One theoretical barrier to wider adoption of joint markets is that the marginals may resolve N/A. If there were some way to guarantee automatically that this would not happen, then joint markets could be automated more.
1,000
3.00This market suddenly looking very relevant with the drama around /MartinRandall/will-ai-wipe-out-humanity-before-th-d8733b2114a8
@Mira The issue for this with me is that no market is really supposed to be intended to N/A, so how should a creator know whether or not to do this?
@BoltonBailey That is definitely not true. I have markets with clear triggers to resolve NA and others that state "this market does not resolve NA". So at least some markets would be tagged and others untagged.
NA: /Mira/if-mira-trains-a-transformer-model
No NA: /Mira/will-a-prompt-that-enables-gpt4-to
For my Sudoku market, I could clearly tag it as not NA'ing.
For my chess market, it used to be eligible for NA until I declared recently that I would be doing the project. Then it could be tagged as not NA'ing.
They know by asking their brain what payout structure they're creating when they make a market. Under what circumstances do which people make money? What should the leverage factor on X event occurring be? etc.
If they don't know whether it could NA, then certainly the "API" isn't going to figure it out by reading their description and tell you. So those markets are a lost cause. But, given that the creator acknowledges a case where it could NA or declares such cases will trigger YES or NO instead, they could just tag the market.
@Mira Ok fair enough, I was being dumb and didn't consider conditional markets. Still, it seems like sometimes there are cases where the market creator ends up N/A ing a market they don't expect will N/A. I was thinking this would be more of a thing where with some form of mod approval or trader consensus a market could be marked this way, but maybe it would be good to just let creators commit to it as you say.