What's the point of randomness?

[Alex]

Randomness has been in RPGs since day one. In the old D&D games, we have randomness because, at the heart, these games were about coming up with plans. For instance? how do you deal with the piranha infested river? Do you jump across? Try to find a bridge? Do you poison the water? Do you use a scape goat, throwing something for them to eat? Do you use a cold spell to freeze the water and walk (skate?) across? etc. Now the thing is, these plans frequently had more than one step. If I knew beforehand if I would succeed or not at each step, I would never need to worry about complications, about a plan failing. Either it would always work, or it would always fail. Of course, the GM might keep that information to himself, but that would make the game worse off. Knowing your chances beforehand, and understanding them help make the gameworld seem more "real", and it helps the players get better at planning and preparing.

Some games have shifted this idea somewhat. Amber doesn't use any kind of randomness, while some indie "story games" use randomness as a kind of an extra "storyteller", that is, something to keep the narrative from following exactly where the players would expect.

You could, if you wanted, make a game without randomness work. Though I have no idea why you would want to do that. The only good reason I can think of is that randomness doesn't work so well with saving and loading, and in certain circumstances people may save scum in order to obtain a nearly impossible result. However, I think adding an ironman option to the game much more sensible.

 

[DramaticPopcorn]

There's no randomness involved in combat. Either you are a better fighter than your opponent or vice versa. If you both have the same ability then I would expect there to be a standoff and the conflict would need to be resolved some other way (or maybe special abilities would come into play (power attack once per day, uncanny dodge, etc.)). The only randomness might be a sudden gust of wind or something, but it still depends on ability because a good fighter can compensate for things like that.

It's not that there is randomness in combat...it's that there are variables out of your control. You can rationalize all you want, but you have no idea when your opponent will attack, which direction it will come from and what other factors might come into play during his attack (a distracting noise, you slip on a rock, a movement in the corner of your eye, his friend comes to help him kick your ass, you lose your footing, he sidesteps an attack, you overextend yourself, etc). None of these things are random (in the sense that something just happens for no reason), but they are definitely variables that mix things up.

So, stop looking at dice as "random," and look at them as putting the variables into play because combat is chaotic and life is chaotic.

I mean, I have a pretty good chance of living through the day if I hang around in my living room watching TV (which means I have pretty good living ability, I guess), but all bets are off if an airliner crashes into my house or terrorists set off a bomb down the street. There are always going to be variables you cannot control. And one of the most important things to realize is no matter how good your supposed "fighting ability" or strength might be, somebody else might just introduce a variable you didn't expect and beat the living hell out of you (ask Goliath how his fight with David turned out).

This is pretty much roleplaying 101. If you cannot bastract yourself like that, you're probably playing the wrong genre.

 

[Servo]

What is a bastract

 

[DramaticPopcorn]

abstract

 

[Galdred]

As other have pointed out, randomness simulates the chaotic nature of combat. The "best knight in the world", William Marshall, and the "best Samurai Dueler", Musashi both main skill were opponent selection : ie William Marshall would wait in the sideline during tournaments and look for winded opponents to get easy wins, while Musashi would carefully cherry pick weaker opponents to force them into a duel, come late to offset their balance.
Both avoided fair fights, probably because the most skilled fighter does not always win, so you'd better take as little chance as possible.
There is a reason all wargames(which try to be war simulations) use randomizer to resolve conflicts instead of a single comparison of strength after all. War is by nature very chaotic (what Clasewitz calls friction).

Strategy, if we follow its military roots(ie, the skill to perform better war "moves" on a large scale. Organization, leadership and logistics are considered a given, and are out of scope, as they are distinct skills of their own) is not only the skill of optimizing piece placement(which we can pedanticly call combinatorics), it also requires managing risk, and more importantly, managing uncertainty (I don't remember whose definition it was, though).
Chess only has one of this component is not a "full scale strategic game". It could almost be considered a puzzle game, as you can consider all outcomes with infinite computing capabilities (it certainly is, if you run a bot against another), and do not care about what your opponent is thinking.

