Yes on one hand, the investment performance is remarkable. On the other hand it shouldn't be, because it's exactly what the CFA program teaches. It just so happens that I'm apparently the only CFA Charterholder that practices exactly what's taught. Everyone else went off to get rich being a hedge fund manager or creating derivatives like CDSs that started the current crisis.

What the CFA program teaches in a very small nutshell is this: Individual investors should not be market timing at all, period, as this is worse than futile. Individual investors should also not be trying to pick stocks (bonds or anything else), unless they have quasi-insider information, like they work there, used to work there, or know someone working there that's not technically an insider that's feeding them accurate and timely info. The only "people" that have enough information, resources, and data to profit by picking stocks (and/or time markets) are large institutions like mutual funds.

So once that huge battle is understood and won (99% of investors still don't get it), and since there are only three ways to make investment decisions, and the most popular two are taught to be no-no's by the CFA program, there is only one methodology left - asset allocation. So I took that to mean that this is what I should become an expert on, because my biz is managing money for investors in a RIA and financial planning setting (my job as an employee at the time).

What you're supposed to is determine a mix of viable asset classes that fits an individual investor's life, and then either fund it with something very diversified like mutual funds, ETFs, or index funds (the CFA program likes index funds the best, as most people can't even pick open-ended mutual funds well enough to beat an index fund). So the questions then become, how do you determine what asset classes to use and how much? Well, I did it in both the Comprehensive Asset Allocation software and the Model Portfolios using the best asset-level portfolio optimizer, educated guesses, and way too much trial and error.

After doing it a million times, I found what works and am sticking with it. I've looked at every else's harebrained investment strategies since the mid-'80s, and found none to be better (I admit I've stolen many ideas from many people (mostly from the hundreds of job interviews I went on), and so I've tinkered with just about everything everyone else has come up with since 1986. The ideas that work, I keep, and the ones that don't, I poke fun at on the site).

Once that war is over, the battle is reduced to what to use to fund the asset classes with? To make a very long story short: Stocks - no, no diversification compared to a mutual fund that may own 100 stocks. Closed-end funds, no - there's no way to screen them to get any predictive value whatsoever because of the large and random premiums and discounts. ETFs, no they are just lame index funds... with fees (and then on top of that, commissions to pay when you buy and sell them. This is all terrible compared to a no-load mutual fund). Index funds, maybe, but the way I pick open-ended mutual funds creams index funds by way too large of a margin about 90% of the time. So all of these questions are now answered - the way I screen and pick mutual funds adds by far the best value to the investment process than anything else I've ever seen, so I'm sticking it with it to the bitter end.

The magic isn't all in the infrastructure (the software), it has a lot to do with finding predictive value in the mutual fund picks (I look at many things, like historical performance, and try to predict that it will keep up long enough to be useful). After millions of trials and errors, I found a way to pick mutual funds that have a one- to two-year predictive value (all mutual funds crap out eventually and have to be replaced, which is why it's critical to keep the subscription going and rebalance. When I say crapped out, I don' t mean it in a smelly sense, but the casino gambling way). It doesn't always work, but the winning picks beat the duds enough to add more long-term value than any other investing strategy I've ever seen. Plus every month I learn something new, so the screening process is continually refined and saved, so it gets better all the time. I had a major breakthrough in '07 when I figured out how to get around the software stupidity and lousy data Morningstar maintains. It seems to be working so I'm sticking with it.

I used to sell these little Morningstar screening filter mcr files that can be used by others in their offline version of Morningstar, but a big NY firm tried to copy my screening process in '03, so I don't sell them anymore. So if the information about how I pick mutual funds is not on the mutual fund recommendations page already, then it's a trade secret now. There's no secrets about the software nor models. Just the opposite, every tiny little detail is explained ad nauseum somewhere on the site. And when you buy the allocation software or model portfolios, not much is protected, so you can see exactly what's going on. Plus you'll get the same spreadsheet I use to calculate returns to account for mutual fund switches, fees, rebalancings, and all that since 1/1/99. So you can audit the returns and figure everything out.