Does the Holy Grail EA exist?

Geko is here.

I am explaining portfolios in a series.

Last time was about diversification of brokers.

This time I planned to write about the diversification of EAs, but before that I will add some “preface.”

That is,

“If there is one excellent EA, there is no need to combine them, right?”

is my thinking on this matter.

Here, let’s refer to an “excellent EA” as a “Holy Grail EA.”

There are various definitions of a Holy Grail EA, but roughly

“win rate over 80%, annual yield of 100%”

as the performance of an EA.

I don’t particularly care about win rate or annual return during the development of an EA (I focus on long-term stable profits),

but from the perspective of operation, if it is hard to lose and the principal doubles in a year, one might want to obtain it.

However, such Holy Grail EAs are hard to come by in the market.

One reason is that developing a Holy Grail EA is inherently difficult.

EA development generally aims first to be profitable using long-term historical data (more than 10 years).

Because an EA that cannot withstand market conditions of at least the past decade is unlikely to yield lasting future profits.

In such a development environment, pursuing the performance of a Holy Grail EA tends to make the stop loss large.

If so, the maximum drawdown becomes enormous, and the probability of failure when the loss rate is 20% increases.

Of course, you can create something close to a Holy Grail EA in the short term with the risk of ruin, but such EAs will not survive in real markets for a year.

And

the second reason is that if the Holy Grail EA’s user base grows, there may be some inexplicable dynamics at work in the market.

This is more of an urban legend level discussion, so I won’t write more about it here.

One thing I can say is that,

if I were to create a Holy Grail EA, I would not publish it publicly.

From the above, I believe an original portfolio is important.