I have very complex, mathy ways to size positions. But since I get asked a lot about sizing, here is a quick, dirty approximation for how it works in practice:
First investment (1.25%): This is better than nothing. Based on my research, it has a positive expected value and its discount to its EV is substantial. Such positions are rolling off and need to be replaced constantly. If it works out well, this sized position can have a real impact on the portfolio. It is a real position, but hopefully won't get us into too much trouble.
2nd (to 2.5%): It is much better now. I liked the idea, but now that things have changed, usually with a price move, it is a better investment and I am glad that I have room to add. This happens about half the time.
2nd (to 5%): Wow. Okay. Didn't see that coming. Sort of shaken. Lost some money so far. Reevaluated all of my premises and they are intact. This was good. Now it is great. That happens one in ten times.
3rd (to 10%): I am crazy or this is a terrific opportunity trading close to its downside. Or, perhaps, both. Now it is time to let this play out. Happens rarely in a low single digit percentage of the time. But it happens.
I just look for/research/own positive expected values and such +EVs can and regularly do have bad outcomes. I want to be able to be wrong hundreds of times per year, maybe a third of the time, and still have a good profit usually, protect principal almost always, and stay more or less as rich as I started no matter what.