Now, what does randomness brings to the table, gameplay wise?
As Richard Garfield (Mr Magic: The Gathering) pointed out in a serie of articles and his book (Characteristics of games), randomness is orthogonal to skill (Tic Tac Toe has no randomness, but requires little skill, poker has very high randomness, and is very skill dependent, while chess is very skill dependent, and has no randomness).
Poker might be considered to be more skill dependent than chess because no limit holdem still cannot be beaten by a computer, while chess is more or less "solved".

Randomness brings several things :
- It broadens the level range of interesting games (you can play Blood Bowl against a weaker opponent, and still lose, so you need to find a strategy that reduces variance against weaker opponents, and one that increases variance against stronger ones, while playing against a much weaker or much stronger ELO ranked foe in a deterministic game like Starcraft or Chess will result a very boring one sided game.
- It makes the number of possible game states much higher(as they are many more different outcome for a single serie of moves), making a brute force puzzle solving approach, or a memorization heavy approach, like Chess openings, or RTS build orders, impractical.
- It adds a layer of risk management (you have to consider many more possible outcomes for a single move).

So in the end, no approach is better than the other, but having randomness or not(and the amount of it) will have far reaching consequences about the nature of a game. I think randomness works better in RPG, as they are by nature simulationnist, and the situations they are trying to represent are pretty chaotic.

 

[odrzut]

The world runs on quantum mechanics. Therefore there is randomness in everything. Even if you could perfectly simulate reality you would have randomness in it.

 

[Twiglard]

Kurwa, a probability distribution doesn't preclude analysis. See also Monte Carlo.

 

[shihonage]

The point of randomness is a jetpack powered by regret.

 

[Galdred]

I forgot about other 'chaos inducing' methods :
Some games use 'mindplay' as a chaos generator :
Rock Paper Scissor itself is not random at all, yet the results are somewhat unpredictable.
Other games use RPS as a chaos generator for tests :
For instance, the Inxile version of the Sorcery gamebooks series use a variant of the firewall game for combat resolution (the original gamebook used dice). I don't remember the system exactly, but it is something like you wage a few points out of your total (determined by your skill), and the one with the higher score deals damage (the damage depends on the amount of points waged by the loser : the more he waged, the more he exposed himself, and takes damage). You only recover a part of the missing point pool every round.
It has no randomness, but it allows for the result not to be deterministic.
Simultaneous turn resolution can act like a quasi randomizer too (like in Frozen Synapse for instance), for the same reason, especially if it is coupled with Fog of War.
A RPG could very well work on such a basis (Inxile's Sorcery comes has some RPG-y features after all).
For instance, you could have several combat moves (feint, block, attack) that perform differently against each other, and have the results determined by the relative skills :
ie feint would beat block, and inflict fatigue damage (or decreasing skill) depending on the relative skill of both fighters, block could make the opponent more vulnerable on his next turn, decreasing his skill for this turn, while attack would deal extra damage to feint.
That's a very simplistic exemple, but it could certainly be fleshed out into a Pen and paper RPG. However, it would make for a poor CRPG, as mindgames are not very fun against computers (as you can hardcode the game theory optimal strategy, making the minigame play out like a random roll, only much slower).

To a lesser extent, this adds some uncertainty in RTS too (expand vs rush vs tech, and fog of war, allow the games not to all play the same).

 

[IDtenT]

Deterministic simulations are boring.

 

[Twiglard]

eh, Random variable ain't what some people think it is.

Time-variant functions, idem. bbl, basis vectors need adjusting.

 

[Gregz]

On Stats, HP, Levels, Experience, and RNG:

 

[Twiglard]

Entropy ain't what it used to.

 

[epeli]

I've wondered why some cRPG rulesets (that have nothing to do with P&P and actual die) and even computer games of other genres limit random chances as if there was a d20 in the play - sometimes min 5%, and even more commonly max 95%.

