TL;DR bet on which word is the most likely to be the next word in a valid haiku.
The Haiku so far:
Brings bayesian slowly
Singularity
Background here:
The next word will be chosen from the sorted list of words, with the highest weighted one first. For each word, we will either choose it as the next word with 50% probability, or skip it and evaluate the next word. All markets will resolve when either 1) a valid haiku has been created or 2) the process has failed to produce a valid haiku, in which case the “failure” options in each market will pay out.
I will not trade in these markets, with the single exception that I will buy 'Haiku fails' shares to disincentivize gaming towards that outcome. Since these are experimental, I may modify the mechanism to avoid unintended consequences/gaming, preserving as much as possible the original intent and mechanism.
Script hash: cfd4bc01f9e89586d071533bc9ddd9be
This is a hash of the script that will make the final word decision, deterministically randomized based on the final state of the market. This is to prove that I have no discretion or control over the outcome. I'll post the script once all the markets using it are resolved.
Related questions
🏅 Top traders
| # | Name | Total profit |
|---|---|---|
| 1 | Ṁ1,520 |
In this round, the saboteurs succeeded in making the poison word the #1 most probable. But in a stroke of luck, the algorithm did not choose that word and instead we got: 'yes'.
So the haiku so far now is:
Brings bayesian slowly
Singularity yes
To be honest, I'm having a hard time seeing where this is going. But it is not yet impossible for it to be a valid, if odd, haiku. GPT-4 completes it as follows:
Brings bayesian slowly
Singularity yes, blooms
In data's vast sea
@Dilon were you waiting for close time? most of the activity is concentrated around the close, which makes me think maybe it should be a shorter market, or perhaps would work better with a random closing time
@Ansel Yeah I was, there’s definitely much less risk if you bet right at close.
A random closing time would make sense for these markets, but it would definitely be more stressful to have to keep checking it, rather than betting all at once.
