▲ 449 points
back
74 comments
This is one of my favorite things on the internet, but it focuses on the positive side of the story which is that groups of people cooperating can defeat a bunch of people who cheat each other. That's a pleasant message.
Unfortunately, I think the corollary is much more important. What this clearly shows is that on an extremely fundamental level, getting cheated or cooperating with people who act in bad faith is what creates the cheating. If you tolerate bad faith, you ask for more bad faith behavior.
If you believe in personal agency and personal responsibility and don't believe in magical thinking, then it shows on a very mathematical level that your own weakness, the ability for someone to take advantage of you without consequences, is what creates defection rather than cooperation.
The lesson is clear, that if you want a world you want to be a part of, then you must become powerful and choose to use that power for good.
>If you tolerate bad faith, you ask for more bad faith behavior.
Yet some bad faith behavior comes from misunderstanding, so to be completely intolerant of perceived bad faith behavior would be the same as entering into the endless "tit-for-tat" cheating loop mentioned on the site
"I'm intolerant of your (perceived) bad faith (because you're intolerant of my (perceived) bad faith)"
Works when perceptions aren't muddied, but that is rarely the case
>The lesson is clear, that if you want a world you want to be a part of, then you must become powerful and choose to use that power for good.
I agree broadly, but that logic is also the cause of a lot of cheating
"I cheat because they cheat"
"I need nuclear weapons because they have nuclear weapons"
Tit-for-tat is a great strategy, but it's not a _dominant_ strategy
I think that's the most important point the website makes
I feel like your comment sums up many life lessons. It's also the reason I stopped believing in extreme leftism as I matured. I think as young idealists we think "Surely if we just gave the government a lot more tax money, they'd solve the problems we all say we care about" and "Surely if you just give freely to whomever says they're in need, they won't cheat the system!"
Later I started to see the patterns where government spends most of the money lining the pockets of the well-connected, and then on the micro level how many people take advantage of any method of unjust enrichment, given the chance, and you start to desire much more accountability from all parties. And yes, things like, say, exhaustive income verification to qualify for benefits definitely hurts those who are playing by the rules the whole time. It's the cost of having trashy individuals in society who exploit everyone relentlessly.
> The lesson is clear, that if you want a world you want to be a part of, then you must become powerful and choose to use that power for good.
This particular lesson we've seen play out time and time again at a state level. Europe, USA, Singapore and even China.
The best thing that happened was corruption was either kept out or effectively eliminated.
When leaders are bad actors, it's pretty obvious, but an explicit example is the Philippines in the 50's-60's and how its economic powerhouse was squandered away.
Today, Vietnam is in the process of cleaning house and showing dividends while Indonesia really could really do amazing stuff if it go its act together.
> This is one of my favorite things on the internet, but it focuses on the positive side of the story which is that groups of people cooperating can defeat a bunch of people who cheat each other. That's a pleasant message.
This is also a reasonable definition of a functional democratic society.
> If you believe in personal agency and personal responsibility and don't believe in magical thinking ...
No man is an island, entire of itself; every man is a piece
of the continent.[0]
Or perhaps a different lesson is that if you want a better world to exist, choose to live it. Sometimes there will be those which do not share this vision and actively work against your choice.
The question then becomes:
Do you allow bad actors to determine who you are
or is who you are orthogonal to what bad actors do?
Becoming powerful might be the easy part
https://www.youtube.com/watch?v=rStL7niR7gs (The Rules for Rulers / CGP Grey)
> getting cheated or cooperating with people who act in bad faith is what creates the cheating
That reminds me of some game-theory stuff, a relatively simple "tit-for-tat with forgiveness" approach does pretty well. It matches a lot of our intuition too: Be nice, punish betrayal, but not too disproportionately.
Not the thing I was thinking of, but found this fun little interactive presentation that goes into some other factors/approaches: https://ncase.me/trust/
You have perfectly described the downfall of western democracies, who are oh so tolerable.
