Meaning of 70% win rate
Gekoo here.
Among the readers who have reached this column,
there may be people who say,
"The automated trading system (EA) is my first time~"
So I thought I’d write a multi-part article with little EA knowledge for beginners.
This time is a continuation about the EA win rate.
In the backtest of EAs being sold on Gogojan, there was data showing a "win rate in the 70% range."
https://www.gogojungle.co.jp/systemtrade/fx/39494
Let’s consider what it means for this "win rate in the 70% range."
*Regardless of EA performance, we will only discuss "win rate" and [TP/170 pips] [SL/130 pips].
A win rate of 70% means that out of 10 trades, 7 will win.
If you trade USD/JPY with 0.1 lots (10,000 units) at 25x leverage, with margin around 6,000 yen at minimum, let’s do a quick calculation.
For the 70% win rate, suppose a sequence of 3 consecutive losses (SL*3) followed by 7 consecutive wins (TP*7).
- SL:-13,000 yen total -13,000 yen
- SL:-13,000 yen total -26,000 yen
- SL:-13,000 yen total -39,000 yen
- TP:+17,000 yen total -22,000 yen
- TP:+17,000 yen total -5,000 yen
- TP:+17,000 yen total +12,000 yen
- TP:+17,000 yen total +29,000 yen
- TP:+17,000 yen total +46,000 yen
- TP:+17,000 yen total +63,000 yen
- TP:+17,000 yen total +80,000 yen
On the 3rd trade, the maximum loss occurs, and you reach break-even on the 6th trade.
In this case, the minimum required margin would be [39,000 + 6,000 = 45,000 yen].
(Because of margin maintenance requirements, in reality you need even more capital to have a buffer.)
So, what happens with the common high-win, low-risk pattern of "small gains and large losses"?
Let’s calculate with [TP/50 pips] [SL/100 pips].
- SL:-10,000 yen total -10,000 yen
- SL:-10,000 yen total -20,000 yen
- SL:-10,000 yen total -30,000 yen
- TP:+5,000 yen total -25,000 yen
- TP:+5,000 yen total -20,000 yen
- TP:+5,000 yen total -15,000 yen
- TP:+5,000 yen total -10,000 yen
- TP:+5,000 yen total -5,000 yen
- TP:+5,000 yen total 0 yen
- TP:+5,000 yen total +5,000 yen
By the 10th trade, you finally reach break-even.
As this example shows, simply looking at win rate without context isn’t very meaningful for trading.
I think it’s better to consider not only win rate but also the "maximum loss," and it’s crucial to understand how much capital is needed.
Of course, the calculation of maximum loss is rarely determined only by [TP] and [SL]; you need to look at the entire logic.
In automated trading systems (EAs), you can estimate required capital from the "maximum drawdown" item in backtest data.