Should I split in "Split or Steal: Can two randomly-chosen shareholders cooperate?" ?
15
188
290
resolved Mar 29
Resolved
YES

The primary purpose of this market is to decide how I should act in @IsaacKing 's market

At some time between noon and 6PM Pacific Time on Mar 29, I will

  • flip a coin to decide the outcome of this market

  • shamelessly trade to profiteer off that knowledge

  • immediately resolve this market

I will then sum the square roots of the number of shares held on each side, excluding my own, and square the sum on each side, getting two numbers Y and N. I will split in the linked market with probability Y/(Y+N) (and steal otherwise). (I will roll my RNG a second time to decide this.)

For example, suppose there are four positions in the market:

  • 1, 4, 4 on YES

  • 9 on NO

Then Y = (sqrt(1) + sqrt(4) + sqrt(4))^2 and N = (sqrt(9))^2, so I will split with probability 25/(25+9).

Get Ṁ200 play money

🏅 Top traders

#NameTotal profit
1Ṁ109
2Ṁ40
3Ṁ27
4Ṁ27
5Ṁ25
Sort by:
bought Ṁ200 of YES

Final results as relevant for how I will act in the linked market:

Y = 2552.42

N = 1163.16

So there is about a 68.7% chance I will split.

Since this turned out to be controversial I'm making sure that I do RNG in a way that is fair and which allows me to do a secret RNG that others can later verify. The following script generates a float in [0, 1] using some entropy of my choosing and a comment that someone else posts.

  • for the first RNG I'll use the next comment on this market. Since I won't trade on this market until just before resolution, this seems fair to me. If the float generated is less than 0.5 I'll resolve YES, otherwise NO.

  • for the second, more important RNG, I'll solicit a comment after this market is resolved. If the float generated is less than Y/(Y+N) as in this market's description I'll split, otherwise I'll steal.

In either case if there are no comments I'll assume nobody cares.

The SHA of my entropy is 7d1a7e659a46543ae0373068dfecbd591d06dda16792c18b5ffd30779f90c15f

The script is:

import hashlib

# hash a secret file I control
with open("entropy", 'rb') as f:
    entropy = f.read()

# and a comment I don't
with open("comment", 'rb') as f:
    comment=f.read()

m_entropy = hashlib.sha256()
m_entropy.update(entropy)
m_entropy.digest()
print("SHA of entropy")
print(m_entropy.hexdigest()+'\n\n')

# use both entropy and comment to compute random
m_secret = hashlib.sha256()
xor = bytes([_a ^ _b for _a, _b in zip(entropy, comment)])
m_secret.update(xor)
m_secret.digest()

# convert to a float in [0, 1]
rand = int(m_secret.hexdigest(), base=16)
max = int('f'*64, base=16)
print(m_secret.hexdigest())
print(rand/max)

This is absolutely unhinged behaviour and you know what i respect it

You didn't make it clear which side is split and which is steal

bought Ṁ50 of YES

@MarcusAbramovitch Split is YES, but I am absolutely confused whether this market resolves to the actual decision

@42irrationalist sorry, that is what i mean. i don't understand how the market resolves

@MarcusAbramovitch I'll do two RNGs:

  • one fair coin flip to resolve this market (so if you have no stake in the other market the fair price for this one is 50%)

  • one weighted coin flip to decide how I act in the other market, with the weighting in the description

predicted YES

@wylderai can't you, like, cooperate, like a normal well-adjusted person instead of creating this godforsaken market

@42irrationalist Issue is a normal person would steal

predicted NO

@xyz in the normal game, yes. in this game, the baseline incentive was supposed to be that they would both have "split" shares that should have helped them. it also depends what people are optimizing for (trading profits, pure mana, using the market to be able to get mana in other ways, etc.)

@MarcusAbramovitch I don’t think shares are a pre commitment to anything. That’s the fun of the market.