There are disadvantages (you need to "sync" with usd price close time) but it's not nearly as bad as you're implying
I still haven't wrapped my head around the details of how combinatorial prediction markets compensate for the betting asset's price volatility (despite reading
Paul's papers), but it still seems clear to me that it is the inferior solution compared to betting with stablecoins directly. For one, collateral levels backing BitUSD can continuously be updated over long periods of time as the collateral asset price drops significantly over this time. I am not sure how this would work in a combinatorial prediciton market, but I would imagine there should be some lower bound for how low the price could drop from the initial price at the start of the PM before the peg would break. Second, there can be considerable delay from the date at which the USD price is reported and the time the payments are actually settled. For example, in Augur it seems this delay is 2 months. Obviously the reporters cannot report a USD price that has not yet happened, so that means the earliest access the winners will have to the collateral asset is 2 months after the fair settlement price was determined. In those 2 months the price of the collateral asset could have dropped even more relative to USD. Also, they would be forced to dump it in exchange for USD as soon as possible after settlement to no longer be exposed to the price changes of the collateral asset, whereas with BitUSD the holders can take their time to incrementally trade fractions of their BitUSD holdings into USD (via the collateral asset if necessary), assuming they even want to bother when they already have the option of just holding it as BitUSD, without needing to worry as much about slippage in the market.