[FAP Turbo Guide - getting the most of automated forex trading] Time to start...
1 view
Skip to first unread message
FAP Turbo Guide Team
May 14, 2009, 5:30:28 AM5/14/09
to fapturboguid...@googlegroups.com
In the April performance review we found out that GBPCHF was actually the second most profitable pair in total generating over 600 pips on 5 demo accounts (it was not traded on the live accounts). The strong results was a bit unexpected to us. This is a perfect example of why you have to break down the data in molecules to see what is really going on rather than just looking on an account statement try to work out in your head what is going on in the account. We highly recommend that you paste your trading data into Excel now and then and perform an analysis similar to the ones we post on this blog. Using the excellent function Pivot Tables in Excel will allow you to stratify the data. Stratification is a great way to find the best ways to improve EA performance. The most important ways to stratify the data in EA trading according to our opinion is:
In addition to this, we believe is a good idea to filter out the losses only and perform a loss analysis to see where the most losses occur. If you are not as excited about stratified sampling and actually find excel crunching quite boring, but still want to make good money in auto trading, sign up to our newsletter and we will give our analyses away for free on regular basis! So, what is the best settings for GBPCHF? To find out, we made a backtest during the period 2008.08.01 to 2009.05.30. We used default settings but extended the trading period from 19 to 06. In the chart below each bar is the sum of losses and wins for each day and open hour.
We used normalized profit as performance measure. The profitability per weekday and per open hour is outlined in the chart below. The most profitable period seem to be around 21-0. Around 04 also seem to be good for some reason. Overall this pair seem to be able to make a good positive contribution on a trading account if traded during the right hours. The two charts below shows wins only and losses only. The data tells us that its important not to start too early with GBPCHF since a lot of losses occur if traded early in the session.
The chart below shows the overall profitability per trade hour.
Finally we decided to split the sample data into two sample periods. This another way to stratify the data. This first sample is for all data during 2008 and the second sample is from all data during 2009.
It was a bit disappointing to see that this pair would have been a lot more profitable during the second part of 2008 compared to the first part of 2009. In 2008 it could have been traded pretty much from 21-06 and would have made very healthy profits. In 2009 however it seem more important to pick the right time period. According to our analysis this pair is best traded during 21-0. Backtesting tells us that it would have been profitable both in 2008 and 2009 when trading it during these timeframes. This fact is further supported by our forward testing in which we have used 21-23. We have also run several backtests trying to find out how the SL level impact profitability. According to our analysis we have found that decreasing SL from 75 to 60 will not increase the number of times SL is hit to a large extent. We therefore believe that SL=60 is a good SL for GBPCHF. This gives a 1:6 ratio between TP and SL. In other words, as long as SL is hit less than about 1 out of 7 trades or so, this pair will be profitable.
Conclusion
GBPCHF was ruled out early in the FAP Turbo community as an unprofitable pair. However, we believe that by limiting the trading hours and a minor adjustment on SL this pair will be profitable. It will not make a huge contribution to the account and it will be a somewhat rocky ride to trade it, but nevertheless profitable. We think it will at least double your account every year. Conservative traders should probably avoid this pair. In the long run however, this pair should make a positive contribution. At the moment it seems like it actually is more profitable than EURGBP which is traded by most FAP Turbo users. From now on, we will trade this pair on live accounts between 21-24 and SL will be set to 60, LRR=10, max spread =8. The problem with this pair is that spreads are usually quite high. It might not be suitable (or possible) to run this one on Alpari Uk. Its better to run it on a good ECN broker where spreads can become really good now and then. Anyway, running a backtest (based on their daytime spread of 7 pips) with the selected parameters on Alpari Uk gives us the following result shown in the chart below. SL was hit 7 times in this test and 152 number of trades was made. This gives us an SL hit ratio of about 1:21 which well below the needed ratio of 1:7. Drawdown in this test is excellent too, only 6,63%.
A final note on statistical analysis
It is very important not to jump into conclusions to early when analyzing data. It is very important to distinguish between what is random and what is systematic. What we want to do is to filter out the randomness in our analysis and try to find out what is systematic and adjust our trading strategy accordingly. For example, the loss only chart indicates that 8 pm on Wednesdays is where the most losses occur on GBPCHF. That Wednesdays during this time is a lot less profitable than Tuesdays and Thursdays is most likely only due to randomness, and one should not draw any conclusions from this. Distinguishing between randomness and what is systematic is extra challenging in trading environments compared to analyzing a process in nature or a relatively static business process. This is because trading environments are so dynamic. However, the better we are in managing this challenge, the more money we will make!
Good luck in your trading, and please feel free to challenge our analysis! Maybe there is something that we did not think of that could further improve profitability, to the benefit for all of us.
