[Note – this post should be especially beneficial to anyone currently developing a trading strategy and/or testing one]
All right, I am still here – cracking on with trading and testing.
A word on the market conditions
The market continues to be rather tight – markets are ranging – FX, indices – other than the GBP crosses which are moving up and down depending on Brexit polls and Brexit poll rumours. Gold and Oil are trending a little but not a lot. I am also keeping a close eye on the Brexit odds on the Betfair exchange. The odds have varied from 1.2 to 1.75 in the past three weeks (meaning that a $1 bet makes a profit of $0.20 and $0.75 respectively). Liquidity and betting volumes on that bet are picking up. Around £5 million of bets were placed in the past 48 hours alone, with total bets now at around £30 million.
Where am I at?
In my last blog post I conveyed some of the frustrations I have been experiencing. Of course, I have decided to redo the 2015 testing to see whether I can get some more meaningful figures. Additionally, doing this additional testing in ForexTester2 (as opposed to spotting the setups on MT4) will provide with trade entry and exit details which will help me to judge trade durations, breakdown by session (London v New York) and clustering of winning and losing trades. It will also generate a lot of time saving for the optimization stage.
Had a good catch up with TraderDi last night (that’s what I will refer to her as from now) who gave me some more pointers and suggestions. We also talked about the some of the psychological aspects and the importance of being strongly committed in order to succeed as a trader. She continues to have a lot of faith and belief in me as a trader. [TraderDi got in contact with me as a result of this blog and because of my article on Trade2Win]
In the last couple of days I already completed a full review in MT4 of the months of March, April and May of 2016 to spot all valid setups, entries, stops and targets. Compared to my first run of that data I spotted 140 trades, with a net return of 26R (a 19.7% return on risk). This compares to 174/43R/24.8% for the first run through. The second time I was more conservative, regularly entered an hour (one candle) later or not at all – thus reducing the number of trades.
In a posts a few weeks ago I pointed out some potential weaknesses associated with my testing – in particular there were 5 points – here I am detailing how I will try to combat each weakness:
Problem#1: Assuming that every single session is traded. Answer: I know that I won’t be trading every single session. I will simply rely on the figure of average expectancy per session, and expect that my total live trades will be less in number than the total test trades.
Problem#2: Slippage on news. Answer: Will close entire positions prior to really big risk events. Will close trades prior to news events occurring during the Asian session.
Problem#3: Other transaction costs such as rollover and widening spreads during rollover. Answer: All slightly more exotic positions (including all NZD and CAD crosses) will be closed prior to daily rollover. Rollover costs will still be ignored as these are minimal.
Problem#4: “Hindsight-Benefit”. Answer: This has been the single biggest issue and is the main reason for undermining my confidence in the testing statistics collected thus far. I will no longer allow this to take place in the final part of this testing.
Problem#4 (part 2): An associated problem here has been lack of certainty as to lining up the historical price data to the news releases – which has contributed to the hindsight-benefit issue. Answer: I will use 1 instrument (outside of FT2) as control instrument to ensure that I have the time correct for each trading session – there will be DST, non-DST and two 2-week phases per year where US is on DST and Europe is not. The control instrument will have historical data downloaded into Excel, which I can compare with the chart on FXCM. This will assist in both setup selection as well as position management and improve the accuracy of the testing.
Problem#5: Psychology issues not present in testing. Answer: None.
Problem#6: Risk/Money Management for multiple positions. Answer: Given that I am only ever risking 1% per trade, that I am continually recalculating the 1% depending on the most recent trading capital account, and that there is a positive correlation between number of simultaneously open positions and profit/loss -I will simply take as many setups as I can see on the charts, and risk the full 1% on each trade.
Group Approach for testing: Rather than testing one instrument at a time, i will do 2 groups of 5 instruments each (5 USD rates in one group, and GBPCHF/EURJPY/EURCHF/EURGBP/USOIL in the other). This will also be more similar to how I make my decisions in live trading conditions.
Multi-Chart Viewing: I will play around with the layout in FT2 to see whether I can see all 5 charts at once – this would speed up the testing time
Hopes for Trading Frequency & Expected Profitability: I am hoping that the 10 instruments will give me 500 trades per 12 month period (the test period is for 17 months) – I am hoping for an EV of 0.15/trade multiplying that out – 500 for 10 instruments will translate into 1,000 trades for 32 instruments, and up to 1,500 for 50 instruments (the ten instruments used are more liquid instruments and thus are more likely to have setups)
If you have read this blog post in its entirety up to this point, then congratulations – you are doing well! Chances are that you are actually seriously interested in the nitty & gritty of the testing and strategy development process. Hopefully you are realising how much detail, analysis, planning and work is required in this process. It’s a lot like work!
If you are interested in other blog posts related to testing, then check out these: