I don't know where this requirement comes from. It is completely unnecessary for a free market. The free market prices in lack of perfect information - i.e. "risk".
Complaining that free markets not being perfect is like complaining that friction in an engine isn't zero. You can't make it zero, but you can get pretty close to it.
I was using free market in a practical sense, as in a market (a pairing of buyers and sellers) that has functional supply and demand dynamics. I wasn't relying on any larger context, like an economy.
I didn't interpret WalterBright's comment as invoking some 'ideal' market either. I don't think talking about 'ideal' anything is even useful outside of an academic context, so I don't assume anyone here, that isn't explicitly talking about theory, is invoking an 'ideal' version of whatever they are talking about.
Thing is, there are generally two different things that people can mean when they say "free market":
1. A "market economy", as contrasted with a "command economy"—ie, more or less the way the US does things, as opposed to the way the Soviet Union did things.
2. An "ideal free market", as described by Adam Smith, which can, in theory, ensure that many systems find a stable, efficient equilibrium that balances consumer desires with producer desires.
The problem is that, however much many (particularly of the Libertarian bent) wish to believe in it, the latter is not real. It is a thought experiment, and it requires a bunch of conditions that don't always apply (eg, perfect information, commoditization, etc). The other problem, of course, is that as I said, the two are often called "free markets" interchangeably without clarification, which leads to much confusion.
You seem to be referring to #1 when you say "there are free markets within the US"; however, what WalterBright is talking about is clearly #2, since he's referring to ways that idealized free markets theoretically operate.