Jon, thanks for your comments (see comments on previous post). I certainly understand and agree with your premise that diversification greatly reduces risk, but I could not have made the case in as concise and detailed an argument. It may be time for you to write a guest article. If interested, just send it in an email and I'll post it.
Your example uses just two stocks vs one, so imagine how significant the decrease in risk if you consider thousands of stocks from different industries with different market caps and from different international economies. And, this extreme diversity is easy and cheap with todays index funds. This is why I indicated I'm comfortable with a higher percentage stocks in my portfolio than conventional wisdom allows.
I'm struggling with one point you make, which is that your average return is only slightly decreased. From what I know, I think there is no decline in average return unless you can predict which stocks will have higher average returns in the future. With the efficiency of markets today, I'm not sure that is possible to do consistently.
I'll point out another thing to worry about, though. Your example talks about two negatively correlated stocks, and the case is clear for that example. Unfortunately, I have found that most stocks worldwide have a significant degree of correlation, which is the basis for my trading in international stocks. I've discussed this with many readers, although I have yet to post a blog on this topic. And, the real trick to understanding the decrease in risk is predicting the amount of correlation of your assets. The big worry is underestimating the risk by underestimating the correlation.