I mostly agree, I just feel the need that valuation metrics rely on past values, which aren't as useful depending on the company you're talking about. If a company has net income X, but a new high margin product launching next month that could easily double or triple that, their P/E will look ridiculous but , in fact, be quite reasonable.
The reason to use past values is because they're known. Anything anticipating the future is speculation, and relies on many assumptions. Check analyst earnings estimates for S&P companies for 2008 and you may be surprised what you find.
I fully believe Tesla can be the number one auto maker, and get self driving working to a sufficient extent to be marketable. But what's the upside? A few X at most. To me that risk/reward is pretty awful... many more attractive companies selling at 1x PS and fast growth/smaller market cap that can go up 10x easily over a few years
