Community The New World Design Poll #2: RNG

[Vault Dweller]

So when are you being lied to: when your THC says 70% but your effective THC is 50% in your last fight because that 70% was based on a hundred attacks or so not twenty or when your THC is 70% and you know that you'll hit precisely 7 times out of 10?

Assume you group THC in sets of 10, and at the beginning I have 70% THC. After 9 attacks where I hit 7 and missed 2 there is only one result for the 10th attack: miss, because that's what's left. so the game is showing me 70% for my last attack, but my actual THC is 0%

You're making the same error as those who complain about 3 consecutive misses by focusing on individual attacks rather than your average. The only question here is how many times you have to attack to get to that average. 100? 1000?

I wonder how many people who ask for truly random rolls would be ok with missing 3-5 times in a row in-game. This isn't challenge, just bad luck.

 

[Rinslin Merwind]

Devs should rig RNG ofc, because in another case there can be situation like in XCOM, where well trained soldier can't hit alien which staying in 50cm from him. These situations ruin immersion imo.
When chance to hit is 70%-80% there shouldn't be a chance of critical miss (when you shoot yourself in leg or gun explodes in your hand), because it's would be ridicule player.

In my opinion "classic" RNG good mostly in tabletop games, where GM can save your ass in funny way and just change how whole adventure flows. CRPG programmed and just can't randomly change situation in favor of player if he fucked up.
Example: Two monks attacking player at start of game, character is trained gunslinger and he trying to shot monks with 90% probability and critically missing (that's "true" random for ya). Things which can happens in tabletop: "shit session too short, I let you live because one monk recognized you as his son, which he thought died long ago"- GM decided to save player. Things happens in CRPG: "lmao you dead" - game decided that you gonna die because "you didn't have luck" to shot some crazy bastards from 2m distance.

 

[Mr. Hiver]

The real issue with the usual RNG based on percentages is that it does not accurately presents complexity of combat.

Its a gross oversimplification, that works as the exact same calculation done each time you attempt to do something - resulting in the same binary outcome - regardless of previous attempts and results.

And there is another thing. It seems the Universe mathematics are broken, because its probabilities that are around 70% that provide most worse repeated results!
Same god damn thing happened in DOS at 73% chance - all the fucking time. Consecutively and repeatedly - while i have never manged to get such strings of consecutive misses with any other percentage. And i tried. A lot.
There 73% seemed to be some weird black swan spot and ive noticed the range for most consecutive misses is close to 70% in other games too.
So someone should fix the fucking quantum.

Until that is done i dont see any problem with adjusting the dumb oversimplification of RNG - in a sensible manner that will still keep the importance of skills and simulate chances. (unlike what Larian did)
The proposed solution seems alright to me, especially if its explained more sensibly, as a part of actual setting and combat - i.e. if we are fighting each other and i miss because you have done some specific move, i will try to compensate for it next time, - so in simple terms, every single move in combat is not happening in complete vacuum separated from anything else.

If its for any reason problematic to just introduce such solution by itself, then another passive skill can be added to produce same results, as something the player can invest points in and so shape and evolve intentionally.

Additionally, i think PoE system did well in that sense with its grazes mechanics, which ameliorate that problem of repeated consecutive misses with high skills.

 

[Darth Roxor]

These situations ruin immersion imo.

oh no wont anyone think of teh immershun

When chance to hit is 70%-80% there shouldn't be a chance of critical miss

there should

Two monks attacking player at start of game, character is trained gunslinger and he trying to shot monks with 90% probability and critically missing

except in any non-retarded pnp the gunslinger is going to have

f a t e p o i n t s

of some sort to get himself a reroll

GM decided to save player

your gm is bad

your players are bad to leave the gunslinger exposed like that

and finally

"you didn't have luck" to shot some crazy bastards from 2m distance.

you are bad to put yourself into an exposed situation like this


diagnosis: people who whine about rng are all fucking bad

Things happens in CRPG: "lmao you dead"

press F9 to reload and consider being less bad in the future

 

[Trashos]

You're making the same error as those who complain about 3 consecutive misses by focusing on individual attacks rather than your average.

Yes, we are talking about 2 different things (average vs next attack). However, when I hover over the enemy and see "THC 70%", I expect that my NEXT attack has 70% THC. Isn't this what everyone would have expected?

The only question here is how many times you have to attack to get to that average. 100? 1000?

