darthbator

What is the value in "Randomness"

Recommended Posts

So I just finished watching a roughly 90 minute interview about the randomization tools and methods used in level generation in "Path of the Exile", a game I haven't actually played. Lots of the level randomization stuff they are doing looks incredibly advanced and while I can't say I learned a huge amount I have new absurd respect for procedurally generated ARPG's. Towards the end of the interview he mentioned that they had been working with that element of their game for roughly 6 years at this stage. This caused a strange feeling inside me (/blush). I've actually always been attracted to games with high levels of randomness. Rouge (like) likes, stuff with loot tables and varying map layouts. In a lot of ways I feel like randomness is the "magic" in the novelty engine of some of my favorite games. At some point I start wondering why? What makes that random component so important?

In this particular time we're totally saturated with games to play, media to watch. There is so much to consume and it comes at such a relentless pace you couldn't possibly see it all. This really makes second play through a rare thing IMO. I think the funny thing is that part of that randomness is still absolute magic. The magic RNG black box is really one of the reasons I am currently in LOVE with XCOM. However a lot if times I wonder how much time and resource is dedicated to creating a "unique" randomized experiences for the player just go totally un noticed and un appreciated. What are some examples of games that really leverage a system like that in a noticeable way? Are there any?

I mean on some level it's depressing that lots of really talented software developers are putting lots of work into stuff like procedurally generating maps for ARPGs that people are going to play through once maybe twice, No one is going to talk about the amazing random level layout in that game. In all honestly I don't think most players will even notice how random the levels are at all.

Share this post


Link to post
Share on other sites

The promise of randomness (or procedural generation) is that every game is unique to some degree. You don't have to play the game multiple times to appreciate its random elements, you just have to talk to someone else who played it. What makes XCOM and FTL great is that your experience in these games is personal to you, and if you have friends who play the same game it makes for great discussion. If you have a story-driven game where everything is authored there's very little scope for discussion since everyone who played it went through exactly the same story doing exactly the same things. Unless the story itself raises some interesting talking points (a rarity in games) the conversation is going to be pretty bland.

Share this post


Link to post
Share on other sites

Yeah, if you watch the X-Com Idle Thumbs Twitch stream you actually get a pretty excellent argument (from JP and Evan, IIRC) for why randomness is so important in games. I can't really sum up what they said, but for me it's just this quality for being able to play the same kind of game for so long while still feeling the sense of discovery and systemic cohesion.

Also, I don't know why you would want randomness to have a "valuable" effect in a "noticeable" way. I think the point of randomness in many cases is so there's a veil in front of the game's mechanic elements like the level design.

Share this post


Link to post
Share on other sites
What are some examples of games that really leverage a system like that in a noticeable way? Are there any?

Well yea, Spelunky, FTL, X-Com and Binding of Isaac are all built on that.

Here's one unique one: Time Fcuk is a web game developed by Edmund McMillen right before Super Meat Boy. It has a user level editor where you can make and submit your own levels, and a portal where you can play other people's levels and vote 5-stars on them.

There's a mode called "ENTER THE UNKNOWN" which opens with the game's fiction intro, then plays 20 user-made levels picked on popularity and ordered by difficulty, and then plays the game's ending.

It's a fun idea.

EDIT-- Oh and there's small-scale random choices in action games that keep them fresh and exciting aswel. A game like Street Fighter reads your position and health and everything to manage what it MIGHT do, and then it's a random number that decides whether it's a block or a backstep. Some amount of unpredictable randomness in whether or not a guard opens a locker in Metal Gear, and whether a boss spits lava or shoots arrows in God of War can really make a big impact on how fun it is to play in those systems. I've played like a hundred hours of the last Devil May Cry cos each enemy has so many different things it can do, and it's always semi-unpredictable.

OF COURSE the random elements are heavily weighted by in-game stats aswel, I should be clear on that. It's interesting to think about though, because the enemies in DarkSiders seem to have one close attack and one ranged attack, if that. Fighting them for the 10th time is really just an excercise in pushing their buttons in the right order. Don't quote me on this I'm still thinking about it.

Share this post


Link to post
Share on other sites
for me it's just this quality for being able to play the same kind of game for so long while still feeling the sense of discovery and systemic cohesion.

