What is the most robust investment strategy? – Voice of the People
What is the most robust investment strategy?
That is the set of investment rules that has proven to be profitable with minimum risks (drawdowns) over prolonged periods of time.
Do those rules exist? Yes, they do.
Just back test the total collection of 14,022 active and inactive stocks in Research Wizard and see that these 14,022 stocks combined compounded 7436%/year with a maximum drawdown of -26% over the past 12 years. That scan takes a full 10 minutes on my new laptop and immediately reveals the errors in the Dbase. But even without these errors, this is the most profitable and safest portfolio approach ever.
But it is not practical. The portfolios are equally weighted and the smallest stocks that only trade a few shares are pushed up by buys much more so than being pushed down by a sell. That asymmetry in those very small stocks is responsible for these tremendous gains, not any fundamental. When you take out these tiny stocks by imposing daily dollar volumes in excess of $500,000 and share prices larger than $1.5 and cap sizes larger than $100 million, you are left with 3446 stocks that suddenly only compound 6.4%/year with the high risk of a maximum drawdown of -60%. When you take the 200 smallest stocks of that collection, you start compounding 22%/year with a somewhat higher risk of -71%. There are only 200 ZR#1’s fulfilling those minimum size requirements. These 200 ZR#1’s compounded 18%/year over the past 12 years with a -54% risk. Hence, not using fundamentals gives you a +4%/year edge over the 200 ZR#1 stocks of the same liquidity.
If you would know the share price a week from today, you could make tremendous amounts of money. Just program in RW the selection rule i5[Recent+nW]/i5[Recent]>0, so that you only select stocks that increase in share value during the next n weeks (n=1, 2, 3, ….). When you apply such forward testing on earnings or margins, you don’t compound steady gains. However, you do compound steady gains when you apply such foreknowledge to estimate revisions. For instance, try i44[Recent+3wks]>0. When parties are holding back that information, they could make enormous amounts of money.
Back testing is only a reliable tool as long as your watch list of back-tested stocks doesn’t change over time. Hence, when a stock of your list becomes inactive, your back-tested results change, as this stock is not available any more. Zacks retroactively started to put back the inactive stocks in its dbcmhist in November 2009. It showed that the back-tested results of mid and large caps were hardly affected by the inclusion of inactive stocks. This so-called survivor bias is more important for the smaller caps.
So are there relatively stable watch lists from which you consistently, over 20 to 32 years, can pick sub selections of up to 50% of those stocks and then steadily compound up to 40-50%/yr with maximum risks between -15% and -30% and hardly suffer from survivorship?
The answer is yes, you can, without using any fundamental or TA analysis. Our 32-year deep back-test programs and quantitative watch-list design just proves that to be true. Does that mean that 32 years of proven past performance holds any promise for future performance? Except for using foreknowledge or flash trading, it is the best you can do. But stocks are like neutrinos. Neutrinos don’t carry any charge and are thus invisible for light waves. Pure mass is only inertial without any wave character. That lack of wave character is also with stocks and prevents you to make any future prediction with any certainty. Heisenberg’s uncertainty principle doesn’t hold for particles without a wave character. It doesn’t hold for neutrinos and stocks. You can only manipulate them when you have the instruments.
Stocks, Forex and Gold 