The only truly correct answer is "infinite". Or at least a huge number which simulates infinity. If that's not what you want to do because of players being frustrated with consecutive misses, that's another matter, and you may be on the right. However, you then have to accept that the in-game information on THC is lying, because it is.

I wonder how many people who ask for truly random rolls would be ok with missing 3-5 times in a row in-game. This isn't challenge, just bad luck.

I get it that many people don't understand statistics, but expect a different attitude from the mathematically-inclined. On the other hand, I didn't mind the AoD trick when I didn't know it existed. Maybe one idea is to use a large enough set that cannot be easily exploited (not 10, say 30). Rest assured that I will still count hits/misses, but other people may not.

Additionally, i think PoE system did well in that sense with its grazes mechanics, which ameliorate that problem of repeated consecutive misses with high skills.

I hate PoE's graze system. "Death by 1000 papercuts", as other Codexers have put it. On the other hand, maybe it will work better on TB. On RTwP, I downright hate it.

 

[Mr. Hiver]

It was just an example of how to solve the issue in a softer way, not a suggestion to copy it verbatim.

Ive argued with morons who want to see RNG removed from RPGs too much already.
But that doesnt mean that the current systems cannot be improved.

Especially because the current system is a gross binary oversimplification of something that in RL has multitudes of shifting and mutually affecting components.

To repeat the actual suggestion i am making, it seems to me that adding an additional passive skill players could invest into - that would achieve results V. Dweller described, would be a good way to go about things.
Either that or make the calculations of outcomes much more complex.

After all, its not just about what you do, but what abilities the enemy has too.
- and terrain, and distance, speed, agility, strength, constitution, atmospheric conditions, intelligence, previous experience, and on and on and on.

edit:

The combat is a messy thing, its not supposed to be represented by a pure mathemathical ideal that produces binary yes or no outcomes.

 

[Blakemoreland Hybrid Boss]

I'm pretty sure one of the RPG lectures hosted by Tim Cain addressed this topic. Part of one proposed solution was to let the first player attack always hit.

Your displayed To-Hit chance doesn't need to be perfectly mathematically accurate. Mathematics shouldn't trump psychology.

Popamole psychology shouldn't trump mechanics. Fuck this decline.

Provide a detailed (but optional) tutorial. Provide a combat testing arena (also optional). Same as in AoD. Provide hints and tips. Make them visible only if a player checks a box in settings.

Players don't like tutorials on the side. The tutorial should be in the start of the game. Banner Saga got this right.

And finally, provide a list of console commands. For those who think the game is "unfair". Let them decide what "fair" is.

 

[Blakemoreland Hybrid Boss]

Notice how players keep using Tim Cain's course on decline and popamole design unironically. Imagine the damage he is doing to the genre.

 

[Rinslin Merwind]

oh no wont anyone think of teh immershun
there should
except in any non-retarded pnp the gunslinger is going to have
f a t e p o i n t s
of some sort to get himself a reroll
your gm is bad
your players are bad to leave the gunslinger exposed like that
and finally
you are bad to put yourself into an exposed situation like this
diagnosis: people who whine about rng are all fucking bad
press F9 to reload and consider being less bad in the future

Well it was my opinion and I didn't pretend that my opinion is absolute.
However I think your argument is rather flawed, since the whole point of fate points that there should be restricted number of fate points in whole game and if situations would be like this (and not only "bad players" can face a fucking ambush on street, it can be scripted) happens a lot it would be rather frustrating.

With your bad/good argument you only managed to prove my point that RNG doesn't give fuck about players skill, bad/good/ugly it's just happens despite physical laws and simple logic.

I mean if you invest in guns you should be able to shoot someone from same distance where even toddler can hit target.

Even if you miss on short distance, you not should be able to shoot yourself in leg, because you have stats in guns and had solid chances to hit target you not just idiot with wrong build. "F9 reload" method shouldn't be spammed in game which position itself as RPG where exist some kind of strategy, not just random encounters.

 

[Mr. Hiver]

Get him, bachelor.

 

[SausageInYourFace]

I have no opinion on this as I am to dumb to comprehend this probability shit. Just make a good game. Thank you for your understanding.

 

[Vault Dweller]

Yes, we are talking about 2 different things (average vs next attack). However, when I hover over the enemy and see "THC 70%", I expect that my NEXT attack has 70% THC. Isn't this what everyone would have expected?

