Red Pill/Blue Pill — The most important Poll in the History of the internet
a/k/a Reflexivity — Part 2
1.
It was a lazy August morning when the 12-year-old dropped the bomb.
Or, rather, when she got her mom to drop it for her. The “bomb” was in the form of a Twitter poll and if you think “bomb” is hyperbolic that tells me only that you survived the blast radius of the discourse that ensued.
Looking at it with the benefit of hindsight you might be thinking, What’s the big deal? Blue won, no one died. Everyone was saved. Sounds like a nothingburger. But the worst part of a bomb is often the fallout. What made this poll so important was what happened after.
As a shell-shocked survivor it would be too painful to go into detail. But I can bear to give you a gloss:
After Blue’s victory, ensuing discourse was about how choosing Blue had been profoundly irrational. A stain on our collective epistemics. Choosing Blue, you might die, while choosing Red you were guaranteed not to, making Red the “dominant strategy”—a game-theoretic term that was thrown around at the time, meaning a choice that always yields the best results for you, independently of what others do.
The Rationality of Red.
This game-theoretic line of reasoning was found to be persuasive. So much so that, in follow-up polls, the percentage of Red grew and grew. It grew so much, in fact, that it even crossed 50%, leading to Blue deaths.
(Look at the original poll tweet again if that’s confusing.)
Now, I want to restate the above to hopefully make clear why this poll was so important. In the first poll, prior to discourse, everyone was saved. In the polls after discourse, not everyone was.
Again, tighter. First poll. Everyone lives. Game theoretic discourse about what the rational choice is. Blues die. Spreading rational reasoning led to worse results.
That makes no sense.
Not unless you’re reflexivity-pilled.
2.
Here’s how Red-pickers see the game:
There’s a game. I see it, you see it. It’s right there in the original poll question. We have common knowledge of what is the dominant option which is always good to choose, no matter what anyone else does. So that’s obviously the correct answer. Simple as.
How Red-pickers see the game.
Simple, yes. But not reflexivity-pilled.
If you’re reflexivity-pilled you’ll realize that the above is a mere fabrication. You’ll realize that “Games”—in the game theoretical sense of formal models of strategic interaction—are just that: models in agents’ heads. A reification. There are no “games”. There’s only agents’ beliefs about what games they’re in. Which also means, of course, that you have no idea what game other agents believe they’re in.
What’s actually going on from Player’s A perspective.
(You might think I’m an unparalleled genius for coming up with this insight. Thanks. So did I. I’ll enjoy that delusion just a bit longer.)
For now, back to Player B. Maybe they’re slow, or made a mistake, or for any myriad of reasons see the game in a way that led them to the “irrational” choice: Blue. Maybe they care about no one dying. Who knows—and that’s the point.
How Player B might see the game.
They might pick Blue which means that the only way they get to live is… if you pick Blue as well. Or maybe they know Red is the dominant option but they don’t know that you know which means that, by the same logic, they have to pick Blue. But if that’s the case then… You see where this goes.
Now, this chain of reasoning matters if and only if you care that Player B lives too. Maybe you don’t. Maybe you’re out only to save your own ass. Like Koro.
3.
This essay is not about anyone in particular so, for the sake of anonymity, I’ll call the person I’m thinking of… hmm… “Koro”.
Koro was one of the main people pushing the discourse towards the rationality of Red.
Have you noticed how, in movies, heroes are always (1) a team, and (2) irrational as hell? In the poll everyone acted the way heroes do in movies: as an irrational clique. That’s how Koro saw it.
His hubris was in not realizing that his “rational” lens was just that: a lens. A way of seeing. NOT “how things really are” and much less “the obviously correct answer”. He put tinted lenses on so long ago he forgot he ever did and what the world looked like before. “You idiots, can’t you see the world is Red!?” His world certainly is, but he forgets that’s something he had to learn to see.
Koro’s red-tinted lenses are why the results of the original poll—the one where everyone survived—made him deeply uncomfortable. He said as much himself. Deeply uncomfortable at Blue team’s “irrationality”. To fix it, Koro launched a persuasive campaign to convince people to choose Red instead. And, as I said, more people voted Red than not in subsequent polls. His campaign worked. But Blues died.
