Following the first research DAO call and Cory’s comment on two tokens prediction market for staking I embarked on a thought experiment to allow downstaking while a proposal is already boosted combining curved bonding and reverse dutch auction. Warning: This is a quite bit confusing (to myself) and probably also to the end user, not trying to offer a practical implementation but just throwing some ideas around for other researchers to pick up.
Phase 1 could be:
A) Keeping the classical upstaking + downstaking of proposal as long as it is in the regular queue
B) Setting a curved bonding market where proposal token can go from 0 to the boosting threshold denominated in GEN, people can buy and sell the boosting token and proposal price moves up and down along the curve. If the curve reaches the treshhold within timeout proposal is boosted if not is not boosted and price goes to zero. Stakers thus pull out money as timeout approaches without boosting or probability of no boost goes up.
Once the proposal is boosted the amounts are locked until resolution of the vote If the proposal finally passes then the upstakers get their stake back+profit from having bought along the curve +dao bounty, if proposal doesn’t pass then stakes are distributed to downstakers according to the preferred mechanism in phase 2.
A reverse dutch auction starts where you can downstake by buying into the proposal dutch x token starting price being equal to the threhold amount of gen. As time goes towards timeout the price falls, at the end of the auction every downstaker clears for an amount of downstake equal to the price of downstake. You thus still can, as in the current model have downstake>upstake and multiple stakers split the upstakes.
If proposal passes having the dutch x token in a certain amount give a member (the proportion they bought/total downstake tokens bought during the auction)*number of upstake tokens
At the end you add a DAO wide bonus and/or can slap a predictor fee. T
Pros: We allow stakers to have skin in the game after boosting period is closed
Cons: The rate of decay in price time or vote dependent while proposals resolution confidence does not moves linearly with time, complicated and unclear for end user.