But do you also expect to miss 3-4 times in a row with a 70% THC while taking damage every time the enemies attack? The question isn't about statistic or probabilities but combat design.

The only truly correct answer is "infinite". Or a huge number which simulates infinity. If that's not what you want to do because of players being frustrated with consecutive misses, that's another matter, and you may be on the right. However, you then have to accept that the in-game information on THC is lying, because it is.

Depends on what THC is. If we take it as your average from 10 or 20 attacks, it's deadly accurate. To have 70% THC, you MUST miss 30%, otherwise your effective THC is 100%. Consecutive misses are a separate issue here. The most important one is representing your THC is accurately. If there's no practical difference between 50 and 70% because "the only truly correct answer is infinite", then raising combat skills from let's say 3 to 4 becomes kinda meaningless, UNLESS it increases your batting average in the course of one fight not the entire game.

I wonder how many people who ask for truly random rolls would be ok with missing 3-5 times in a row in-game. This isn't challenge, just bad luck.

I get it that many people don't understand statistics, but expect a different attitude from the mathematically-inclined. On the other hand, I didn't mind the AoD trick when I didn't know it existed. Maybe one idea is to use a large enough set that cannot be easily exploited (not 10, but, say 30). Rest assured that I will still count hits/misses, but other people may not.

I agree that 10 might be too small a set. I was thinking 20 but 30 might work too.

 

[Drowed]

So when are you being lied to: when your THC says 70% but your effective THC is 50% in your last fight because that 70% was based on a hundred attacks or so not twenty or when your THC is 70% and you know that you'll hit precisely 7 times out of 10?

Assume you group THC in sets of 10, and at the beginning I have 70% THC. After 9 attacks where I hit 7 and missed 2 there is only one result for the 10th attack: miss, because that's what's left. so the game is showing me 70% for my last attack, but my actual THC is 0%

You're making the same error as those who complain about 3 consecutive misses by focusing on individual attacks rather than your average.

I can't agree that it's the same here, VD. A person who claims foul to miss 10 consecutive times with a chance to hit 70% is complaining about the fact that a random chance is occurring at random - and that doesn't make sense, at least mathematically. A person complaining that they will necessarily and invariably miss a chance to hit of 70% is rightfully claiming that the event will not be random - that last miss has already been determined and nothing can change that. It is essentially the opposite of random.

You are literally creating a model where past events interfere with future events: the fact that you can only miss 3 times out of 10 with a 70% THC means that if you have already missed 3 times in 7 tries, you will necessarily hit the next 3. You can be 100% sure of the future results, and if you can be sure of the results, this is literally the opposite of the definition of random.

The only question here is how many times you have to attack to get to that average. 100? 1000?

Random? I mean, the point of being random is that, precisely, it's impossible to determine how long will take to the average to show itself. It could be in 10, in 100, or 100000 or even billions of billions of tries.

I wonder how many people who ask for truly random rolls would be ok with missing 3-5 times in a row in-game. This isn't challenge, just bad luck.

This I fully agree with. The point here for me is a design issue: it's not the semantic question of trying to change the meaning of what would be "random" or not, but to decide if it is good design to use a really random system on a THC. How much luck or chance should influence the outcome of a battle. I don't see any problem in using a formula that determines that the averages will always be expressed in "10 events", for example, as long as this is clearly described in the mechanics of the game. The whole point for me isn't the exact formula that you will use in the end but rather that it should be clearly described to the user.

 

[Binky]

Players don't like tutorials on the side. The tutorial should be in the start of the game. Banner Saga got this right.

Can you skip the tutorial in Banner Saga or are you forced to play it every time you start a new game? If you can't skip the tutorial they didn't get it right. Can't speak for "players" like you do, but I don't appreciate having my time wasted by being forced to play through a tutorial every time like I'm some amnesiac.

As for console commands, I don't see the problem. Label them cheats, cheat mode, fuck around mode, mods. Lock achievements if you use them, mark players that use them, etc.

 

[Vault Dweller]

I'm not arguing the definition of random. I'm asking what your THC is supposed to represent. One answer is that it's merely a modifier in a random draw. Another, equally valid and much more practical answer is that it is your average and then you simply must hit X number of times and miss Y number of times to maintain that average.

 

[undecaf]

Should we rig the RNG to meet players’ expectations?

Absolutely not.

