The intel guy (Bob something or the other) gave a lecture at stanford a couple of months back.
He explained basic econ, there should be 3 distinguishable groups, and people will most often pick the one in the middle.
For example Celeron<p4<EE. (though EE is too expensive to be even considered correctly)
People will see the celeron, think meh?, then see EE and think way too much, and settle n the p4.
In the case of newcastle, its a64<fx. There is no real middle. In that case a64 will get less money than say the p4 setup, as there is a chance because there is no middle ground, that the buyer may just skip the buy.