My view on RNG

[eddieballgame]

Have recently "rediscovered" this fine game. Am very impressed with all the hard work put into Wesnoth.
I have been reading a lot of the comments on the pros & cons of Wesnoth's randomness in battles. Personally, this is one aspect of the game that makes it so much fun.
As an avid Backgammon player I am quite aware of the "cruelty of the dice". As most serious BG players know, BG is a game of "position & probabilities. Thus you always have to be aware & willingly ( not always! ) to accept a bad result even in a superior position. One of the few games where a weaker player can defeat even a World Class player. However, in the long run the better player is always favored, even if so slightly.
Winning is fun, but in Wesnoth...losing can be fun when you realize it is just a game & you lost not because you did not play well, but for those damn dice.

[Tom_Of_Wesnoth]

Without the element of luck Wesnoth wouldn't be, well, it wouldn't be Wesnoth.

[ancestral]

This has been discussed to death in the past, but I’ll throw in my quick take right here.

In battle, historically, as a commander, you have little control over all the variables. Often you cannot be perfect, and you’ll make decisions that don’t always pan out. That’s war for you. Nondeterministic and never predictable. What makes a great officer is one who accounts for the unknowns, and take calculated risks. That, I believe, is the secret to playing The Battle for Wesnoth.

[Tom_Of_Wesnoth]

I remember seeing a era a few years ago that took out the RNG. Is it still around for those who want it?

[tekelili]

This has been discussed to death in the past.

I agree. The reason why this debate always becomed "ethernal" is because there is always someone that tries to create cathegories of "good games" and "bad games" based in the axiome that "a good game should always have the best player as winner". Such axiome is so obsolete like state "a good geometry should never allow pararels meet each other".

Games with probability involved in outcomes are like fuzzy logic: You can do "right operations" and still be uncertain about the value of a proposition. Some people will claim that a system where you can´t know if a proposition is false or true (even knowing initial values of all variables) is not a valid system, but that is just lack of knowledge about what is a "valid system" imho. Those people often use to have wrong ideas about what is a "deterministic enveiroment" claiming that BfW is not, when in fact BfW (as fuzzy logic) is deterministic, because the value of variables (probabilities) is always determined by system rules and is never "unknown".

Imo, a good game is one that allows better players take better decissions, and for sure BfW allows that. An example of bad game would be roulette.

[iceiceice]

This has been discussed to death in the past.

I agree. The reason why this debate always becomed "ethernal" is because there is always someone that tries to create cathegories of "good games" and "bad games" based in the axiome that "a good game should always have the best player as winner". Such axiome is so obsolete like state "a good geometry should never allow pararels meet each other".

Games with probability involved in outcomes are like fuzzy logic: You can do "right operations" and still be uncertain about the value of a proposition. Some people will claim that a system where you can´t know if a proposition is false or true (even knowing initial values of all variables) is not a valid system, but that is just lack of knowledge about what is a "valid system" imho. Those people often use to have wrong ideas about what is a "deterministic enveiroment" claiming that BfW is not, when in fact BfW (as fuzzy logic) is deterministic, because the value of variables (probabilities) is always determined by system rules and is never "unknown".

Imo, a good game is one that allows better players take better decissions, and for sure BfW allows that. An example of bad game would be roulette.

+1

Imo the *most* interesting thing about wesnoth (at least mp strategy) is how to manage the luck and how hard it turns out to be (at least for me). In games like risk or civilization, things are very binary, each unit-on-unit has some randomness but not really that much. In wesnoth, esp. in mp games, there is some significant luck, but the interesting part is that the luck is not contained, it spreads everywhere in some sense... if I have loyalists fighting rebels for instance, if I have bad luck on one side of the front, it may force me to use a horse on one side when I would rather have used it on the other side, but maybe I should use a horse on the right first just to see if I get more hits before I decide to shift units... and is there a wose in that tree? hmm I wonder... These questions and considerations can rapidly get extremely complex it seems.