I mean, there's nothing wrong with limiting the extremes of random rolls or using simple abstraction, but why d20? It feels like cargo cult design.

 

[IDtenT]

Well, that's no different to every fantasy setting still having two sexes.

 

[Burning Bridges]

Many things we do require both skill and luck.
Things like shooting. Or playing tennis. Even the best players in the world need a bit of luck to hit perfect shots.
There is both skill and randomness involved.

 

[Average Manatee]

I've wondered why some cRPG rulesets (that have nothing to do with P&P and actual die) and even computer games of other genres limit random chances as if there was a d20 in the play - sometimes min 5%, and even more commonly max 95%.

I mean, there's nothing wrong with limiting the extremes of random rolls or using simple abstraction, but why d20? It feels like cargo cult design.

It's just a good point at which you can say that something will almost certainly happen but still have a non-negligible chance of backfiring. 90% is not reliable and 99% is too close to 100%.

Statisticians like to talk in terms of 95% confidence intervals, I have to wonder if there is some link.

 

[Baron Dupek]

Some spice against predictable actions. Limited and used wise make game interesing, but like most features is not used wise....

 

[Bibbimbop]

Arguably, dice have no place in a single-player setting where people will just reload. Rather than an actually important feature of computer RPGs, it's more likely just a hold-over from the initial table-top days of gaming as a group activity.

The first wave of table-top war games probably grew out of the dice games which have, completely unembellished by storylines and rulesets, been the preoccupation of groups of idle men for millennia. Dice actually are fun in a group setting like this, because there is no temptation to "save-scum" when other people are keeping the outcomes honest.

In single-player, it usually provokes tedious reloading and encourages OCD bad habits

 

[DraQ]

Simulationist angle:
A perfect simulation would require a dedicated universe to run, therefore any practically achievable simulation is imperfect and of limited scope. Factors outside of simulation's scope can produce seemingly random variations in behaviour of simulated system. This is independent of whether our own or game's universe is deterministic or not.

Gamist angle:
Introduction of random factor force player to plan robustly by breaking reliability of any single given plan. This broadens space of possible courses of events and makes the game more interesting.

Narrativist angle:
...and both help building up tension and upholding good narrative.

</thread>

 

[Twiglard]

:mondblut:

Edit: wow it worked!?

 

[Skall]

I prefer systems that rely on minimal randomization when it comes to combat (roguelikes and other highly-random systems are another story). As mentioned above, having no randomization within a given conflict in a single-player game doesn't necessarily mean all of the variables are known and it's just a puzzle. And even if it is, few people can memorize and optimally play through all the permutations of chess or other complex systems.

I understand that certain games are designed with risk management in mind, e.g., Jagged Alliance, and it's up to the player to plan for various contingencies. I also acknowledge that things not always going the player's way can create memorable, internal stories: "My three veterans had a weakened soldier surrounded, dead to rights, but in a mad dash he clipped one of 'em and only two made it back."

Still, it's not something I typically enjoy.

I'm sure we've all had stretches where we missed three 80%+ checks in a row only to be nailed by a 23% attack in return, which also passed its 5% crit check. At times like these it's easy to fall into the gambler's fallacy, but even with proper understanding of probability, my gut reaction is always frustration. A big reason why actually stems from all those tools and tactics meant to alleviate randomness. If it's always adventurous to lay down a smoke grenade, break line of sight, seek cover, attempt a flanking manoeuver, conserve AP for reaction shots, etc., why ever go for a low-chance attack? However, playing smartly like this means that I never feel like I got lucky because I never took those low-probability chances, while the dumb AI occasionally got rewarded for bonehead moves.

 

[28.8bps Modem]

You should go necro the tangut prophecy thread. That's a good time for everyone.

 

[Skall]

Looks like it judging by the end of that thread.

 

[DraQ]

However, playing smartly like this means that I never feel like I got lucky because I never took those low-probability chances, while the dumb AI occasionally got rewarded for bonehead moves.

That's called "having superior numbers". They need to only get lucky once, you'd need to get lucky all the time.