The week has been productive. Traded Oil futures for a couple of hours each day as part of a trading room. Completed a significant amount of back-testing.
Balanced it out with regular exercise and lots of fun away from the office (including competing in a 6-hour inline skating race and chasing down street art in East London).
Trend-Following Strategy Back-Testing
Finished the short and long versions of the trend-following model and have run it over several instruments for a range of time periods from 2004 to 2016.
Thus far, this produced 2,800 trades. The testing shows that the strategy has no edge, with the overall profit/loss being roughly equal to the transaction costs. The good point here is that, with these size of sample, I can be confident that the result is somewhat statistically significant. And of course that the defined setup has been executed consistently without vagueness or discretion.
Am aiming to run the model through the futures data for several instruments today – ES (the S&P 500 index), CL (Crude Oil), Copper, Gold and TY (the US 10-Treasury Bonds). Following that, the testing sample will touch not only FX, but also indices, commodities and fixed income. The assumption is of course that the strategy can be applied uniformly to a wide range of instruments.
It’s been very beneficial to discuss some of the intricacies and details with the traders at the hedge fund here. The guys have extensive experience as far as testing and some of the relevant maths are concerned.
Following that I will:
- check how some of the trades look on the chart (as opposed to being pure numbers in a spreadsheet).
- reconsider how I have defined the trend – and thus how the bullish, bearish or neutral bias is determined
- reconsider how I define the pullback
- reconsider how I assess whether the pullback has likely come to an end
- [the Oil Trading Room experience is very helpful in this regard]
Yes, this will take a long time – however I feel I am doing it properly and professionally, and that it will actually to meaningful and useful conclusions. Yes, I am aware of over-fitting risks associated testing – but thanks for reminding me!
In recent months, I have been harping on a lot about testing – here are links to some of those blog posts:
- Back-testing and trading confidence
- Back-testing trading strategies
- Searching for Alpha in my testing…
Oil Futures Trading Room
4 trading sessions into the week, it’s been a profitable thus far. All in all, have managed to accumulate around 300 ticks in profit since starting in the room several weeks ago. Let’s hope it continues like that. Have upped the position size to 4 CL contracts per trade since the beginning of the week (with approx US$1,200 risk per trade). Reduced position size by 50% for Thursday’s session because of the OPEC meeting in Vienna – that was good risk management. As expected, volatility was very high for the session with price taking a real dump in the New York session.
As opposed to most other members of the trading room, I am literally trying to copy the lead trader’s trades, without taking any trades based on my own analysis.
It is very interesting to observe how the other members trade, and the type of emotions they display depending on the outcome of the trades, and additionally what ideas they pursue for their trading setups.
My final thought on the trading room is that I should start to document all the trades that I am taking, including the reasoning and analysis shared by the lead trader – this should be a fantastic method for accelerating my learning since the lead traders goes to great lengths explaining his thoughts for taking a trade. Additionally I need to collect more trading statistics on the executed trades. Until now, I have not bothered with documentation on the basis that I was copying somebody else’s trades.