EA craftsman’s EA course 【014】 Do not be fooled by the ratios such as win rate, profitability, maximum drawdown rate, and other such metrics in FX automated trading!

Ratio Expert

As discussed in the 11th EA lecture "About Revenue Amount Ranking," what matters is the ratio, not the actual amount or real numbers.

For example…

I earned 100 million yen in one year!

Even so, if the required margin to achieve that is 10 billion yen, or if there is a drawdown on the order of hundreds of millions during the process, you wouldn’t want such an EA, would you?


So, the tip for choosing a good EA is about ratios rather than the amount itself, such as how much was earned with how much capital. Capital efficiency…

We’ve talked about that, but this time there’s a point about ratios that you should pay attention to as well.

Vrije自在

For example…

Maximum drawdown less than 5%!

EA with 1000% profitability!

Suppose there is an EA that sells such things.

When you hear these numbers, do you think, “I want it! I’m curious!”?


Well, even if the drawdown is less than 5%… how much is being earned to compensate for that? What about the margin?

If you’re thinking that, it’s okay.

Sure. A drawdown under 5%… you can adjust the margin to make it work.

For example, suppose there’s an EA with a maximum drawdown of 490,000 yen.

To keep it under 5%, you’d set the initial margin to more than 10 million yen. In other words, it’s flexible.

Moreover, what if that EA earns only 100,000 yen? To earn 100,000 yen, you’d have to tolerate a drawdown of 490,000 yen along the way… that’s a very poor capital efficiency, isn’t it?

Apparent Capital Efficiency

The same goes for profitability.

Even an EA that earns only 100,000 yen a year can easily create an impression of “1000% achieved!” by setting the initial margin extremely low, like below 10,000 yen.


Of course, in this case there are barriers such as potential capital shortages along the way, but if the timing of starting the operation is favorable, it can be achieved surprisingly easily by luck.

Especially in backtests, since the drawdown timing is predetermined, you can conduct tests that avoid it, making manipulation possible.

Furthermore, by setting an extremely borderline margin that won’t bankrupt within that period and outputting the test results, you can create the appearance of a highly capital-efficient EA.

In other words, judging by amounts like “I earned this much!” or by rankings is foolish…

More than that, even ratios, percentages, and rates cannot by themselves determine the superiority of an EA.

■ My developed EA concept and operational policy

EA Craftsman’s EA (Three Arrows) is here