I would hardly laud that as a smashing victory -- 20 out of 31. That's still very competitive, if you ask me, not withstanding AMD hasn't even shrunk to .13 micron yet.
I wouldn't take Xbit's review too seriously. They (like HardOCP for some reason) don't think it's "fair" to use AMD's SSE patch that's available to the public. Not only that, but Xbit used a lot of SSE2 optimized tests that are used about 0% of the time in any real world or even specialized situations. Using a SSE2 optimized benchmark is not what's "bad", it's using an SSE2 optimized benchmark that virtually no one in any significant, or minor for that matter, position uses.
Oh well, I'd still say the NW did it's job, it wasn't going to beat the AXP's anyway, so I'm just awaiting other processors and platforms at this point.