Even Plato described this already in The Republic.
The board game "Diplomacy" teaches lessons like this, also how fairness works even between unequal powers.
"Si vis pacem, para bellum".
OR, you can organize.
The most power-hungry among us often exhibit sociopathic behaviours, and those disproportionately occupy leadership positions. Most good things in our world (democracy, healthcare, welfare) came from building consensus, and weren't gifted by powerful individuals who rose to the top in an effort to bring fire to the masses. To the contrary, when they get their way, welfare and healthcare are cut to give tax breaks to the powerful, and democracies fall to concentrate power more effectively in the hands of the few.
I reject this simplistic Nietzschean analysis of the world. I reject the will to power and the Übermensch. Good societies are built on large groups of people cooperating, not individuals becoming powerful to do good.
I admire your bravery, but, do you understand that you risk losing everything in much of the western world for posting such heretical thoughts?
Vertasium has a great video talking how Tit-for-Tat (Copycat) wins as a strategy (and how there was a math competition that proved it as well)
https://www.youtube.com/watch?v=mScpHTIi-kM
---
This seems like a nice rebuild of the math competition performed years ago (as talked about in the video link above).
Direct link to that part of the video: https://youtu.be/mScpHTIi-kM?si=yzZxyeYw4cJA-i37&t=583
Tit-for-Tat with occasional Forgiveness and occasional Defection (abuse of trust) often performs better than pure Tit-For-Tat, especially in the presence of random error.
Tit-For-Tat falls into "permanent mutual Defection" tar-pit when playing against a Tit-for-Tat-like opponent that Defects once (perhaps in error) and is non-Forgiving.
Humans are pretty good at repeated-game theory, intuitively.
The success of tit-for-tat is a common misunderstanding; it is only the most successful within the makeup of Axelrod's tournament. The OP explains that the strategy is entirely dependent on the environment, and tit-for-tat may not always be the optimal strategy.
No conversation on game theory is complete until someone brings up Golden Balls, and in particular this amazing moment (warning: terrible audio quality). https://www.youtube.com/watch?v=S0qjK3TWZE8
I must say this is amazing. The psychology and manipulation makes me realize how poor I am regarding trust even when the other side is pushing for some unconfirmed equilibrium.
In the game I acknowledge that I was aligned to the "Simpleton" strategy (before it was outlined). Looks like simpleton might actually be applicable in a more general sense too which is a little disheartening.
The Evolution of Cooperation is one of the best non-fiction book I've ever read. Through basic algebra it lets you in on appreciating such a deep and profound idea.
Another great book on the subject is The Joy of Game Theory by Presh Talwalkar.
I remember that in older formal game theory tournaments, a punish-once single-retaliation strategy won out. It was unconditional on copying, simply that if the opponent cheated, you cheat back once and then forgive until another cheat. Another form of Golden Rule approach. But I think those tournaments were under simpler conditions than the one here.
I like the incorporation of miscommunication, and being able to change the parameters.
Once weakness in those tournaments is that every interaction has the same stakes, so it doesn't model the possibility of "con artists" who behave generally at low stakes, but cheat at high stakes.
I remember that as well, but only because it's mentioned in book 8 of The Expanse.
Found this ~10 years ago. Still one of the best things i've ever come across on the internet.
Similar - this is where I found out about Komiku's music; I must've heard to thousands of hours of this while writing code
So this little game actually amplifies the distinction between "game theory" and (let's call it) 'relationship theory'. In the former you rely on strategy. In the latter, you rely on established trust.
You run the game once and at the end you are given 'character' headsup on the participants. Next time around playing the same game, you know who is who.
p.s. In effect the distinction can be generalized as 'depth of priors' for the 'bayesian game'.
Are you talking about "one-off games" vs "repeated games", or something else?
Repeated games are part of "game theory"
I disagree with the phrasing right the first question it asks.
