Why 42 average strategies are safer together than apart

I run 42 trading systems on one account. First I measured each system alone. I took each one's worst losing period and wrote down the loss. Then I added all 42 numbers together. The total was 4,192 USD. Then I ran the same 42 systems together, over the  same period, with the same trades, as one portfolio. I measured the worst losing period of the whole account. It was 432 USD. That is 9.7 times smaller. I did not change any code. No system became better because it joined the group. Only one thing happened: their bad weeks did not arrive at the same time. 

This effect has a name. It is called diversification, and Harry Markowitz described it in 1952. It is the most boring edge in trading. There is no secret indicator and no clever filter. You only need systems that are different enough from each other. When the breakout systems lose money in a quiet market, the systems that buy dips are earning. On wild days it works the other way around. Two honest notes. First, the combined account still loses money in some weeks. The losses are simply small enough that I can wait without stress. Second, this only works if the systems are truly different. If you run 42 copies of the same idea under different names, you gain nothing. How to measure that difference is a topic for another post. I trade this with my own money.

Over the next few weeks I will publish the rest: how to measure the difference between two  systems, why I do not use my historical drawdown for planning, and what a truly diversified portfolio looks like as a picture.