I think randomness in a game speaks to the strengths of the mechanics. That "systematic cohesion" should be a point of pride for any developer that makes a game like this. It means the rules of the game are strong enough that they support many different expressions of them. If those expressions (one's experiences) are meaningful to the player, all the better.

In a game like FTL or Binding of Isaac the randomness of enemy encounters can only be countered by the consistency of your character progression. The choices you make build on each other. So often a choice is a bad one, turning into a retrospective climax, leading only to falling action and the player's demise. These games aren't testing if you can memorize the exact layout of a level - they are testing you, and your decision making process.

Share this post


Link to post
Share on other sites

I mentioned in the difficulty thread that foresight is a means of circumventing difficulty (though not in an unfair manner). Randomness in video game designs prevents foresight to some extent or another, so it keeps things relatively fresh.

Share this post


Link to post
Share on other sites

Man, I thought this was going to be a talk about "random humor" and expected to see people dump on "randomness". Oh well....

But to speak about the actual topic, if there is one kind of randomness in games I hate, it's random drops. It forces you to grind and "artificially lengthens" the game. That's one thing I like in "western" RPGs, you don't have to 200 goblins to get the snotroot of handsomeness, you just find it growing in a field. Of course, "western" RPGs have random drops sometimes, but since you can't grind in them, it doesn't matter and the randomization in these games is usually how much money they drop and if they have a potato you can eat or not.

Share this post


Link to post
Share on other sites

This discussion reminds me of an old Thumbs cast where they mentioned Firaxis tweaked the probabilities in Civ to make random events "fit" better with how players generally evaluate probabilities, e.g. people get pissed off when something that is 80% probable does not occur, or they have a run of misses, so they cook it so that doesn't happen as much.

Given that the public understanding of statistics and probabilities is already deep in the toilet I see this kind of manipulation of randomness as irresponsible - it reinforces incorrect ideas about what these figures mean and how randomness behaves which I guess could adversely affect the choices people make in daily life. Any thoughts on this?

Share this post


Link to post
Share on other sites

Given that the public understanding of statistics and probabilities is already deep in the toilet I see this kind of manipulation of randomness as irresponsible - it reinforces incorrect ideas about what these figures mean and how randomness behaves which I guess could adversely affect the choices people make in daily life. Any thoughts on this?

I think it's probably not Firaxis' job to singlehandedly overcome widespread fallacies when estimating probability. Humanity is wired to distort probabilities, to overestimate the chances of something catastrophic happening and to underestimate the chances of something happening that we can prevent, as a survival mechanism. It's not purely a question of education.

Share this post


Link to post
Share on other sites

I think it's probably not Firaxis' job to singlehandedly overcome widespread fallacies when estimating probability. Humanity is wired to distort probabilities, to overestimate the chances of something catastrophic happening and to underestimate the chances of something happening that we can prevent, as a survival mechanism. It's not purely a question of education.

I used Firaxis as an example, I'm sure other studios do this too and I didn't mean to single them out as bad guys or anything. I get what you're saying but there are decisions in life that don't boil down to gut instinct and experience can be informative. Also the distortion applied seems to depends on your outlook: people who play the lottery overestimate the chance of something good happening and also believe they can predict it.

Share this post


Link to post
Share on other sites

Given that the public understanding of statistics and probabilities is already deep in the toilet I see this kind of manipulation of randomness as irresponsible - it reinforces incorrect ideas about what these figures mean and how randomness behaves which I guess could adversely affect the choices people make in daily life. Any thoughts on this?

I think it's probably not Firaxis' job to singlehandedly overcome widespread fallacies when estimating probability. Humanity is wired to distort probabilities, to overestimate the chances of something catastrophic happening and to underestimate the chances of something happening that we can prevent, as a survival mechanism. It's not purely a question of education.

I sort of agree with both of you, and don't have a solid opinion on this question, which frustrates me because I think it's a really interesting one. Both in terms of how and why we distort probabilities in our own minds, and in terms of what responsibility, if any, creators of cultural artefacts have to make their works beneficial to society. When creating a piece of entertainment, is entertainment the only goal? Clearly for some it isn't, but do creators have an obligation to educate or persuade or promote consideration by way of their works or is it simply a bonus? It's a question I've been asking myself quite a bit over time in regards to the games industry, but it's also come up recently here in particular as we've been discussing which kinds of books are worth reading for the book club and so on.

