It actually is ideal, given trusted API/Oracles, a timeout period to ensure mis-published data is corrected promptly, and a few other non-obvious gotchas.
You could even have trusted middle-men that collect a Fee iff ambiguity is discovered and an appeal is wagered against the escrow.
Augur tried this and doesn't seem to have gained a lot traction doing so. Any staking schemes that were developed were relatively easily gamed and that led to distrust in the system.
We, like traditional financial exchanges, define the contracts upfront with clear rules on how they will be settled, then follow the letter of the law very strictly when settling a market.
In general, we've found market certainty to be more important that accuracy: ie. having pre-defined rules that everyone can agree on and that are pre-set is more important than those rules being the "correct rules" (not that having correct rules isn't important).
If it is set up so that the users can put in bets and appoint judges, then it can't ever be rugpulled. Individual judges could turn hostile. But if the judges get paid for their work, a great track record might be worth more than any rugpull.
You could even have trusted middle-men that collect a Fee iff ambiguity is discovered and an appeal is wagered against the escrow.
I'm sure it exists and got rugpulled already.