This is not very interesting. Anyone who has attempted even a cursory inspection of market returns understands this rationale. The more interesting problem is

This is not very interesting. Anyone who has attempted even a cursory inspection of market returns understands this rationale. The more interesting problem is to attempt to train an investment curve based on some normalized metric for market value, which would be superior to both the lump-sum and DCA strats. What he's not showing you is that gyrations in the market have a tremendous impact on the long-run return outlook and this is extremely detrimental to the lump-sum strategy in particular. DCA is slightly better but still not perfect. You therefore want to pool cash over certain periods of time and invest more, relatively speaking, during periods when the market is trending lower. This is a tough model to construct because there are certain subjective features (how to normalize market value properly) and more complicated variables you have to account for like a negative ROI on holding cash. All that said, I actually built a model fitting this description a while back and it works swimmingly.