As for the other part, why we skew chances and whether it's actually valuable to do so... I think I see some of my day disappearing into research on this.

Share this post


Link to post
Share on other sites

This discussion reminds me of an old Thumbs cast where they mentioned Firaxis tweaked the probabilities in Civ to make random events "fit" better with how players generally evaluate probabilities, e.g. people get pissed off when something that is 80% probable does not occur, or they have a run of misses, so they cook it so that doesn't happen as much.

Given that the public understanding of statistics and probabilities is already deep in the toilet I see this kind of manipulation of randomness as irresponsible - it reinforces incorrect ideas about what these figures mean and how randomness behaves which I guess could adversely affect the choices people make in daily life. Any thoughts on this?

Most people probably interpret statements like "This operation has 99 % success rate" as "There is no risk whatsoever, but don't sue us if your internals are already rigged to explode" and not as "Every hundredth patient dies". Same goes for things like "I'm 95 % sure about this". So yeah.. people's understanding of probabilities is fucked, and I can totally see how some players can't understand how the guy can miss a shot if he has a 90 % chance to hit.

As for the original question, I'm with The Claw on this one. While pretty much every scene in Uncharted was carefully crafted to maximize awesomeness, I rarely felt anything at all because I knew there wasn't anything unique with my experience. But when I was running away the first super mutant I had encountered in Fallout 3 and accidentally led him into a village where it slaughtered most of the inhabitants, I was awestruck. Can this really happen? Or that time when I was closing in on a downed UFO in XCOM without encountering any resistance and then all at once I was attacked by Mutons, Sectoid Commanders and a Cyber disc. You remember that? Of course not, because that experience was mine!

Share this post


Link to post
Share on other sites

That conversation reminds me of Who Wants To Be A Millionaire where they choose the 50-50 lifeline and it removes 2 random false answers.

EVERYONE KNOWS that really they deliberately leave the two answers the contestant's actually stuck on.

Or is it actually random, and that's enough for people to project a made-up motive onto it?

In the same way that the lottery numbers that come up are always right next to the ones you've got, it's fun how random numbers can really become their own character by accident.

Like if you have a really good run in FTL then you get really tense because you KNOW, STATISTICALLY that there must be some horrifying nightmare around the corner.

Share this post


Link to post
Share on other sites

Why use percentages for probabilities if they are not accurate? Just display the probability less precisely, even with just a description.

Share this post


Link to post
Share on other sites

EVERYONE KNOWS that really they deliberately leave the two answers the contestant's actually stuck on.

Or is it actually random, and that's enough for people to project a made-up motive onto it?

Forever ago, when I worked internships in Toronto, I'd take the train to work with my Dad. We share a love of Backgammon, so one weekend I hacked my PSP and wrote a Backgammon game for it. (see: forever ago statement above)

The dice rolls were completely random. Like, literally (rand() % 6) + 1 random. My Dad was *convinced* that I had the dice rolling in my favour. :)

Share this post


Link to post
Share on other sites

This discussion reminds me of an old Thumbs cast where they mentioned Firaxis tweaked the probabilities in Civ to make random events "fit" better with how players generally evaluate probabilities, e.g. people get pissed off when something that is 80% probable does not occur, or they have a run of misses, so they cook it so that doesn't happen as much.

I believe there are "fair" ways to do this. The implementation details matter.

Here's what I think: if there's an 80% chance of something happening, in the purest terms that means that random(0, 100) < 80, right? If that's the case, it's very possible that your RNG will continue to output high numbers, and in reality your chances end up being something like 20%. To me, as the player, this feels unfair/frustrating.

What I'd like to try is an algorithm that normalizes the output to the stated % over time. So lets say the odds are 80%, but after 3-5 rolls the actual % chance is something like 30-40%. Over time, the algorithm will make the chances more likely (raising it up to, say, 95%) so that after 10-15 rolls, your actual chances end up being 80% or close to it.

I'm sure this kind of adaptive algorithm is how the numbers are "cooked", but it feels somewhat fair to me.

