When I look at the formula p - (1-p)/E, and we take p to be .5 (coin flipping), there is indeed a chance to win in this game, because of the occurrence of E. I thought you meant coin flipping, a dollar for a dollar. But apparently (and following the link) there is actually somebody who wants to bet more than one dollar against my dollar! Of course you can win at that game! The surprising part is that this seems to appear to you as something surprising (or something that should surprise us...)
The formula is intended to give an optimum amount to risk, given a certain probability and a certain return. Very well. Now I don't suppose there are a lot of cases were p is close to 1 and E is high. In any case, it is easy to see that from the formula follows: never risk a greater part from your money than the probability of winning, because (1-p)/E is always positive, under reasonable assumptions. So what we see here is that you don't invest everything, but stay partly in cash. That is a very well known form of risk management in investing; too much volatility may kill you. I don't see where this flies in the face of standard investment practice and/or wisdom.
The interesting thing is the optimization of the amount to risk. Now that is very, very nice.
But didn't you mention earlier, regarding risk avoidance, that there should be companies with p = 1 (or very close to one) and high E? Looking for such stocks seems to be a rather different strategy than the one outlined here. Or did I get you wrong?
Regards,
Karel
