Live Trading, running an algo, reconciling the two, more coding, building a market context model. Long hours. All that was part of my week. Read further for a detailed discussion of these.
Am sharing some of these details because firstly it helps me to clear things up in my head a little, and secondly, I have found more and more regularly that blog readers are contacting me to share that they have come across similar issues and it can often lead to the start of a new alliance or working relationship.
Live Trading – rule-based manual execution
Traded the pin-bar based trading strategy that I am in the process of building into an automated strategy. Have been doing that for the past 3 weeks, taking a total of 35 setups as well as a few unfilled orders. Looking for setups on the 4-hour and daily timeframes, so it is not taking up a lot of time each day. Have settled into a good routine for doing that.
Running Algo live in demo
Am also running the automated strategy in demo version on the TradeStation platform to check for false positives (trades entered by the strategy that I would not want) and false negatives (where the strategy fails to trade when I would like it to).
Created Trend Filter & Market Context Model
Thanks to some code that a trading colleague shared with me, I was able to effectively finally build a market context model that is largely based on ‘market structure’ (see Michael Voigt [in German] and Pepperstone [in English]). It boils down to comparing high and low points in a market to judge on whether a market is trending up, trending down or consolidating. This took just several hours across 2-3 days – very happy about that.
My colleague also gave me some guidance on how I should structure my code more effectively (now that my code bits routinely already reach 200-300 lines) and how to use ‘functions’ and ‘projects’ more effectively. Thanks a bunch!
Incredibly valuable Feedback Loops
Comparing and reconciling the manual trades and the trades taken by the strategy is giving me a feedback loop for making further adjustments to the code. Additionally the automated strategy is alerting me to potential setups thus helping me in checking for setups (across near 40 markets), similar to a trade signal generator.
Likewise, in running the market context model on TradeStation, I can look for situations where the model is giving me a trend rating that I don’t agree with, to then allow me to refine the code (for example it gave a sideways market condition for the S&P for much of November, which I was not happy with). Again, it’s a feedback loop. Again, it also helps me to more quickly judge the trend condition of a market by looking at the custom indicator that I created, and which I have integrated into the buy/sell decision-making progress of the trading strategy.
Insights from the Feedback Loop
Examples that came out of the feedback where being more precise in setting the parameters used to judge the pinbars in terms of their size (using a transaction cost multiple as well as ATR-multiple) and their shape by analyzing their wicks and bodies. Further, I studied the price action preceding the pin-bar trigger candles to then articulate my reading of the price action into code as much as possible.
Becoming more proficient at extracting raw data from the historical prices has been helpful in data analysis and building tweaks. Obviously it helps if one likes working with numbers, statistics and probabilities.
‘Psychology’ and discipline much less of an issue
My mental game has become a lot more rule-based and process-driven because I constantly have to force myself to articulate my thought processes into code. I have to be ultra specific in what I am looking for. Thus, when then looking at the charts in real-time, the decision as to whether to trade or not, whether to fluff around with an open position and so on, become incredibly easier.
A lot of the ‘psychology’ just disappears – incidentally, this is part of the message I was trying to convey in a webinar I gave recently on how to overcome fear and greed in trading.
It’s become fairly easy to build some processes and routines around the live trading – something that has been drummed into my head for a long time by fxtrader_Paul, one of my trading coaches.
Strategy Development Status
Even with all this work, the strategy is still producing many false negatives, i.e. missing trades. Maybe I will just have to accept that. The upside is that I can let the strategy run day-in day-out 24 hours a day (whilst I relax in my hammock on a tropical beach).
Anyways, that’s enough bantering from me – plus my beautiful girl-friend is waiting for me to get home for a nice evening in!
Finally, as a quick bonus, here’s a link to two excellent in-depth articles that I came across recently, written by independent trader Felix based in Japan:
- From SmartforexLearning: Trading strategy based around pin-bars and momentum
- From SmartforexLearning: Discussion on FOMO (“Fear of Missing Out”) and how to tackle it