I don't know but my sense is that if you carefully arranged battles in wesnoth with arbitrarily large numbers of units, you could make it so that an individual piece of randomness from a strike at one side of the battle is still highly strategically relevant even to units on the opposite side of the battle, (for a master class player maybe), so the relevance of luck can travel arbitrarily far ... My impression is that this isn't true of other games. It reminds me of the strange things physicists say about supercritical fluids that achieve unbounded interaction range... then they start to do things like start boiling suddenly when poked, or climb up the walls of a container ...

Maybe I'm wrong though, I'm not really very good at this game

Spoiler:

foo
foo
foo

(note that I am not a physicist nor did I claim to be :p )

[pyrophorus]

I remember seeing a era a few years ago that took out the RNG. Is it still around for those who want it?

Yes, the "No Luck Era" and "No Randomness Mod" are available on the addons server.

[Pentarctagon]

Merged the "Position & Probabalities" thread with this one, since they're both about the RNG.

[Sacho]

Apologies for the lack of structure to my post, it was mostly stream of conscience Also, only halfway through it I realized I should be using BB codes rather than markdown emphasis.

1. Players who complain don't want you to tell them that they probably play wrong, or that unit xyz is replacable or expendable. They get even more frustrated because nobody acknowledges that the RNG can screw you in such unlikely ways, that it doesn't seem fair and, in fact, isn't.

Okay, here you go - I acknowledge the RNG can screw you. When you *minimize*(note - this is not *remove*) bad luck, you reduce the situations where RNG can screw you, but sometimes there's nothing you can do. Then you have to reset. That can feel very frustrating. However, you need to also acknowledge all the *GOOD* things the RNG does(the uncertainty factor making a fairly simple rules-wise game be very deep and strategic). It seems like you're more leaning towards "remove all frustration" from the game, but this also tends to remove most challenge, and as such, makes the game boring. Again, the RNG screwing you is not fair, it can happen, sometimes(very rarely) there's no way around it, this is all balanced by all the *positives* the RNG brings.

2.1 If there is an enemy land-unit in shallow water (20%), if I attack, it is unlikely that I will hit more than once in 3 attacks, very unlikely that I will hit more than twice. It is also unlikely that the one attack that connects will be the first, it will rather be the third (so even a 1/73 hp unit in shallow water will damage the attacker). If I place any unit exposed in shallow water, the enemy will most likely hit every single time, be it 3, 5 or 8 attacks.

Uhh....by the nature of the RNG, none of these statements are true. Like a further poster mentioned, you are probably seeing perception bias at work here. As an aside though, players *can* have unreasonable expectations of a 80% hit chance. For example, hitting *all three* attacks is an unreasonable expectation - it's about a ~50% chance. Luckily, wesnoth produces *awesome* graphs of your expected damage(this is probably my favourite feature of the game).

-snip-

Okay, rather than stating the same thing over and over, I'll try to assume that you mean these are the *rare* bad situations where the RNG screws you. But your conclusion doesn't seem to support that. I'm sure we all have horror stories, but actually, the more I play wesnoth, the more I find myself getting *lucky* rather than *unlucky* - as I've probably grown more pessimistic in my assessment of what my units can do. In fact, I more often notice how that elven fighter I throw at the enemy as a sacrifice makes a heroic stand and lets me completely steamroll the now poorly positioned opponents. It's true that when you play conservatively, the situations where the RNG screws you will probably be more common than the situations where the RNG suddenly turns a losing situation into a winning one(since you'd be trying to preempt having losing situations in the first place) - but at the end of the day, over a long period, the statistics window shows that I have average "luck", and I'm fine with that.

3. The true Problem with the RNG is that it follow a pattern which may even be calculated accurately but goes through any unit, enemy or not. What I mean by that is this:

-snip-

This is *NOT* how an RNG works at all. It's a common misconception that an RNG "balances" bad rolls by providing good rolls, but this is not the case. There is no guarantee that after you miss 10 times in a row, you(or your opponent) would suddenly get good rolls - in fact, the coin tossing analogy is absolutely correct - the Nth number the RNG generates has no relation*(not technically, but practically..) to the N+1th. If you fail to hit a unit with 4 mage attacks, you CANNOT assume that your next mage will hit. This is *wrong*. This is *not* how RNG works. Your next mage will have the *same* chance of hitting, regardless of how many times you miss first and how many goats you sacrifice. This misunderstanding of the law of large numbers is probably why casinos are so successful.