And, to be clear, his conception of rationality isn’t wrong. Red is the dominant strategy—if payoffs are fixed, beliefs are independent, and no one conditions on anyone else’s reasoning.
But in human systems—in reflexive systems—those conditions fail. And that’s when pushing for Red stopped being “rational”, and started becoming corrosive.
In a reflexive system, “Red is rational” isn’t just a neutral description—it’s an intervention. In a reflexive system, the moment you broadcast a supposedly normative model of rationality, you’re no longer neutrally analyzing; you’re playing.
Koro believed our Blue heroes were living in a deluded story, failing to realize that he too is in a story—a hyper-rational mythos of how agents “should” behave which punishes deviation as “irrational.” The difference being that his led to worse outcomes.
4.
So that is what Koro failed to realize. But what have we realized?
Hopefully we’ve realized that the narrative we believe about what is going on changes how we model things, which changes how we act, which changes what happens.
This is what heroes intuit, even if they can’t say it: in reflexive systems there is no pre-determined correct choice because players don’t sit somewhere outside the game. They’re an intrinsic part of it. And it is through their present choices that they make a past option retroactively right.
Villains’ “rationality” is myopic because it fails to take this into account. It fails to take into account that they’re creating the conditions they mistakenly see themselves as rationally responding to.
Or, said differently, villains are merely rational while superheroes are, appropriately, superrational. That’s why villains “low-key make sense”. Heroes act in a way that is, from the outside, indistinguishable from the way that people who don’t understand that Red is the dominant option act. But they do so to save them.
Failures of theory-of-mind would explain ignoring this heterogeneity and hinging your choices on “if everyone would just”. “If everyone would just pick Red.” Yea, good luck with that. Different people will choose different options for all sorts of crazy reasons you can’t even imagine and you need to account for that in your choices to save the irrational from themselves. Even when they act against their best interests. Especially when they act against their best interests.
The other reason not to do so would be not caring about others’ outcomes. Now that would be villainous.
5.
And, sadly, with that, we reach the end when I can nurse the delusion of my originality no longer. Thank you for this time, it was good while it lasted.
Early on I said that it was thrown around that Red was the “dominant choice”. Hopefully I’ve demonstrated why that wasn’t the case: it would have been the dominant case if a bunch of conditions obtained that didn’t, because we were within a reflexive system.
One conclusion you could come away with is “Game Theory Bad”. I want to dispel that right now.
The problem wasn’t “game theory” as much as it was too little game theory or “pop” game theory. The misapplication of game theory is why Red got the worst outcomes. People did better when they knew none of it than when they knew a bit.
As the saying doesn’t go: a little knowledge is an infohazard.
But that is why we should listen instead to people who have a lot of it. My “original brilliant thought” just recapitulates something figured out by a Russian researcher born more than half a century before me. He even did a series of lectures on it called … drumroll… “Lectures on Reflexive Game Theory”. No comment.
But it’s not just the Russians. When I said above that “superheroes are, appropriately, superrational” I wasn’t just being cringe: I was referencing a game-theoretic term first coined by the American Doug Hofstadter in 1983. Oh, and Germans, too, can’t help but have derived this all the way back in 1785.
My lack of originality pains me, but I was born too late to fix it. So what remains for me to try to make my place in history? Only to spread the ideas of those that came before me. Which is, of course, what this essay is really about.
In true reflexive fashion, this essay is both description and a self-aware intervention—a deliberate attempt meant to undo the damage caused by mistaking a local heuristic for Universal Rationality.
And, if it works, it’ll make itself obsolete. Wait. But then there goes my place in history—
Addendum May 2026
I wrote this essay 5 months ago. The idea was that different people model the “game” differently and that you have to account for that in your calculations. Recently there was another round of discourse about it, kicked off by Tim Urban. What was different about this round is that, this time, LLMs made it trivial for people to show just that, through the memes they made about it:
Excuse me as I take a short victory lap…