Predictors bias towards success of proposals

Right now anyone can boost a proposal by putting sufficient upstake all at once and leave no place/time for downstaking, and thus no skin in the game to call for “negative voters” during the boosted voting period.

Suggestions:

- Order predictions: start with only upstake, and only once crossing the boosting threshold invite/allow downstaking for some short period before boosting.
- Allow downstaking for all boosting period but then grade the downstakers according to their time of stake.
- Allow staking “against the consensus” (and possibly grade staking in proportion to its “opposition”). In particular downstaking when voting goes Yes.

Check [work document](https://docs.google.com/document/d/1LX5q-lIv8w_1XJ0LJIq_MpY1r3ITT7y_pBU1JWktNhM/edit#) (DM for comment/edit permission).