1. Lemming portfolio theory. The consensus is always wrong. Facebook is a media love orgy right now.

Lemmings at the Gate

2. Private valuations are opinion, not science. A few years ago the Winklevoss twins used Microsoft’s $15 billion valuation to settle their lawsuit. Meanwhile, Facebook’s board appraised the company at $3.7 billion. Is there more than one valuation in 2011?

3. Goldman Sachs takes care of Goldman Sachs. Is the $50 billion valuation “God’s work” or more engineering from the organization that brought us the Abacus fiasco?

4. At 100x on earnings this multiple is lost in space. Danger, Will Robinson.

5. Growth drives high multiples, but measurements are far from perfect. Here’s what Inside Facebook writes about Quantcast, a firm that tracks Facebook’s stats.

In what appears to have been an unannounced recalculation, the firm is now showing Facebook 129 million monthly unique visitors. That’s the most it has ever tracked, according to the site today. But when we looked last month, Quantcast was showing Facebook had plateaued at 137 million monthly uniques. Apparently there was some sort of methodology change.

6. IPO hurry-up strategy—prime the pump and sell the hype. Facebook’s US traffic during November fell from 133.5 million monthly visitors to 132.7 million. Growth is coming from overseas, but there is no certainty that international revenues will keep pace with those in the US. Is the $50 billion valuation step one of a carefully orchestrated strategy to go public—just in case ad income plateaus sooner than expected?

7. Goldman makes money even when others don’t. Its fees from this deal total $160 million. Its take from a Facebook IPO could be just as big. So how much of the bank’s $450 million investment is really at risk?

8. Smells bad. In order to participate, high-net investors must agree to 4 percent front-end fees, lock-ups on their stock, and zero transparency. I was a HNW stockbroker. I’ve seen this movie before, and the client gets screwed.

9. At 25x earnings, the Google comp argues for a much lower multiple. What was that about a 100x multiple?

10. Caveat emptor striga Architeuthis dux. That’s Latin, more or less, for “buyer beware of great vampire squid.”