We have had, however, suggestions earlier in this discussion from others that appropriate buy and sell signals might have helped in getting one out of companies which had become overvalued, thus preserving profits, and allowing one to get back in once the bubble was deflated.
I think almost any valuation or pricing tool you can name can give you that result. I say "think" because I can only claim "knowledge" of whatever I personally test or study. But I can't see any fundamental reason why most ot them should not offer that benefit. However, I reiterate that none of them will maximize profits, nor will any of them result in much nimbleness. A rock tied on ot the end of a stout stick is an improvement on a hand-held rock, but it still isn't an impact driver.
I am suggesting that a large scale study is probably using standard published P/Es and such and thus is analyzing numbers which I believe to be highly unreliable since they have not been subject to the kind of scrutiny and selection discussed on this thread.
Sure, individual numbers are bound to be inaccurate (especially for earnings). But in the composite, what is the error? If it is random, then the trends hold. If it is non-random, the trends will hold unless there is a very negative correlation. Whatever - the conundrum faced is that if large scale, you have more error with an individual data point; if small scale or over a selected subgroup, you have statistical problems and possibly confounding variables. This isn't and won't be like engineering. That is about the first thing an engineer or scientist will notice. It is important to explicitly acknowledge that. Afterwards, we can think about whether we can improve our hand-held rock by tying it to a stout stick.
Hmm....maybe a hammer isn't the tool to use for comparison. Maybe it is better to say that a fixed wrench or a socket is better than an adjustable wrench, but will never be a torque wrench?
- Pirah |
