So lets have a simple look at some numbers:
What you can see here is a initial state of chances 25% each. Usually people will trade when they disagree with the price. However - the trading size is limited because too big trades would move the market too much.
So lets assume a trader don't want to move a market by more than 3% per trade. In this case that would be a trade of $13. (based on the initial funding of $415) Furthermore we have to assume that the market maker will loose his full initial funding since trading will be allowed beyond the resolution.
Now lets assume a fee of 1% (that is what serious sport betting exchanges take). That would be $13*1% = 13cent per trade for the market maker. So to break even about 3200 trades would be needed.
Super popular makers can make this easily but a lot of topics don't. For those the crowd funding option makes total sense to me. All contribute in the common good of a) knowing the likelihood and b) having the option to make profitable trades against the MM.
(One side note: Augur wants to give half of the fees to the oracle. This would mean 6400 trades would be necessary to break even. Or to put it in different words: if the MM breaks even he pays the same amount to the oracle as the initial funding. This is ridiculously expensive for the MM for the simple services of bringing one fact to the block chain.)