This article is determined mainly for beginning traders because
experienced traders know traps of backtesting process very well. But as I
found, there is not very much information about it in the literature or
web pages about trading. Everywhere you can read that
backtesting is essential for your trading and that a lot of people are
going to underestimate it . But to be honest, I can’t remember a single
good article where was written “How to backtest” and of what to be aware
during this process.
My thought on this theme starts when one friend of mine asked me to show him how to trade. But trading isn’t something what can be learn this way. You can’t think that you can watch some traders few times in action, learn his/her strategy and just copy it and be a good trader.
No no, not even exact science like mathematic can be learn properly so
easily. So I tell him that he should stop just reading about trading and
start doing it. Start with demo account, place some orders and start
with backtesting some basic thought (when you beginning you can start
with whatever you want).
But beginners do not believe that trading can’t be learnt easily. So he wanted me to show him my system. And he wanted to backtest it on his own. Those of you who already have some profitable system know that even simple system can be quite complex with all its details, filters and so on. So I show him my basic thougt. I have already known that this thought in its basic form cannot be long term profitable, but I did not tell him this fact.
Few days later he called me back with a big smile. He tested whole
idea on one year period (it is intraday system) and he was very
profitable according to his bactest. It sounds strange to me because I
know that it isn’t profitable (after spread deduction) during this time
period (according to backtest of mine). So I asked him to send me his
backtest file. He had only few trades during first two months and almost
all of them was winners. Another three months have much more trades and
ration between winners and loosers was much
equal. Then there was some months with few trades (mostly winners) again
and last two month were full of trades – mostly loosers. What does it
mean? You can tell that there are different periods during the year and
sometimes you get more winners and sometimes more loosers. You can
uphold correctness of this backtest by lot of common rights. But in fact
none of these upholds will be right this time. Because friend of mine
fall in something what I call „Backtesting Traps“.
And what are these Backtesting Traps? I don’t want to write about well known facts like sufficient number of trades, neccessary papertrading after backtesting, benefits of manual backtesting or anything common. Backtesting traps are something else. In fact it’s simple process of learning. At first your mind must be „set“ to see your patterns properly before you can do your backtest properly. And this part is something what lot of traders mostly underestimate. Then you can hear/read things like „Trading always looks great only in backtests…“. There are three Backtesting traps which you need to overcome to be sure that your backtest has some value.
First Trap
You are just learning to see your patterns when you start bactesting
some strategy. And it takes some time (count of trades) before you can
see your pattern in all possible shapes. That is to be reason why my
friend had only few trades during first backtested months. He
just did not see it in all shapes, but after 50 trades he began to see
it in various situations and that is why the number of the backtested
trades rose.
It is OK and sometimes we can learn from it that some “shapes” of our
pattern are more successful than others (this is how are created
different filters). And why is it a trap? Because when we backtest for
example 12 months period we can make some evaluation. But usually first
1-3 months of our backtest are completely useless for such thing because
of this “learning period”. It is very simple to avoid this trap. Just
retest all your backtests once again. I know it takes a lot of time. But
it is definitely worthy. After this retest you will
see some very important things. At first you can be sure that your
backtest is more accurate and at second you will see situations where
you decide in the different way than before. And these situations
usually make the problem dutiny the live trading. Focus on these
situations and decide what to do in them.
Second Trap
Second trap is simple to describe, but it is harder to avoid it. Many traders are usualy focused on the winners situations during backtesting much more. It sometimes leads to the situation when you overlook some looser trades and it leads to the other distortion in backtest results. Try to be very objective. It is better to take more losers (at unclear situation) during backtesting then to ignore some of them – there is no delete or back button when you lose your live trade. This second trap results in much more winners than losers in backtest of my friend.
Third Trap
Last trap is the worst one. And it causes Trading always looks great
only in backtests…“ very often. Even if you are very objective during
your backtesting, even if you can recognize all shapes and possibilities
of your pattern, even then there is some trades which you usually miss
(and they are unfortunatelly mostly loosers).
Why? Because it is sometimes very hard to recognize all places where you can put trade when you looking on whole picture (in the middle of the graph) – especially when you have routine in backtesting your strategy, then you are looking faster – and less accurate for this situations:
Try to imagine that you are entering on cross of moving averages. On the left picture there is no entry at all. But on the right picture it doesn’t seems so obvious. This is the biggest trap during backtesting. And it is almost impossible to avoid it by looking in the middle of the graph. If I want to avoid it then I look only on right end of the graph and then pushing graph candle by candle and writing my trades down. This procedure helps me to be objective (I can’t see how trade ends) and let me avoid third backtesting trap. It prepare me better for live data.
These three traps are main reason for bad results of my friend’s
backtest. And also reason for lot of „Trading always looks great only in
backtests…“ situations. Try to think
about these Bactesting traps during your next backtesting. I hope it help you improve your trading.
I wish you lot of pips and success