 

[Blakemoreland Hybrid Boss]

Should we rig the RNG to meet players’ expectations?

Absolutely not.

What if the players' expectations are averages that are more realistic? Players shouldn't be able to have three misses in a roll with 70%. That's not how these things work in the real world. I got it now. This is a very subtle issue. It makes perfect sense to modify THC. I will change my vote.

If Vince end up implementing this, it will not be a consolation for popamole players either, since they can't make a decent build to save their lives.

 

[thesheeep]

I'm all for rigging, especially in single-player games.
If you can make the rigging "reliable", it could still be used in MP games, as well.

I prefer the method of "counting together" probabilities and when too many misses occured, the next one is a hit no matter what.
I.e. if the accumulated percentage reaches 200%, the hit is guaranteed. So four misses in a row with a 50% chance would mean the fourth one would be a guaranteed hit, 10 misses in a row with a 20% chance, etc.
This 200% could also be subject to difficulty rules. Make it 100% on easy mode or something like that.

I don't give a fuck about the mathematical definition of "random". It has to be fun.
Missing 3 times in a row with 70% is not fun. That is all that matters.
And of course, let players enable/disable it in the options. This should be obvious.

One question is if you want to be fully open about it - for example, telling the player clearly when his next attack is a guaranteed hit - or hide it.
Or if you count this per player or per unit. Like, the player can miss three times in a row with 70% with different characters, but a single character can not miss three times in a row with 70%. Per player would mean you can never miss with such a chance three times in a row, no matter the characters involved.
Not sure about those myself.

 

[Jim the Dinosaur]

The whole point for me isn't the exact formula that you will use in the end but rather that it should be clearly described to the user.

Describing it would defeat the point of the psychological effect it's meant to achieve. If you have a counter in the UI saying 7/7 hits, 0/3 misses, then the player adapts his strategy (e.g. not attacking any longer with that character), which he's not supposed to do, because the AI doesn't adapt to its shifting odds either.

Players shouldn't be able to have three misses in a roll with 70%. That's not how these things work in the real world.

there you go, that wasn't so hard

 

[Blakemoreland Hybrid Boss]

In the real world the probability of each action can be altered by a series of factors, including stress, focus, etc. It’s not like a series of coin toss.

 

[gaussgunner]

Why are we even talking about 70% THC? That sounds like a lack of confidence. In real life, if you raid a rival gang with odds no better than 70% per attack, you're just asking for them to kill your dudes. 70% is ok if you have some AP to burn and no better moves. But don't complain if you keep missing. 70% sucks. If it's important to get a hit, use buffs or aimed shots or something to push your THC up toward 100%. If you can't do that when it counts, you're not ready for that fight.

This is pure speculation, not about any game in particular, but I wonder if there's an issue of balancing or AI tuning. In an encounter with equally matched enemies, if they fire away whenever they get a >50% shot, they're gonna miss a lot, so then you can take 70% shots and think you're doing good. But sometimes you're gonna miss 5 times in a row. Maybe if the AI went for 90% shots, you'd have to play better and miss less, then you wouldn't be complaining.

 

[gaussgunner]

In the real world the probability of each action can be altered by a series of factors, including stress, focus, etc. It’s not like a series of coin toss.

Those factors would affect your THC. They don't break randomness.

 

[Drowed]

If you have a counter in the UI saying 7/7 hits, 0/3 misses, then the player adapts his strategy (e.g. not attacking any longer with that character), which he's not supposed to do, because the AI doesn't adapt to its shifting odds either.

Well, maybe it should? It sure would be interesting that way.

 

[Blakemoreland Hybrid Boss]

Drowed I agree that changing the THC to diminish frustration is an illusion to make things more fun, but they are more realistic on a psychologic level.

gaussgunner The point here is that each event has a different probability affected by the previous act. They are interconnected. Depending on the person, three misses practically ensure that the next attempt will be a miss.

 

[Jim the Dinosaur]

If you have a counter in the UI saying 7/7 hits, 0/3 misses, then the player adapts his strategy (e.g. not attacking any longer with that character), which he's not supposed to do, because the AI doesn't adapt to its shifting odds either.

Well, maybe it should? It sure would be interesting that way.

Why would that be interesting? You're having a shootout with a mobster and suddenly he runs away; you'd assume that has a reason and not just him sensing his "luck-fatigue" setting in.