"About your observation about there being a big premium that daytraders miss by not holding overnight risk. You should take a closer look at your data. Re-run the spreadsheet and this time start from about September 2002. You will see the relationship almost breaks down. Shorting overnight or going long over night yields more or less the same return. Before 2002, the point you made on your blog was quite valid. The interesting question then is why this relationship changed. Presumably, the answer is because most daytraders were eliminated by the bubble bursting." - sc

This is a great observation, but it only shows half of what is going on here. September 2002 was a major market bottom, so let's also look at the results for the person who carried risk through the entire period, September 2002 until present.

• Buying the close and selling the open: $100,000 becomes $100,598.40
• Buying and holding: $100,000 becomes $171,178.80

This tells us that the public was not willing to pay a premium for overnight liquidity. We might conclude that they were frequently getting out of positions and taking small gains because they were willing to pay premiums (higher prices) to people who for the most part bought and held and in doing so improved liquidity over a four year timespan. The people who did most of their buying near the market bottom bought when supply was relatively abundant and demand was relatively scarce...