Choosing individual stocks without any idea of what you're looking for is like running through a dynamite factory with a burning match. You may live, but you're still an idiot.

I wait until an investment idea is so good, it hits me over the head like an anvil.

My goal is to buy a company at a low multiple to normal earnings power several years out and that the company earns good returns on capital at that level of normal earnings. A holding period of more than one year also works quite well as the factors are persistent in years 2 and 3.

Here is part of the tradeoff with diversification. You must be diversified enough to survive bad times or bad luck so that skill and good process can have the chance to pay off over the long term.

The more confidence I have in each one of my stock picks, the fewer companies I need to own in my portfolio to feel comfortable.

Over the long term, despite significant drops from time to time, stocks (especially an intelligently selected stock portfolio) will be one of your best investment options. The trick is to GET to the long term. Think in terms of 5 years, 10 years and longer. Do your planning and asset allocation ahead of time. Choose a portion of your assets to invest in the stock market - and stick with it! Yes, the bad times will come, but over the truly long term, the good times will win out - and I hope the lessons from 2008 will help get you there to enjoy them.

The big picture is: the main thing you should be concerned about in the future are incremental returns on capital going forward. As it turns out, past history of a good return on capital is a good proxy for this but obviously not foolproof. I think this is an area where thoughtful analysis can add value to any simple ranking/screening strategy such as the magic formula. When doing in depth analysis of companies, I care very much about long term earnings power, not necessarily so much about the volatility of that earnings power but about my certainty of "normal" earnings power over time.