Red or Blue?

They don’t need to be conspiring to kill you. It’s entirely possible that 26 randomly chosen humans, whatever they agreed previously, might decide they can’t take the risk of death of pressing blue. You are betting your life that a majority of a group of strangers will be willing to press the “maybe die” button, because that’s what they told some people they have never met before they were going to do.

It’s insane to put all of you through that rather than just choose red. Even if everyone does press blue as agreed that’s a hellish situation to go through for no reason.

My various examples of added incentives, from money gain for blue to finger loss for red, are trying to convey my intuition that it should take very little for red-pressers to see how extremely safe they can become due to the incentive’s nudge toward blue as a default, even if they didn’t see it as a default before. (I should be clear that all of those, if inverted so that blue pays penalties and red gets rewards, make me switch from blue to red out of the equivalent thought process; I hate that it works out that way, but it naturally has to.)

I haven’t had much luck getting that across, regardless of the website. Generally, though, any “physical punishment” version – every red-presser has to lose a finger, every red-presser has to drink some quantity of urine – seems to make it a little clearer than the monetary ones do.

How about this: it’s the classic original, but the one extra step required is that if you want to press red, you have to win a coinflip; losing that counts as pressing blue. Do you press red?

So that’s an interesting one. It takes maths :wink:

My actual first thought is actually it makes red more attractive. Most likely no one is dying no matter what anyone chooses. For anyone to die at all it would mean less people choose blue than the difference between the red and blue coin flips (and that there are more red than blue flips). If Google ai is right that’s around 45k (the standard deviation? My stats are not all that) . So if more than 45k people choose blue we are probably safe.

But it’s still twice the risk choosing red over blue. Whatever the small risk of death is it’s twice as high for red as blue. So maybe red? But on the other hand the chance of you personally being the difference between people dying and nobody dying is FAR greater here, like by like 6 orders of magnitude. You still probably won’t but it’s a possibility. And if anyone dies then it’s around half the world that dies, that’s a world ending number of people. So yeah maybe blue?

Did this swap the colors?

I find it extraordinarily unlikely that 26 people, having agreed to press blue and knowing that everyone else has agreed to do so, would go back on their word for no good reason, also knowing that doing so is likely to kill a bunch of people.

You find that “entirely possible”. And you call it insane to disagree.

Okey dokey.

Again:

You have little faith in people’s integrity. I have little faith in their competence.

Yes I meant to say

Whatever the small risk of death is it’s half as high for red as blue

So you are going to go with the “hey everyone! I know we we don’t need to, but lets all press the death button!” argument. Okay dokey, thats totally logical there is nothing objectively wrong with this argument :roll_eyes:

Has anybody on the blue team acknowledged the functionally identical poison pill (antidote supplied only if 50.00001% or more take it) scenario and said they would take the poison rather than just not?

Have you considered the possibility that not everyone thinks as you do?

Yup that’s why it’s objectively the best outcome. in this case you all need to agree (the only way there will be a large risk of someone dying in this scenario is if we fail to agree).

You can argue “hey we should all press the death button that can kill us!” Or we can argue “no one should press the death button! Let’s all press the button that is guaranteed not to kill you!”. Which argument is more likely to convince everyone?

The poison pill one is not functionally identical at all, because the question is functionally changed every single time the setup is changed. That’s what I meant by my example earlier. “Is Poe’s The Raven seven words long?” is interchangeable with “Is Poe’s The Raven exactly seven words long?”. But we can’t derive that “adding the word exactly changes nothing”, because “Is this question seven words long?” has a very different answer from “Is this question exactly seven words long?”

If you set up the problem in a certain way, then you can expect everyone else to have the same setup, and expect them to expect others to expect others, etc. Every subtle change makes the coordination dynamic completely different.

A really important element is what this or that version presents as a “do nothing” default. One logical counterpart of your pill one is: everyone is offered an injection of antidote, and told that if and only if a majority take the injection, the whole set of people will be injected with the poison, against their will or no. If you say “Yes, of course I choose the injection”… I don’t know what to say. You could have just done nothing. That would have been the same as ordinary daily life.

For the pill one, you and 100 really smart people had to coordinate, then “Don’t take the pill” works great as coordination. But wait, you say! That’s an admission of red over blue! Yes, because “What if one person messes up” simply isn’t going to happen. Our physical instincts are part of the equation here. The opposite scenario, where everyone takes the pill followed by the antidote, requires a collective defiance of physical instincts, not merely button-pressing, which is 100% “neutral”. That’s what I’m getting at when I say: what if choosing red also entails taking a sip of urine, or something. Do you really intend to get dozens or hundreds of people to do that, unanimously, “just to be on the safe side because death was on the line.” That’s a situation where everyone Could Just Not.

Now, of course, you say “But if the pill version is applied to the human population, lots of toddlers and confused people might take it!” Yes, and that’s horrible. But even then, coordinating a majority to take the pill-and-then-antidote as well will be much, much harder then when it was merely pressing buttons. You all have to automatically trust the physics of the situation: that the antidote will work perfectly, you won’t be sick afterward, etc. I don’t know what, if anything, is a parallel to the trust problem in a pure-button version. That is to say, supposing everyone unanimously presses blue, what bad outcome do we fear remaining possible according to the human imagination? We’re not told how the death is applied at all. Maybe red-pressers just assume the blue buttons are electrified or something?

I posted why we might need to. When you missed it, I quoted it for emphasis. When you missed it again, I quoted it again, and explained why I thought you might be missing it repeatedly. I’m not sure what’s going on here, that you paraphrase me as saying “we don’t need to” after those three repetitions. You might disagree that we need to, but to say that’s what I’m going with? Very very strange.

The likelihood of both arguments convincing everyone is a minuscule rounding error from 0%.

Yup your going with “we should press the death button!” That Is objectively worse argument than “none of us should press the death button!”

“Objectively”?

Yes saying we should press the “death” button is an objectively worse argument than saying we should press the “not death” button. That’s a pretty uncontroversial statement.

The likelihood of convincing 50 people to press the death button is close to 0. There absolutely nothing a stranger can tell me that will convince me to press the “might die” button when pressing the “definitely won’t die” button has the same outcome.

Convincing 50 people to press the “won’t die” button is straight forward. Are you going to stick to your guns and press blue if 49 other people say they will press red? 40? 26? 24? 20? There is no way in hell once a handful of people say they will press red it won’t convince the others to do so.

Say “Well done, you logically took the choice that protected you. You’re a smart cookie!”

Oh you got that across just fine. Better I think than my attempt of what if you knew that someone you know well, that would have guessed would press red, pressed blue?

The majority of us have a default that while it might be close, a majority would also press blue and all live. To press red in that mindset is cheating on an understood social contract. A minority have the opposite default.

I agree with you that the question of how much of a, and what sort of, nudge it would take for each to flip sides is interesting. I suspect there is nudge level that would have the polls flip from most often solidly blue to almost all red, and the reverse, extremely dramatically.

It’s not uncontroversial, not least because of your framing. But having tried unsuccessfully four times to convey an idea to you, I’m convinced I’m unable to convey the idea successfully, so I’m out.