But first we should talk about backtesting. I do backtesting in three passes. The first pass is called eye-test. It’s quick. You need to draw your pattern and see if it works or not. The second pass is optimization pass. You change something in the pattern and see how changes affect the profitability. This pass may have multiple cycles. For example, I want to know how bats profitability changes if B doesn’t touch 0.50XA. I need to test the history period and find bats with this criterea. It’s cycle number 1. Then, I want to know how bats profitability changes if C-point goes beyond 0,886AB. Again, I test the same history period finding bats with this criterea. It’s cycle number 2. And so on. Whatever idea about the pattern you have you test in this second pass. Alternative entries, alternative exits- all rule of engagement are tested here. And finally, when I have all the ROE tested I do the last, trading pass. It’s done as if I was really in that moment trading this bat)). I draw equity curve, write down the time of trades, count average win and average loss, profit factor, expected payoff and so on. In this pass you mostly test your money menagement strategy. If you want to optimize your strategy, again you will need several cycles). After backtesting is done print euity curve drawing, print it and place it right in fron of your eyes. This curve is your ideal result. You are to compare your account equity curve with your backtesting equity curve to find similarities and differences. If the curves are different – it means you do execution errors, you are trading NOT the same way you have been trading in the third pass.
My realtime trading equity curve was not the same as my backtesting equity curve. I mast have been doing some execution errors.
You can read more about trading errors here:
After correcting my trading style the problem didn’t dissappear. Equity curve looked bettern but not as well as backtesting curve. Then I began to scrutinize my backtesting trades. What was so special in my backtesting trading that I cannot repeat in real life? And I found it. It was frequency of trades. In backtesting I had 1-2 per day. In real trading 3-6. I began to study my backtesting more carefully and found lots of bat patterns that I didn’t noticed when I was doing backtestings. After checking these bats I found most of them to be loosers. Why the heck I didn’t notice them? The answer is in the mindset I had when back testing. I just looked at the chart and picked up bats that did catch my eye. I was relaxed and had fun. Whereas in realtime trading I usually FORCED myself to find bats (especially after a series of losses: “I NEED some bats to enter the market. If I don’t enter the market I won’t get my money back”). When you force yourself to SEE a pattern, stalking every tick, strained you will trade EVERY pattern coinsiding with your ROE. But when you are relaxed, feel fun and joy, your mind blocks your perception from seeing those patterns which are typicly loosers. It’s quite simple. Every looser causes emotional pain. Subconsciously you know what type of pattern causes more pain. Your mind is simply tries to aviod pain distracting your attention or somehow else. This type of pattern just don’t catch your eye. If you don’t see anything in the chart at first glance then skip it. Do not try to force youself to see something.
Continuation is in comments...
To be successful in trading, you must have CLARITY OF THOUGHT! When you are trading a lot and following too many pairs, you will find that you will have much more confusion, anxiety and worry. All enemies of good trading.
I worked out five criteria to distinguish bat pattern types:
1. It’s quite simple. Is a bat bullish or bearish?
2. Is XA a straight move or is it two-three impulse move? If it’s a straight move, then the pattern is Simple (S). If not then we’ll contain Head and Shoulders (HS) pattern withing the bat.
3. Do we have a flag, channel or wedge within BC leg? If yes, we have a Flag (F) bat. If not, we have a Non-flag (N) bat.
4. Does A and C points correlation form a Single top/bottom (S) or a Double top/bottom (D)
5. Is CD leg a Perfect (staight, P) move or is it Imperfect (several impulses leg, I)
Now you can guess that bull HSFSI bat type means a bullish bat pattern which has a several impulses XA leg, channel in BC leg, A and C don’t form a double top, and CD is several impusle move also.
Applying these 5 criteria I detected and described 32 Bat pattern types. To simplify the matter I will only speak about BEAR types from now on. So we'll consider only 16.
On the chart you see two types of BATS. Picture 1 is SNSP type. And picture 2 is HSNSP. They differ only by the type of XA leg. Now I will describe them in details.
SNSP type is a so-called perfect Bat. All its legs are straight. Pattern traders like this type because it looks like in a manual or a textbook. Traders think that if the patterns look perfect it should work perfect. But it's not true. Let’s look at some features of SNSP type:
1. It’s frequency of appearance is 26/200. It means that among 200 bats you’ll have 26 SNSP. It’s not very often. For example, HSNSP appears twice more often.
2. On eurusd+gbpusd+audusd+eurgbp portfolio it has 46,15% profitability. It means that among 200 trades you’ll have about 92 wins and about 108 losses. It’s ok for bats because of risk to reward. Trading only this type you will still have a slightly but steadily rising equity curve.
Now HSNSP type. It’s not as perfect as SNSP type because it’s XA is not a straight move but still it looks good also. Here’re its features:
1. Frequency of appearance is 41/200. It’s one the most common patterns (second place, silver medal))). As you see it appears almost twice more often then SNSP type.
2. On the portfolio note above it has 41,46% profitability. You’ll have about 117 losers among 200 trades. This pattern is not ok. Usual bat pattern risk to reward makes equity curve of trading HSNSP fluctuate about breakeven.
Well, as you see, SNSP and HSNSP types work not very well.But they are frequent enough. I just deleted them from my trading plan. The other way to deal with these patterns is to lower your position size when trade them.
I strongly recommend you to do a backtesting of these types by yourself before you make some conclusions.
Well, it’s all for today. Hope to see you next Wednesday at my next edition devoted to other types of BAT. If you find the information valuable – click like. If you don’t wont’ to miss my posts – subscribe. If you have some suggestion, queations or whatever – leave a comment.