I don’t think anyone at the office has told me they’re on manifold, but there’s a reasonable chance someone with insider info can bet here - something to bear in mind.

Released today

“jokes aside, this is really cool!”

People seem to be having fun with it.

The usual suspects on the strategies team have been placing dodgy bets (betting inf., nan, etc) to find bugs and finding ways to give themselves negative balances.

So far one legitimate market created by regular users who didn’t work on the market (and 7 dodgy markets deleted by admins).

I’ve already observed one bet placement battle that wouldn't have been necessary if we had limit orders (not implemented yet).

Overall going well

We didn’t end up releasing today, but it is now merged and will go out tomorrow morning 💪

Some more info.

We’ve implemented the market system in the company’s chat bot, so people don’t have to go to (and remember) a website to interact with the markets. Market creation can be announced publicly with a command sent to the bot in public channels (like !market create <yes or no question>). All commands have help commands explaining them.

We’ve put some effort into making the interface easy and simple (e.g. no mention of shares anywhere).

We’ll release some time in the next 3 working* days. I’ll post updates on release day, 1 week after and 2 weeks after but limit replies after release other than at those times.

Could you add some colour to what a "success" would be? eg what about if you get 10 markets created by one person, and 5 more create one each, each market gets around 5 bets, and activity in the second half of the month is low until you send a reminder email, at which point it bumps up to around half the previous high, and drops off again in a few days.

Or what if there's a controversial market about executive compensation, you get 100 bets, and then you're asked to remove it, and people lose interest?

I probably sound super skeptical but I'm curious about this as I've thought about doing it in my workplace but not sure what to expect.

@fwbt I won’t put numbers on it, but if I think people will continue to use it (even just for fun, even just occasionally) then it counts as a success.

If it looks like it’ll be forgotten about and abandoned (e.g. no one seems to think it’s cool, no one places any bets after the initial release) then this will resolve to no.

Something in between will be based on my subjective judgement, but the bar is higher than “was it worth the hours it took to bang out the mvp to try it” - it has to look like we’ll get some value out of it.

I’m really not expecting any controversy around it - everyone is pretty chill.

@QuestionableQuentin You put a number on it when responding to Bruno :D Good luck!

Maybe calibration would be an interesting metric?

@BrunoParga For sure, but I don’t think we’ll get enough bets placed within 30 days to plot a reasonable graph. I think a dozen markets created and traded on in that time would be a strong result (more than sufficient for Yes).

I would be surprised if 30d is enough time to tell much of anything? I guess it tests whether people create an initial batch of markets and place some bets. But there will barely be time for many markets to resolve or Check if there’s any interest beyond initial novelty

Reasons to think no:

• It’s a small company, around 30 employees. There are concerns not enough people are available to participate to make the markets useful

• Some skepticism has already been expressed before the project started about its utility to the company

However:

• Its an automated trading company, so you’d think people working here would already have some interest/curiosity

• Prediction markets are cool (?)

After the initial yes bet for motivation I won’t trade on this market