I will run a script once per day that does the following:
Draw a random number between 0 and 999 (inclusive).
If the number is 4: Resolve this market to the Bitcoin price in Euro (BTC/EUR) at that time (queried by API).
If the number is not 4: Do nothing.
In other words, every day, this market will have a probability of 1/1000 to be resolved at the (then) current Bitcoin price in Euro. If the market is not resolved by resolution date, I will resolve it manually to the then-current Bitcoin price in Euro. If the resolution criterion gets triggered and BTC/EUR is outside of the range of this market (0-1,000,000), this market resolves to the next closest bound (e.g., if BTC/EUR is 2,348,368, this market resolves to 1,000,000). If BTC/EUR leaves these bounds and no resolution criterion is triggered, nothing happens.
My hypothesis is that this market should track BTC/EUR pretty closely if Manifold is efficient. But I probably haven't thought all possibilities through, so interested to see what happens.
Related questions
@Quate Started it at the middle of the range to incentivize some trading. Not sure, maybe I should have initialized it with the current value? On the other hand, I'm pretty sure it will get down to the true value quickly. Right now it's free mana, basically.