It says "Let's say the other player cheats, and doesn't put in a coin. What should you do?" and the two buttons are "cheat" and "cooperate". But if the other player doesn't put a coin in then not putting in a coin is not "cheating". It is simply not playing the game with that person.
Cheating would be where you say you will put in the coin (or have already put the coin in) but not doing so.
This is really the best ways to explain basic game theory to anyone. It wasn't until some excellent Profs in Grad school that I finally understand some of the most basic concepts, but would always find it hard to explain to friends and family.
I also think game theory is one of the most important philosophies/life-lessons to understand as you go through life and this is an excellent resource to get people started on the basics.
Radiolab had a story about this idea that I enjoyed.
https://radiolab.org/podcast/104010-one-good-deed-deserves-a...
Radiolab's overdramatic style has the ability to ruin even my favorite topics.
This web site was so enlightening. I imagine it could often be difficult to put theory to practice as in real life there are so many variables to consider and many of them we can't quantify. Some of the demos illustrated complete opposite results due to a 1% change in the miscommunication variable. But I will say the 3 main takeaways they give you were more general and certainly applicable to real life.
One of the things I take is that if you keep the game the same but turn the punishment for both cheating really high cooperation blossoms even with high degree of miscommunication. Bring back hanging and quartering for petty crimes!
In all seriousness Game Theory fails in reality because it cannot account for the players changing the game.
I think that detective isn't the correct way to exploit copycat. A character which cooperates, but cheats at random, and keeps defecting if you retaliate, could push the balance towards cooperators. Grudgers, not forgivers, also win in high error scenarios for some reason, as long as enough rounds are played.
If the creator is reading this, I was having problems with the viewport on my tablet. I can't scroll around the page and in portrait mode, stuff disappears off the side of the page, and in landscape mode stuff disappears off the top and bottom of the page.
https://ncase.me/trust/notes/ ( NOTES: "The Evolution of Trust" )
If you want the take-aways, click the next-to-last navigation circle on the bottom of the screen. I won't paste the spoilers here, because I think it would detract the experience.
Fantastic! Top notch explainer and engaging game.
To me it’s important to say that tit-for-tat and the Golden Rule are not the same. My understanding of the two are very different.
This makes me think about The Dark Forest and the chain of distrust that results from lightspeed communications at stellar scale. The universe is harsh...
And the stakes are instant annihilation! so there is no room for mistakes, eliminating almost all other viable positions
One of the most "old but golds". I smile every time this makes it into the top here.
this is pretty cool, very nice explanation of concepts I hadn't touched since undergrad econ
The simulation didn't cover the problem that the US is having right now -- intentional miscommunication. Unfortunately, there's a reason some countries employ warehouses full of trolls and propaganda spreaders. If you are losing the "game" spreading chaos will level the playing field. It will take the people of those countries to stop their leadership from acting in bad faith before things improve.
This is also the reason some social media outlets have become dystopian hellscapes.
Game Theory is a vast concept and I believe this work would be better classified under Prisoner's Dilemma (https://en.wikipedia.org/wiki/Prisoner%27s_dilemma), which is by definition:
The prisoner's dilemma is a game theory thought experiment involving two rational agents, each of whom can either cooperate for mutual benefit or betray their partner ("defect") for individual gain. The dilemma arises from the fact that while defecting is rational for each agent, cooperation yields a higher payoff for each.
..but if you cheat & they cooperate, you gain three coins at their cost of one. (score: +3 vs -1) Therefore: you "should" still CHEAT.
Yes, technically and mathematically, it's 100% correct but morally, ethically and/or emotionally, it hurts.. a lot. Personally, I would never, ever do that!
playing time: 30 min • by nicky case, july 2017
very cute!
Now add communication between players to repeated trials.
spoiler alert: is copycat strategy the reason united states feels like it's becoming more like the rest of the world, more authoritarian?