4. Even if you're the most talented of logical dudes or dudettes on the planet, when you are beginning to figure out this game, the RNG can screw your better approach.
-snip-
So, in conclusion, a good move/strategy/tactic you figured out can appear as unfavorable and maybe, because of that, you won't try to do this again, even if it was the most right thing to do in the given situation.

In general your statement is correct, of course - it's basically stating "you can't know the chance% of a random event from a single experiment". This is absolutely true. However, your example actually shows that you can use your knowledge of probability to determine that your strategy was indeed, solid, and you were just unlucky - instead of changing it, you decided to do the same thing again.

You can calculate the odds and compare them to the odds of another approach, and determine which one is better - no matter if playing it out works or doesn't. And over a large number of games, you can tell whether your strategic approach is more successful - when you start winning campaigns more often, when you have to reload less and less often, etc.

5. And lastly, even if for 1 scenario the RNG does not screw you, it still feels unfair, but not because of the RNG. Realize this:

5.1 The (computer-)enemy does not give a flying hell if he dies. He can do the stupidest riskiest moves and succeed, which, if you would attempt them, would cost you precious high-level units you might very well need in the next scenario or would lead to you just plainly loosing the scneario. The Enemy does not care about EFB either.

5.2 The enemy usually outnumbers you, which means that the enemy just gets more RNG-Rolls and therefor an initially better chance to succeed at anything. (Ever fought an army of level-0 goblins, or zombies, and had really bad luck?)

5.3 A turn-limit is ALWAYS in favor of the enemy. He can ALWAYS play slow, you ALWAYS have to be fast(er) (except the rare "defend" maps)

These things don't seem like complaints about the RNG at all, and you can bring similar complaints against almost any game of the genre.

Realistically, writing an AI that is either aggressively hunting single kills, or passively defending via superior numbers is where the genre is stuck at. Both tactics are ways to artificially(hehe) inflate the difficulty in order to make it more interesting for the player. The campaigns are all developed with the AI's attitude in mind. If at some point, we manage to create AI that can match a player's ingenuity, you'd probably see campaigns of equal numbers and objectives, but currently, the AI *needs* these handicaps in order to present a challenge. I think this is an interesting discussion to continue(both in how to develop campaigns and how to make the AI better), but it has no bearing to RNG, really. You could have a chess AI that will risk 5 pieces to capture 1, and it would present the same frustration if players were really attached to their pieces.

There are other points to this, but you get the idea.


These all are the reasons why this game is

A) Hard to get into for someone who starts out un-lucky (When I started playing many years ago, I lost the tutorial. twice.)

This seems possible but you need to provide more concrete proof. After all, wouldn't a player that starts out lucky(let's say wins the tutorial on the first try), but then gets unlucky after quit just the same? In general, if the RNG provides a frustration to you, then you'd be frustrated playing Wesnoth.

B) leads players to save-abuse the hell out of the RNG

Or challenge themselves to develop rng-resilient strategies? I admit I completely understand your frustration. When I first started playing Wesnoth, I would get emotionally attached to units, I'd not want to lose (almost) anything, and I'd frequently save-scum to avoid poor results. Eventually though, I disciplined myself to treat poor RNG as an extra difficulty, trying to power through scenarios and even campaigns regardless.

There's an interesting split here to examine, in my opinion, but it is handled very well by Wesnoth. If you just want to "progress" through a campaign(see the story, develop characters, etc), Wesnoth provides you with good tools - you can start at a low difficulty, and the save features are immaculate(fast save/loading, you can save at practically any time, etc - there's many games that have poor save/load functionality and fixed difficulty!). Then if you treat campaigns as just consecutive scenario puzzle challenges, you might be fine with taking on a higher difficulty and "taking in stride" unexpected losses, trying to finish despite them.

