In my opinion, other than knowing how to bid if you're a potential buyer, the type of auction doesn't matter.  Mathematically, it tends to wash out in the end (for instance, people bid more if they know they won't pay the highest bid), unless people have access to unusual information.

The most important detail governing the success of auctions is the size of the market.  If you don't have a critical mass of bidders, the power of the auction starts to break down.  That's why there aren't millions of small online auctioneers instead of one eBay.  If you want to sell your life-size Ronald Reagan cardboard cutout, you need to find people who have the same (or more) interest.  A healthy auction community brings you together;  a too-small auction community doesn't. 

One acceptable proxy for very large volume is a large targeted volume, where the bidding audience is spiked with hand-picked, interested parties for each of the items.  In my opinion, it's the job of a high-end auctioneer to go out and bring those "qualified" bidders to the table.

One concern I had about all of the domain auctions -- which is still a relatively new market -- was "where are the premium domain buyers?"  Secondly, do non-professionals -- who might want to buy a particular brand for their real-world business -- know where to go to find these domains? 

Others seeem to share the same worries.  In his fascinating post "domain roundtable scorecard" reviewing the Domain RoundTable conference, Adam Strong at Domain Name News says:

"In his predictions, Jay [Westerdal, CEO of Roundtable organizer Name Intelligence] expected heavy bidding and higher prices as the names were ran up, but the reality wasn't like that prediction. Nearly 3/4 of the names he wrote about simply weren't sold. The total results at the auction were 167 out of 450 domains sold, meaning 37% were sold. As was the case with the other recent Moniker auction events, many of the domains at DRT sold right at the reserve prices. Sellers are being encouraged by Moniker andDRT to lower reserves in order to "get more bidders". Getting more bidders shouldn't come at the cost of the person selling the domain though, should it? That's the job of the auctioneer or event producers. The game plan to lower the reserves to stimulate activity certainly doesn't seem to be doing that. In many cases names are selling at their reserves with only 1 bidder."

Adam points out that Moniker's Traffic auction, which was probably the most successful large-scale domain auction (by revenue, and % of domains sold), only sold 54% of its inventory, so there's obviously still not a huge amount of liquidity out there.

Says peter_stargazer on the domaintools.com site:

"If i would go ahead and set the reserve to the minimum $1.000 in order to maximize the odds for making it into the auction as so often is recommended, i could in theory recive lots of bids and in the end selling it for more than $20k. But from personal experience with other high profile domain auctions. I have learned the hard way that as a seller, i MUST Assume that my domain will only receive one bid thus selling for the minimum reserve. And therefore i need to set a reserve price that wont make me cry. And if this means that no one will bid on my name... So be it. In the end a domain is only worth what someone is willing to pay for it. It's the eternal debate of setting the reserve to high or to low..."