Today I received an email:

Hi Markus,

I bought your "The Complete Guide To Day Trading" book. In your book you said you have a setup that produce 90% winning trades. What is your setup for day trading?

It seems that this reader misunderstood what I wanted to say. I used this as an example for a highly curve-fitted trading strategy that will fail in real trading.

Here's what I say in the book:

"Losses are part of our business. A trading system that doesn't have losses is "too good to be true." Recently, I ran into a trading system with a whopping winning percentage of 91% and a drawdown of less than $500. WOW!

When I looked at the details, though, it turned out that the system was only tested on 87 trades and – of course – it was curve-fitted. If you run across a trading system with numbers too good to be true, then it's probably exactly THAT: too good to be true.

Usually you can expect the following from a robust trading system:

• A winning percentage of 60-80%
• A profit factor of 1.3-2.5
• A maximum drawdown of 10-20% of the yearly profit

Use these numbers as a rough guideline, and you’ll easily identify curve-fitted systems.

You need to understand that winning percentage is a function of the ratio between your stop loss and your profit target. If you increase your stop loss and decrease your profit target, your winning percentage will go up.

As an example:

If you set a stop loss of $200 and a profit target of $300, then your winning percentage could be 60%.

If you increase your stop loss to $2,000 and decrease your profit target to $100, then your winning percentage might climb up to 90%. But this wouldn't really make a you more profitable.The winning percentage increases, but the average profit per trade decreases. Therefore most of your profits might be eaten up by commissions.

Most trend-following systems have a winning percentage of 55%-65%.
Trend-fading systems typically have a winning percentage of 65%-80%.

Make sense?