Somewhere the ones who never had an issue in the beginning must admit that they were lucky, that their enjoyment of the game today might very well not be if the first 5 scenarios they ever played gave them an RNG-middle-finger.

These days, as I try out(or more often, watch reviews/let's plays) of games, a game only piques my attention if I/the reviewer loses the game and is forced to reload multiple times. I've often quit out of games which don't present any challenge in the first few scenarios, so my position is opposite yours, for what it's worth. I distinctly remember failing the HttT campaign multiple times(on the first few missions) before I ever managed to progress through it.

Even disregarding my complete opposition to your argument, what you're describing is very, very unlikely. As long as players pick lower difficulties when they struggle, I think Wesnoth provides a wide spectrum of campaigns to suit practically *anyone*. I also want to really bold this point, because I think it's very important - the way Wesnoth provides easy access to statistical data makes losing due to RNG very easy to identify, unlike other games, where it's hard to tell that you're getting screwed by the RNG because none of the formulas are available to you. I would much more likely quit one of those games, after I employ what I feel is a solid strategy and mysteriously lose anyway, rather than Wesnoth, where the mystery is immediately revealed - it was the RNG.

It also seems to be very obvious that the RNG is faulty if you consider, that save-loading as a practice works and can lead to hilariously lucky replays. Just imagine that save-loaded replay stacked against the player and you'll get the Idea why some are in rage when they play and create accounts like "Ihatewesnoth" to [censored] and moan. I believe them. Every single word they say. I believe that they started the game open minded and had an experience leading them to believe, that there is no enjoyment to be had.

I'm not sure what the fault of the RNG is here? Seeding the RNG differently after each load is intended(exactly to enable this practice!).

And I know that it has nothing to do with having played Fire Emblem, or FF-Tactics, or Tactics Ogre or even Yu-gi-oh (the GBA/NDS games), which I all played before and after. Those games also screw you on RNG time and time again, but it doesn't appear as frequent as to lead players to [censored] and moan about those games.

Now that's complete BS. Hell, even in this thread, people talk about the stat growths of Fire Emblem making characters unusable, and having played all of these, I've read piles upon piles of threads about how unfair the RNG can be. This is not a feature specific to Wesnoth - some people are really frustrated by RNG, some aren't. If you are really frustrated by it, you can use one of the zero RNG mods! This is the beauty of an open source, moddable game, like Wesnoth. Meanwhile, criticizing the vanilla game for catering to a certain playstyle is what's not fair - it was never intended to please *everyone*.

Especially young people could use a more fair wesnoth, one which has a build in "fail-counter" which triggers one unit to hit with all attacks once, when the fail-counter reaches a certain number (all this hidden, of course.) Call it "Wesnoth: The childsplay version". Oh well, after all this years there is one thing I know for sure.

Aha! This is actually a very interesting suggestion. It's a biased RNG implemented in games like warcraft 3. I think it's an interesting mod to explore, but it would significantly complicate the vanilla game(you would have to constantly change the %s for each attack). I think that would be fairly confusing to new players, but I'm really interested in a possible modification for it. I have a few good friends who have told me that they don't like playing Wesnoth specifically due of the lack of a biased RNG, because having one vastly reduces their frustration in dealing with it.

Please take no offense out of this, I know how much work this game is for alot of you, and I know that there is a consistent fan-base which adapted to, understands and enjoys the game. It just pains me in my chest everytime when there is no Understanding to be found whenever someone does not enjoy it, be it because he's young or unlucky.

I think this is disingenuous. On this board, whenever RNG is discussed, there is plenty of sympathy for players who show situations where the RNG screws you over - after all, we all experience them! But you treat the opposition towards removing RNG as a lack of understanding. I completely agree that the RNG can screw you, but I do not agree that this warrants its removal. I also completely understand if this makes the game frustrating and less enjoyable to you - but I also know I don't enjoy whole categories of gaming, yet I don't go there to tell them how their game must be tailored to my tastes instead. Plus, Wesnoth even provides possible remedies with modding!