Randomness or the Perception of Randomness?
May 27, 2024 5:40 AM   Subscribe

 
no dice
posted by HearHere at 6:08 AM on May 27 [1 favorite]


I'm skimming... am I correct that this is this about the sometimes counterintuitive differences between random outcomes and random paths? E.g. the kind of thing you see with genetic drift in small populations, or random walk theory in economics, where a path created by a series of random decision points can look like it's a pattern that has a sensible narrative explanation?
posted by clawsoon at 7:02 AM on May 27


What’s more, it’s reasonable to expect that this margin of error would probably halve each time you doubled the number of rolls.

Imagine someone showed up and started talking about Shakespeare. And calmly asserted that Shakespeare wrote most of his plays while living in France, as if it was a fact everyone knew.

(It takes 4x as many samples to tighten the distributions by a factor of 2 - standard deviation, which is basically what that is measuring, scales with the square root of samples).
posted by NotAYakk at 7:18 AM on May 27 [5 favorites]


Imagine someone showed up and started talking about Shakespeare.

Shakespeare? I'm pretty sure we're talking about chess.
posted by clawsoon at 7:24 AM on May 27


clawsoon, I think it's about that, more or less, as well as how unprepared we are to deal with true randomness. A bit like the conversation a week-ish back about lying in games, and when, where, and how inaccuracy affects narratives.

NotAYakk, I don't understand. Are you saying that the calculations are incorrect, or are you saying that he's drawing incorrect conclusions from his calculations, or is the Shakespeare comment a comment about the conclusions that a person might assert based on looking at randomly clumped data? Something else? (Sorry if I'm missing nuance here -- numerically I do little beyond basic math, algebra, or stats these days. With a weird little offshoot into math necessary for network analysis, where a program does most of the calculations.)
posted by cupcakeninja at 7:25 AM on May 27 [1 favorite]


Shakespeare? I'm pretty sure we're talking about chess.
"what's the first thing you remember?" is dialogue i'd forgotten from the beginning of Rosencrantz & Guildenstern Are Dead
posted by HearHere at 7:29 AM on May 27 [4 favorites]


cupcakeninja, just imagine someone was writing a post about Shakespeare, and in it they confidently talked about how everyone knows that Shakespeare wrote their plays in Brittany, France, and goes on a bit about the beautiful countryside that inspired him.

Then goes on and talks about the relationship of Romeo and Juliette to post-modern pop art.

The statement about where Shakespeare wrote his plays not really relevant to the pop-art thesis, but it is also ridiculously wrong. And it is really jarring.

That is what "if you double the samples, you halve the error" claim feels like in this post. Really jarring and makes me not trust anything the person says about statistics or randomness without checking. It might contain valid points, but if this person told me the sky was blue I'd go outside and look.

I mean, they claim those images are produced by a certain kind of random walk, but after *that* kind of error I have no reason to believe anything they say on the subject. I'd find it credible that they drew them by hand. Why? I don't know, but I can't prove they didn't.

Verifying their other claims, or finding a flaw in them, takes work. And I don't want to do the work when they start off with displaying really fundamental errors in how random sampling works; if that kind of fundamental error is an acceptable error, any other omissions or errors aren't going to be surprising information.
posted by NotAYakk at 7:44 AM on May 27 [7 favorites]


Ahhhhhhh, thank you. That's very helpful, and it makes me question what's here, too. Hmm. As a writer and a player of games, I can draw various conclusions from both right and "usefully wrong" articles, but it does make me question the utility of the article for srs bzns game design.
posted by cupcakeninja at 7:48 AM on May 27


NotAYakk, I've sent the standard deviation part of your comment to the website's author, with the hope that they will find it interesting and/or useful. Hopefully their contact form works and they actually read it.
posted by clawsoon at 7:59 AM on May 27 [1 favorite]


cupcakeninja, another good thread on randomness is here. thanks, as ever
posted by HearHere at 9:32 AM on May 27 [1 favorite]


For example, let’s take a string of five results: 4 – 7 – 8 – 10 – 16. Even probability means that this is just as likely to occur as another valid set of results: 10 – 10 – 10 – 10 – 10. Yet, if you were to show those strings to someone, they would have little hesitation in describing the first as random and the second as decidedly not random.

This is very much true. I tried to convince someone that they might as well buy a lotto ticket with numbers 1-2-3-4-5-6 because it's just as likely as any other ticket. I failed.

This article is a bit dense, but I do want to eventually read the author's other articles on this topic.

