministic mode, shamans won't slow more often than they do at the moment (slow is still subject to the RNG). Against 50% defense, the thunderer will do 9-1. So instead of one hit one miss for two subsequent attacks, the thunderer now will do 9 damage each time. Shouldn't affect balance at all.

On the other hand, you can of course design an algorithm in such a way that balance will be affected. This is what we have to discuss and hopefully prevent.

[pauxlo]

Against 50% defense, the thunderer will do 9-1. So instead of one hit one miss for two subsequent attacks, the thunderer now will do 9 damage each time. Shouldn't affect balance at all.

I'm not this sure about this.
For example, imagine a (single) thunderer trying to remove a single goblin spearman from a (normal) village (suppose there are no other units around, and the goblin does not attack himself).

Here are the probabilities for the goblin to be dead after x turns of attack (rounded to whole %):

Code: Select all

turns     now   deterministic
   0       0%          0%
   1      40%          0%
   2      64%          0%
   3      78%          0%
   4      87%          0%
   5      92%          0%
   6      95%          0%
   7      97%          0%
   8      98%          0%
   9      99%          0%
  10      99%          0%
  11     100%          0%

(I hope I didn't miscalculate ... the last "100%" really is about 0.99637.)

Simply, the chance to do really much damage is sometimes more worth than every time doing the expected damage.
This not only matters when trying to get someone off a village, but for example also for breaking through a line, when the defenders can rotate away and the attackers can't attack from more than two positions: being able to hit with one or two thunderers really helps (even if they do not hit every time).

[grrr]

I like those design goals, and I'm also fine with Sauron's mod. It seems like a variation on my minimum damage idea. I'd prefer to have a flat minimum damage, though.

Woah what?! Another wesnothian who sees the need and elegance of min dmg? I once even made an era with the idea that once I get the stupid % dmg modifiers out of the way, I could easily implement minimum dmg on top of that (my example has always been: "Grunt 2-9 x2 melee dmg", meaning if it misses, it still does 2 dmg, see http://forums.wesnoth.org/viewtopic.php?f=15&t=23160 for the era).

[Zarel]

I'm not convinced that reducing the RNG influence will automatically result in a balanced game.

And the elegance of minimum damage enforcement and Sauron's system is the same - it will have none of the disadvantages you describe.

[Death]

I saw someone mention something about "no game has ever done a karma system," well, I know of one that has, and very successfully.

Go take a look at Tactics Arena (it's free). It uses a 'karma' system different than the ones described here, but the basic goal is the same: avoiding localized 'luck spikes'.

The difference is that instead of having a karma value for the attacker, it's used for the defender, ie. take the defending unit's %-to-be-hit, and modify it based on previous hits or misses accrued. (blocking is distinguished from misses)



that's all i can contribute. i'm having a tough time following this (nifty) thread.