For more information on FAP Turbo, please check the official homepage: http://www.fapturbo.com
FAP Turbo Guide Team
--
Upplagd av FAP Turbo Guide Team på FAP Turbo Guide - getting the most of automated forex trading den 5/14/2009 01:36:00 fm
- Open trade hour (during what hours is the EA most profitable?)
- Week day (what days is the EA most profitable)
- Currency pair (how is each individual pair doing?)
- Week (is there any trends in profitability over time?)
- Broker (what broker is most profitable?)
In addition to this, we believe is a good idea to filter out the losses only and perform a loss analysis to see where the most losses occur. If you are not as excited about stratified sampling and actually find excel crunching quite boring, but still want to make good money in auto trading, sign up to our newsletter and we will give our analyses away for free on regular basis! So, what is the best settings for GBPCHF? To find out, we made a backtest during the period 2008.08.01 to 2009.05.30. We used default settings but extended the trading period from 19 to 06. In the chart below each bar is the sum of losses and wins for each day and open hour.
We used normalized profit as performance measure. The profitability per weekday and per open hour is outlined in the chart below. The most profitable period seem to be around 21-0. Around 04 also seem to be good for some reason. Overall this pair seem to be able to make a good positive contribution on a trading account if traded during the right hours. The two charts below shows wins only and losses only. The data tells us that its important not to start too early with GBPCHF since a lot of losses occur if traded early in the session.
The chart below shows the overall profitability per trade hour.
Finally we decided to split the sample data into two sample periods. This another way to stratify the data. This first sample is for all data during 2008 and the second sample is from all data during 2009.
It was a bit disappointing to see that this pair would have been a lot more profitable during the second part of 2008 compared to the first part of 2009. In 2008 it could have been traded pretty much from 21-06 and would have made very healthy profits. In 2009 however it seem more important to pick the right time period. According to our analysis this pair is best traded during 21-0. Backtesting tells us that it would have been profitable both in 2008 and 2009 when trading it during these timeframes. This fact is further supported by our forward testing in which we have used 21-23. We have also run several backtests trying to find out how the SL level impact profitability. According to our analysis we have found that decreasing SL from 75 to 60 will not increase the number of times SL is hit to a large extent. We therefore believe that SL=60 is a good SL for GBPCHF. This gives a 1:6 ratio between TP and SL. In other words, as long as SL is hit less than about 1 out of 7 trades or so, this pair will be profitable.Conclusion
GBPCHF was ruled out early in the FAP Turbo community as an unprofitable pair. However, we believe that by limiting the trading hours and a minor adjustment on SL this pair will be profitable. It will not make a huge contribution to the account and it will be a somewhat rocky ride to trade it, but nevertheless profitable. We think it will at least double your account every year. Conservative traders should probably avoid this pair. In the long run however, this pair should make a positive contribution. At the moment it seems like it actually is more profitable than EURGBP which is traded by most FAP Turbo users. From now on, we will trade this pair on live accounts between 21-24 and SL will be set to 60, LRR=10, max spread =8. The problem with this pair is that spreads are usually quite high. It might not be suitable (or possible) to run this one on Alpari Uk. Its better to run it on a good ECN broker where spreads can become really good now and then. Anyway, running a backtest (based on their daytime spread of 7 pips) with the selected parameters on Alpari Uk gives us the following result shown in the chart below. SL was hit 7 times in this test and 152 number of trades was made. This gives us an SL hit ratio of about 1:21 which well below the needed ratio of 1:7. Drawdown in this test is excellent too, only 6,63%.
A final note on statistical analysis
It is very important not to jump into conclusions to early when analyzing data. It is very important to distinguish between what is random and what is systematic. What we want to do is to filter out the randomness in our analysis and try to find out what is systematic and adjust our trading strategy accordingly. For example, the loss only chart indicates that 8 pm on Wednesdays is where the most losses occur on GBPCHF. That Wednesdays during this time is a lot less profitable than Tuesdays and Thursdays is most likely only due to randomness, and one should not draw any conclusions from this. Distinguishing between randomness and what is systematic is extra challenging in trading environments compared to analyzing a process in nature or a relatively static business process. This is because trading environments are so dynamic. However, the better we are in managing this challenge, the more money we will make!
Good luck in your trading, and please feel free to challenge our analysis! Maybe there is something that we did not think of that could further improve profitability, to the benefit for all of us.
For more information on FAP Turbo, please check the official homepage: http://www.fapturbo.com
FAP Turbo Guide Team
--
Upplagd av FAP Turbo Guide Team på FAP Turbo Guide - getting the most of automated forex trading den 5/14/2009 01:36:00 fm
Reply all
Reply to author
Forward
0 new messages