RE: Kurlak as an analyst
Say what you will about Kurlak (and I've had my cheap shots at him), but he is the only analyst who makes available his computable models (something Intel itself has been unwilling to share with us even in retrospect). If you disagree with his assumptions, you can make your own. If you disagree with his conclusions, you can dig back into them and find out where he went wrong. His bad misses on Q3 and probably Q4 result from something, but what? In his July 15 forecast he projected sales of $5,950 million for Q3 and in Sept 15 $6,460 million. Where did that $510 million come from? He forecasts unit mpu sales of 21 million as of the summer, and estimates 22.8 million as of the fall. He missed Celeron's 2.8 million altogether, and this looks like one major source of his shortfall. The other, I guess, is his estimated price of $550 for Pro/Xeon, and, in the future, the forecast of $450 for 1999. Sounds awfully cheap to me.
Instead of barking at this guy, I think it would make sense to download his model from askmerrill.com (you'll have to register but it is free) and see if we cannot collectively build a more reliable model for forecasting Intel's future revenues and profits. |
