For events where the likelihood changes not constantly but very big at one point in time automated market makers with an uptick protection might be more efficient.
Lets say we have an event: Will an earthquake happen in 2016 in San Francisco. Or: Will the FED increase the interest rate.
This events can change the likelihood of one outcome to "1" in a second. Sure - the one who "reports" on the outcome first by buying the winning shares should rewarded but it might not be necessary to give him the full remaining balance of the market maker funding.
The market maker funding should make it easier to get markets started but there is no reason to "trough money away". So it would clearly be favorable to have some kind of uptick protection.
One mechanism could be that a price movment is limited to x% - lets say 5% per hour. If this limit is reached the market maker freezes and people can do bidding on the max price they are going to pay.
So lets take the earthquake example. The price would be constantly at 4% and than surprisingly an earthquake happens. The one who realized it first (and still has a internet connection and no other trouble) would buy shares of the winning outcome and move the prive to x + 5% = 9%. Than the market would be frozen. Normally people could buy shares at the price of 0.09 in this state. Instead there will be a bidding period where people do a bid for shares. Every bid would extend the bidding period. So if at least two people are bidding the should move the price close to 1.
One open questions:
How should trading continue if the bids go up to 0.5 but not to 1. If trading would continue at 0.5 the market maker would lose its property of "bounded losses".