Thursday, July 5, 2012

Randomness

The resident genius that runs in my offline campaign does not like computer-driven random number generators.  They do not, in his correct estimation, produce 'true' random numbers.  The most flexible, familiar one on your system would be in your excel program, in which you can obtain a random number either by typing either formula:

=randbetween(x,y) - which will give you whole numbers between x & y
=rand()*x - which will give you all numbers between zero and x.

There are ways to play with these formulas, but I'll let you look it up elsewhere.  I mean only to address the problem that these number generators (which use the clock in your computer as the initial 'seed') are only good for ONLY tens of thousands of random numbers.  Hm.  I don't know about you, but tens of thousands sounds like a lot.

If we presume that the time it takes to roll a die, watch it bounce on the table and stop, read the number, and snatch up the die preparatory to roll it again is only one-and-a-half seconds, it would take you four hours and ten minutes to roll the die 10,000 times.  This is assuming the die never takes too long to roll, and never rolls far enough away that you need to lean forward (wasting precious tenths of a second) and that you never drop it.  So, under those circumstances, it is possible to make 10,000 rolls in the space of an ordinary D&D session.

But even if we argue that the second set of 10,000 numbers is precisely the same as the first set (because we have a really crappy number generator), it hardly makes a difference ... because those die rolls will be applied to so many possible tables that the repetition becomes meaningless.  The only concern would be that you or the players knew what the next number in the sequence was going to be - and that they modified their behavior to take advantage.

Try an experiment.  Write out a list of 100 d20 rolls, print them and hand them to each player.  Do not keep a list for yourself.  If at all possible, don't look at the list you give your players.  Tell your players that these are going to be the next rolls that everyone is going to you - DM included.  Then run your game, designating when your players should roll dice to hit, save or whatever else as usual.  Do not let your players choose the order in which they attack - go right to left around the table, or whatever usual method you use to keep order in your game.

I'm guessing that foreknowledge of the rolls won't be that much of an advantage.  I'm also guessing the players will get frustrated at the enemy "stealing" their rolls.  They might be able to strategize a little, since they will hold a physical list in their hands which they can think about in advance.

Imagine how hard it would be to make it work if they had only their memories to rely upon, and could not compare memories with each other.

The reality is that we're not complex enough to take advantage of a repetition even if there was one.  If there was some Rain Man at the table with perfect memory, and he was restrained from shouting out the numbers ahead of time (I know a number of people who would like Hoffman gagged), he wouldn't have that much of an advantage, since those really good numbers were going to be stolen from him most of the time.  Perhaps some genius would like to write and tell me how they'd manage it - for myself, I think that one more elf standing behind a tree taking one unexpected bow shot would totally fuck up Rain Man's whole plan.

Tell me how I'm wrong.

From my experience, I believe the game itself is unpredictable enough to counter-balance any lack of randomness that might be present in a pseudo number generator.  There are too many human variables influencing the exact course of action from minute to minute that doesn't exist in the static computer codebreaker/statistics manager that needs 'real' random numbers.

Perhaps there's a future in cyphers based on four D&D players fighting a red dragon that the Russians are working on right now.

ADDENDUM:

You can try this experiment too.  Tell your players you're going to use a random number generator for every roll all night.  Make lists of 150 x d20, d12, d10, d8, d6 and d4 rolls, and use that list over and over again - using multiples for when more than one die is rolled (3d4 for instance).  Don't tell them you're doing this.

See if anyone notices a repetition.  If they do, see if they can make use of it in the course of one session.

(You can always justify these things by handwaving their justification with magic)

7 comments:

Adam Thornton said...

I tried this at my table, when my players threw the universal "Free Will / Determinism" knife switch to the other position. (An actual thing in that game; I'm big on reification of metaphor, although I stole that particular thing from Graham Nelson's Curses).

As you surmise, it'd didn't make any perceptible difference (we just had a sheet of paper with columns full of dX rolls, which got crossed off as they were used).

Eventually they got creeped out and put it back.

Richard Todd said...

Excel's prng was pretty bad, but I believe it has since been improved. Even a low-quality pseudorandom cycle length will be expected to in the billions these days. Realistically, if you tested sequences mathematically, I'd bet these numbers are actually more random than the output of cheap plastic dice + crafty players. Still, there is a certain "fun factor" involved in throwing your own dice rather than having a laptop spit out the numbers. I kinda prefer it, myself, even though I know the computer can do a better job.

Oddbit said...

For non-players with large amounts of information to process (14 archers or kobold warriors or the like all attacking) Excel for calculating the data is great in all ways unless the GM rolls in front of the players. The dice offer a lot more drama and anticipation.

Now this post combined with a comment in the previous post make me wonder about a big state machine for determining the results of a round of combat. The GM enters the enemies actions, then the players enter theirs, then the combat plays out on a screen like some kind of sports play... That would be an interesting way to use computers for gaming.

Carl said...

I have played and DMed sessions with no dice, using a dice roller app on a smart phone to generate the random numbers. I saw two different issues.

The actual randomness of the numbers is one thing, and my guess is that the random number generators are more "random" than dice throws. At no point did anyone question the validity of a random number from the app or claim to notice a pattern. This actually seems to happen more when people are throwing dice.

The experience at the table of throwing the dice vs. using a random number generator is another thing, and renders the first issue more or less moot in my estimation.

You can't shake a number generator in your hand and blow on it. You can't throw it soft so it barely rolls or hard so it bounds around for an eternity. There is no moment of suspense as the rotation slows and the final tumbles produce a result.

Carl said...

Now that I think about it there are probably ways to make using a random number generator a more tactile and rewarding experience. The idea that popped into my head was printing out that list of 20,000 random numbers (or whatever) in a boxed grid small enough that you couldn't read it from dart throwing distance. Then put it on a dart board and have fun, re-throwing if you hit a number that has already been played.

That actually sounds like a ton of fun. Of course you will get players scoping out the position of those natural 20's, but trying to hit that small of a box with a dart is going to be considerably worse odds than 1 in 20!

jbeltman said...

Carl, you could do that but just bring it out each time the group gets into a fight in a bar, since darts = bar!

Tom said...

I suppose my problem with random number generators is
a) I'm not savvy enough to know if one's been tampered with.
b)I know that several of my players OTOH ARE that kinda savvy.

With physical dice the pool of people able to 'load the dice' is quite a bit smaller (down to one), and my confidence in the ability to detect such tampering is quite a lot greater.

Other than that I have two players who use the things on their tablets at the table, while myself and another player prefer the traditional dice. Remarkably, no-one has yet cried out 'Heretic' or 'Luddite' and attempt foul murder/ refused to play.

But then I don't even have a cell phone, let alone a smart phone or tablet.