Backtest: Renko reversal
This is the backtest of a Renko based chart. The system will enter on first green bar, after a series of red bars.
First please read my post on How to properly backtest Renko charts.
System Info
ES Renko 4 ticks chart.
DataFeed: IQFeed
Software: Investor/RT(affiliate link)
Contract traded: 1
Direction: Long Only
Backtest period: [2021–05–06 18:00–2021–11–02 17:00] — Almost 6 months.
System hypothesis and entry signal
After a series of red down bars, we come into oversold territory and wait for a green reversal bar to enter.
Entry condition: At least 6 down bars, then a green bar.
Exits
Stop loss: Close of a red bar below the open of entry green bar: exit on close.
Profit: Price moved in our direction of 2 green bar: exit on close.
I like these equidistant stop/profit: if the system has an edge, it should hit the target more often.
Filter
Time: Only enter between 0800 and 1700.
Sample trades
Results
Gross Profit/Loss: -8750.00 $
Number of Trades: 819
% Profitable: 44%
Average profit per trade: -10.68 $
Terrible result. Of course, this system will need refinement. When system results are so bad we can say we stumbled on serendipity: Reversing the rule and entering short instead of long would give a positive outcome, and quite good! I might create another post with the exact same system but short.
So why is the initial hypothesis so bad? As you can see on the image, a drop of 6 consecutive bars doesn’t place us in an oversold context…. but in a downtrend!! The system in fact tries to buy small pullbacks in a downtrend.
Filter variation
What if we require the reversal bar to have increased volume?
Time: Only enter between 0800 and 1700.
Volume of current bar >= Average volume of previous 3 bars * 4
Sample trade
Results
Gross Profit/Loss: -1350.00
Number of Trades: 62
% Profitable: 21.77%
Average profit per trade: -10.68 $
You can see on the chart I added a white line over the volume bars: this is the 3 bars volume average. This trade is a losing trade. After a severe down move, a huge increase of volume on the reversal bar is not even capable of making a 2 bar move… it’s the same problem as with the above test: the system tries to play a reversal, but more often than not, a big down move calls for more down movement. The equity curve is not shown here, but it is terrible.
Also, the filter is too restrictive: only 62 trades in 6 months this is not acceptable. I tried lesser volume factor (*2, *3) but the system results are still unacceptable.
Filter variation
What if the reversal happens specifically at the low of the day?
Time: Only enter between 0800 and 1700.
Low of the previous bar (red bar before green reversal bar) is low of the day.
Sample trade
Results
Gross Profit/Loss: -3225
Number of Trades: 199
% Profitable: 41%
Average profit per trade:-16.21 $
The blue line on the chart represents the low of the full session (1800->1700). Unfortunately, entering on a reversal bar, after a big down move, at the low of day doesn’t improve the system.
At this point, I could keep refining the filtering criteria, or change the exit conditions. But I will stop for now. This system was looking for a precise, immediate reversal to catch profit and get out quickly. The hypothesis and the context chosen for exits proved to be wrong.
Questions? Backtesting ideas and variations? Ask below in the comments.
