cpu speed is king vs high fsb/htt or loose timings
That's true in some situations...if we're assuming that using a divider is going to have a higher fsb then running 1:1. When fsb speeds are constant that isn't true. It's also not true for 10-15% above 1:1 in many cases. For instance, when i oc'd my 3800 to 2.7 on a divider, it scored very similarly to my 2.5 1:1 score in 3dmark 05. Also, in comparing other systems to my own system (1:1), using the search tool provided by futuremark, there are 0 systems that score better then i do with the same parts and a divider (while having a higher cpu speed then i do).
Parts to search by:
X2 processor ALL
6800 series in SLI
The only scores the beat mine, are 1:1.
Outside of 3D type benches, like SuperPI, this doesn't hold as true. It's more like 5-10% instead of 10-15%.
So while what you say has truth in it, you might want to add a little information, so you don't accidentally misinform people.
3dmark means nothing, ingame tells a better tale vs that program.