OK, thanks. Now I understand your intent. I think this broad conclusion is missing consideration of a few parameters, like standard deviation and dollar-weighted rate of return.
A 100% stock portfolio has a much higher standard deviation (magnitude of ups and downs) than a blended portfolio. This doesn't matter much for someone with a 10 or 20 year horizon, either investing regularly or just letting the portfolio coast. Rip Van Winkle wouldn't have been bothered in the least. And it's statistically the highest yielding portfolio. BTDT, got the tee shirt.
For someone who is drawing from the portfolio, however, 100% stocks is a potential disaster. This person is forced to sell stocks whenever they need money, including at such inauspicious times as October 20, 1987. (The market went down 22% on October 19 in one day, the equivalent of about a 5,000 point drop of today's Dow.) Other drops since then have been slightly less dramatic but 2007/8 was numerically worse. The effect of these draws is this: That money will not benefit when the market recovers because it is no longer in the market. This means that the dollar-weighted rate of return that the investor receives will be less than the time-weighted return that you are focused on. Maybe much less.
For people who are willing to ride the bucking standard deviation horse, the best strategy I have seen is this: Forget ratios. Figure out about how much money you will need from your portfolio in the next three or four years, put that bucket of money into cash, cash equivalents and maybe a short bond ladder. Then, as time goes by, replenish the bucket when times are good (like now) and use the bucket as a cushion when times are bad. Admittedly, this is a little like market timing but of a very mild and low risk sort. It is arguably better that being forced to sell stocks regardless of market conditions.
IMO you ought to re-think your 10% and 5.5% numbers too. I don't know what period you used to come up with these numbers but from an historical perspective 10% nominal dollar return is unrealistic. IIRC a linear regression since 1929 will give you something around 7%. Also, if your period was dominated by the last 10 years of near zero fixed income returns this, too, is atypical. If history is any guide, both stock market yield and fixed income yield will regress to their long-term means. This regression to the mean does not detract from the 100% stock argument per se, but it does make the numbers far less exciting.