Share this post


Link to post
Share on other sites

What I'd like to try is an algorithm that normalizes the output to the stated % over time.

Try this:

random(0, 100) < 80

I guess to achieve the behaviour that people generally expect you could just do something like success = (numRolls % 5) == 0 but this would not be at all random and would be immediately exploited by players.

Share this post


Link to post
Share on other sites

Most players hate proper random in games. There's a bias in their perception of how the games rules behave, and they'll attribute absolute nonsense to game mechanics when things are unpredictable.

We've had an issue where players believe that the AI has a higher chance to critically hit them. And even though this is measurably false, it became the accepted theory. According to them, an unrelated patch broke the game, and the AI was better and unfair. Then one day we had to restart the server to load a new datasource for archival reasons, and the short maintenance window was enough for them to think we applied a patch, and then everything was "fixed".

It's truly bizarre. The unknown is not something that people get comfortable with.

What I'd like to try is an algorithm that normalizes the output to the stated % over time.
This is kind of how randomness is handled in some accounting programs. When you're left with a fraction of a cent, if you just randomly distribute the extra income to one account, then you may be favoring a single party more than others. This becomes a big deal when you have payments split between monthly payments, and broken up amongst individual items. It's far from sexy, but the unallocated income it could generate is enormous. See Superman or Office Space.

Share this post


Link to post
Share on other sites

Try this:

That's true ... eventually. It probably wouldn't be true over the course of a single game.

Share this post


Link to post
Share on other sites

Here's what I think: if there's an 80% chance of something happening, in the purest terms that means that random(0, 100) < 80, right? If that's the case, it's very possible that your RNG will continue to output high numbers, and in reality your chances end up being something like 20%. To me, as the player, this feels unfair/frustrating.

I don't understand this point at all. The chances of you missing/failing three times in a row is 0.8 %, isn't it? How will ones chances "end up being something like 20 %"? And as said, random(0, 100) < 80 ensures the 80 % probability over time.

Share this post


Link to post
Share on other sites

Maybe an example would make my point more clear?

Let's generate 5 random numbers between 0-99: 88, 78, 82, 19, 91.

Only two of those numbers are <80, so only 2/5 "hits" happen. Even though we're testing for <80 (ie 80%), effectively our hit rate is 40%.

Obviously, if we kept generating random numbers, like say, 50 or 500 or 5000 of them, our effective hit rate would be 80%.

My point though, is that many games don't generate *that* many random numbers, and so effectively the game would feel unfair.

What I propose is an algorithm that normalizes the random numbers to the stated % over a shorter period of time, more closely tuned to the rules of the game itself, not to an abstract "as we tend to infinity".

Share this post


Link to post
Share on other sites

Maybe an example would make my point more clear?

Let's generate 5 random numbers between 0-99: 88, 78, 82, 19, 91.

Only two of those numbers are <80, so only 2/5 "hits" happen. Even though we're testing for <80 (ie 80%), effectively our hit rate is 40%.

Obviously, if we kept generating random numbers, like say, 50 or 500 or 5000 of them, our effective hit rate would be 80%.

My point though, is that many games don't generate *that* many random numbers, and so effectively the game would feel unfair.

What I propose is an algorithm that normalizes the random numbers to the stated % over a shorter period of time, more closely tuned to the rules of the game itself, not to an abstract "as we tend to infinity".

So does that mean that the next 10 are going to be guaranteed hits in order to produce 12/15=80%?

Share this post


Link to post
Share on other sites

Yeah. I'd think that a system where the chance of hitting (secretly) depended on the past events would be even more confusing to people than the correct way of calculating probabilities.

Share this post


Link to post
Share on other sites

Not really. "Random" means something different in popular parlance than in technical actuality. If you put your iTunes on "shuffle", you expect to hear a different song from a different artist every time. That's "random". Except it's definitely not random, in actuality. True randomness would have multiple songs from the same album coming up, maybe in sequential order, or even the same song playing four or five times in a row, on rare occasion. This would prove itself to be random if you listened to infinite songs, but you don't listen to infinite songs. You listen to a dozen or so and then turn it off, so that dozen damn well better be completely different, right?

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now