This was brought up at Austin bitdevs last night. It's a really interesting way to create a prediction market for potential soft forks to Bitcoin without relying on centralized exchanges.
tl/dr:
- Two people make a bet about whether a certain opcode will be activated.
- Both put btc into a timelocked transaction.
- The person who wants/believes the new opcode will be activated has the transaction pay to their address using the proposed opcode.
- Other side of the bet has the transaction pay to their address using normal means.
- At the time the timelock is reached, the bet is won by whichever path the money follows.
Now you have a publicly viewable way to generate reliable signal about how much people want a certain opcode to be adopted.
(It's a lot more complicated than this, so read the write up in the link, but I think I've got the rough idea of it. Also it's specific to the CTV proposals, but maybe you could generalize the concept.)