Whether you go for perceived "randomness" or true randomness, you're only playing tabletop RPGs wrong if you aren't having fun. This reminds me of another split: whether you want to play the adventure straight and finish the quest or you want to seduce the kobold mayor and burn down the tavern, either way is fine as long as every player is in agreement on the style they're playing.
posted by AlSweigart at 11:02 AM on May 27 [3 favorites]


The connection between RPGs and statistics goes back at least as far as Gary Gygax's authorship and professional background. My first introduction to distributions and dice rolling came from the 1st edition AD&D manuals.
I've always had a fascination with procedural generation of games and stories, and this article plays into that. Sometimes, as a DM, I enjoy rolling up random events or locations (and ChatGPT has been a very interesting addition to that approach).
But in the end, I've learned that most players are far more interested in the story of their characters and the situations they encounter - and not whether the world is 'playing fair.' They would rather I look them in the eye and baldface lie about what the written adventure or any dice table says, in order to make a good story. I've tried hard to apply that to my DMing techniques. Certainly, when it comes to combat rolls, I consider it not just my prerogative, but my duty, to fudge the results to suit expediency and story.
posted by Flight Hardware, do not touch at 11:10 AM on May 27 [2 favorites]


I tried to convince someone that they might as well buy a lotto ticket with numbers 1-2-3-4-5-6 because it's just as likely as any other ticket. I failed.

Total derail but: if all numbers are equally likely, then the most important thing to do when choosing a lotto number is choosing one least likely to be chosen by someone else. "The kind of code an idiot has on his luggage" is at best a double bluff, hoping nobody else is as dumb as your pick.

Anyways, as a child, the true question of GM randomness wasn't whether true randomness or perceived randomness was better, but whether a GM who always rolled dice behind a screen could be trusted to not lie and kill off their annoying brother's character.
posted by pwnguin at 11:19 AM on May 27 [6 favorites]


"The kind of code an idiot has on his luggage" is at best a double bluff, hoping nobody else is as dumb as your pick.

If there's one thing I've learned from watching Taskmaster, "I'll pick this because everyone else will assume that everyone else is picking this, therefore nobody else will pick this" is always a failing strategy.
posted by clawsoon at 11:25 AM on May 27 [8 favorites]


"I'll pick this because everyone else will assume that everyone else is picking this, therefore nobody else will pick this" is always a failing strategy.

If you and your friends like thinking about this sort of iocaine powder type of question, the collaborative game Just One is like a cross between that and taboo.
posted by aubilenon at 12:48 PM on May 27 [3 favorites]


I'm not going to jump into a discussion of probability math. I'm a retired game designer, after all.

But I appreciate the call out in the article to:

https://anydice.com/

Which is a great tool for crafting dice-based experiences.
posted by chromecow at 12:55 PM on May 27 [5 favorites]


This piece claims that the numbers 1 2 3 4 5 6 are played by about 7,000 people for every draw of the UK lottery.
posted by Rumple at 4:01 PM on May 27 [4 favorites]


NotAYakk, I've sent the standard deviation part of your comment to the website's author

They were happy to get the info and will update the post when they get a chance.
posted by clawsoon at 5:30 PM on May 27 [1 favorite]


Not D&D, but somewhat relatedly, one of my favourite implementations of "randomness" is the weird Warhammer 40K CCG where every card was simultaneously a fighting unit, a battle action, and a dice roll, depending on the context it was drawn or used.

"Heavier" units tended to be balanced with lower dice rolls while "lighter" units tended to have higher dice rolls, I suppose to represent that lighter units were more agile, and could better exploit any opportunities or weaknesses the enemy represented. The card could either be used as a unit deployed to contest a location (kind of like Marvel Snap), or it could be a "battle action" which represented off-field support.

So when some ability goes, roll 4+ on a D6 to get this effect, you'd draw the top card of your deck and look at the dice roll then discard it.

The dice rolls were pseudo-random, because you had a choice and some control in crafting the probability distribution of your deck.

The coolest part of it was the Eldar, which are a psychic race with limited ability to foretell the future. In exchange for their units being generally weaker and less numerous, several of their units had the ability to "see" the future, either by sneaking a peak at the top card of their deck, and some could even change the future, by allowing you to put that card on the bottom of your deck if you didn't like it. So they won by winning dice rolls more often, or losing the ones that mattered less, which was incredibly flavourful to the lore.
posted by xdvesper at 6:53 PM on May 27 [2 favorites]


« Older The Vast Jalapeño Conspiracy   |   I’m furious that they are responding at all. Newer »


This thread has been archived and is closed to new comments