Dear all,
Scope and purpose of this thread is to discuss, backtest and implement pivot based entry-exit-money management and position sizing systems.
To start with, I have coded a system in Amibroker for entry-exit and money management by drawing inspiration from Saint and Karthikmarar.
Notes about this system :
1. It works for intraday on all time-frames, not suitable / tested for EOD data. Backtest results show it working very well for particular time-frames than other time-frames.
2. Users will have to enter / adjust parameters like total capital, maximum loss per trade as % of capital, room for stoploss as % of close below pivot and sensitivity of pivot.
3. If the expected pivots are not seen, reduce the parameter sensitivity of pivot.
4. The mechanism of finding pivots is based on zig() function and shows values little different than pivots.
5. Designed for continuous trading. i.e. short=sell and cover=buy
6. When long, watch crack of close in green line ONLY. Ignore red line.
7. When short, watch crack of close in red line ONLY. Ignore green line.
8. You will have to tweak the parameters as per your trading style.
9. Since stoploss lines are plotted continuously, they sometimes move down (when long) and up when Short. This is against stoploss principle that it should move in one direction only. This has rarely affected the results (surprisingly), workaround will be implemented soon.
10. This method sometimes shows unreliable entry-exit on gap up / down opening.
Snapshot , code and back-test results, backtester settings snapshot are attached.
Backtester settings : All backtests are done using
capital - 100000/-,
sensitivity for pivots - 0.4,
max. loss per trade as % of capital 1%
room for stoploss as % of close below pivot 1%
Important regarding pivot definition : Since this method is based on pivots, formation / definition of pivot, which is controlled by sensitivity of pivots, plays an important role here. By keeping rest of the settings same, even if we decrease sensitivity for pivots to 0.1 from 0.4, the net profit in backtest zooms from 4% to 393%. Both these examples are also attached for reference just to underline how important and critical is this factor. This alone could be one reason why when everyone follows Saints pivot method, their bank balance is different from eachother.
Imp. regarding backtesting : Although our system indicates permissible quantity to buy / short, backtester ignores this and buys / shorts max. quantity using available funds. Hence although we are ensuring max. loss per trade is not more than 1%, backtester shows more loss than expected. Once this system is finetuned, I will write custom backtester to incorporate our settings.
Comments most welcome.
Best regards,
Abhay
Scope and purpose of this thread is to discuss, backtest and implement pivot based entry-exit-money management and position sizing systems.
To start with, I have coded a system in Amibroker for entry-exit and money management by drawing inspiration from Saint and Karthikmarar.
Notes about this system :
1. It works for intraday on all time-frames, not suitable / tested for EOD data. Backtest results show it working very well for particular time-frames than other time-frames.
2. Users will have to enter / adjust parameters like total capital, maximum loss per trade as % of capital, room for stoploss as % of close below pivot and sensitivity of pivot.
3. If the expected pivots are not seen, reduce the parameter sensitivity of pivot.
4. The mechanism of finding pivots is based on zig() function and shows values little different than pivots.
5. Designed for continuous trading. i.e. short=sell and cover=buy
6. When long, watch crack of close in green line ONLY. Ignore red line.
7. When short, watch crack of close in red line ONLY. Ignore green line.
8. You will have to tweak the parameters as per your trading style.
9. Since stoploss lines are plotted continuously, they sometimes move down (when long) and up when Short. This is against stoploss principle that it should move in one direction only. This has rarely affected the results (surprisingly), workaround will be implemented soon.
10. This method sometimes shows unreliable entry-exit on gap up / down opening.
Snapshot , code and back-test results, backtester settings snapshot are attached.
Backtester settings : All backtests are done using
capital - 100000/-,
sensitivity for pivots - 0.4,
max. loss per trade as % of capital 1%
room for stoploss as % of close below pivot 1%
Important regarding pivot definition : Since this method is based on pivots, formation / definition of pivot, which is controlled by sensitivity of pivots, plays an important role here. By keeping rest of the settings same, even if we decrease sensitivity for pivots to 0.1 from 0.4, the net profit in backtest zooms from 4% to 393%. Both these examples are also attached for reference just to underline how important and critical is this factor. This alone could be one reason why when everyone follows Saints pivot method, their bank balance is different from eachother.
Imp. regarding backtesting : Although our system indicates permissible quantity to buy / short, backtester ignores this and buys / shorts max. quantity using available funds. Hence although we are ensuring max. loss per trade is not more than 1%, backtester shows more loss than expected. Once this system is finetuned, I will write custom backtester to incorporate our settings.
Comments most welcome.
Best regards,
Abhay
