Re: As to providing real data I'm the only one out of us two who has. AMD's own data as to their own low product mix, analyst reports which back this up and which *directly* attribute the low asp to and finally Fab36 only producing Semprons and single-core A64s.
I asked for either data or a plausible scenario. Your data was ambiguous to the point where it could have just as easily proven my point over yours, so I asked you to refine it, or stick with a plausible scenario. Once again, the anecdotal evidence from AMD and the analysts can just as easily infer a 50% product mix to Sempron, which is much more low end weighted than Intel's Celeron mix, which is in the range of 30-35%.
Re: You on the other hand have looked at the proportion of OEM offerings which tell us nothing about what quantity each is sold in
No, but it doesn't take an expert to point out that HP offering far more Athlon options* relative to Sempron would be absurd if Athlon only accounted for 10-15% of the volumes. If you are going to justify that case, you should at least respond with a logical reason why they would do so.
* In the Pavilion line, HP has 3 Athlon offerings and 1 Sempron offering, and in their Presario line, they have 2 Athlon to 1 Sempron offering. Their Media Center line has 2 Athlon options to 0 Sepron. Your scenario would imply that 85-90% of volumes comes from 2 out of 9 product lines. That's a huge disconnect.
Re: Finally the official price list which is forward looking and perhaps tells us more of what next quarter would be like and tell us little about what parts in vendor's inventory has been sold in the last quarter. You would have been better of looking at a price list at the beginning of last quarter if you are going to play this asp game in which you do make some rather elementary fundamental mistakes on your basic assumptions.
Fine, but the earlier price list shows even larger prices on some of their X2 and Athlon 64 skus, which would have lifted ASPs even further. Go ahead and compare. There are very few major differences: