How it works
On the following page we'll explore how our model and our rules actually work, giving examples that show examples of trades in action and explain the logic behind them.
We use code to generate 285 different buy and sell signal on 95 assets across industries, sectors, and asset classes. We trade 4 different sets of rules, which adds extra diversification. Two systems are based on continuous volatility adjusted pricing lines, one is based on the 40 day price breakouts (new high or new low), and a 120 day breakout.
Average True Range
System 1 which is the original system we built back in 2019, which is an ATR (Average True Range) based long trend following system. It plots a continuous price line that follows the asset price (depending on the position) and generates buy signals after moving a certain distance from a previous low, and closes buy signals when it moves a defined distance from a previous closing high.
System 1 is long only, and mainly focused on individual equities, crypto, and metals. Below are a few examples of this ATR system in good and bad (performance) periods, as well as through heightened volatility periods (like March '20).
What makes this system unique, apart from being long-only, is that each signal is fitted to be the most profitable ATR line for that asset historically, which means of the four systems this is the only without fully universal rules.
Apple was trending nicely in to the Covid crisis in 2020, our system said that trend was ending / slowing down (at least temporarily) on the 25th of Feb. We closed our trade, took a decent profit and and avoided a 35% decline during the pandemic market crash.
When conditions improved, the system told us to go long again on the 15th of April, and went on to capture a 71% move, before pulling back in September 2020.
This is a good example of our system working at it's best.
After these trends, we entered a rockier period with a series of small trends and false breakouts, (below), that roughly evened each other out in the P/L. The system is designed to take controlled losses, capping each failed breakout at 0.2% of the portfolio, but letting winners stay open to run up profit, and pay for all the losers and make a profit.
Long and Short Average True Range
System 1.2 works in a similar way to 1, but assets go long and short. Otherwise the same rules and logic applies. Whereas System 1 trades equities, indicies, crypto, and metals, System 1.2 trades currencies, commodities, and bonds.
Below shows Corn with its ATR line signal. Commodities have been back in the headlines in 2022 with covid-induced supply constraints driving up prices. Our system caught those trades, which paid for a decade (or more in some cases like Coffee) of small losses.
40 Day Breakout
System 2 works completely different to 1 and 4. It captures the 40 day breakout and closes on an identical 40 day closing low. Very simply, for long positions it opens a trade when the stock/asset closes the previous day above the previous 40 days highest close. Likewise, it opens a short if the asset closes below the previous 40 day closing low.
These parameters mirror original trend following strategies from the 80s that used a 40 day open, 20 day closing low, or a 20/10 rule. I've back tested this strategy, and whilst it isn't as profitable as longer term time frames (which are more commonly used these days at firms like Chesapeake), such as 200/250 day close, it closes the trade quicker than those systems and gives back less profit on trend reversals, which adds important diversification to our model when accompanied by longer term models in aggregate.
System 2 trades equities, bonds, crypto, metals, commodities, currencies, on the long side, and commodities and currencies on the short side.
Below shows a 40 day breakout in bitcoin on the 14th of October 2020 at 11428 USD. It closed 6 months later in late April at 51,187, capturing an enormous trade and protecting from the sell off back to 30,000 USD. On Bitcoin, the 40/40 breakout (long-only) has a 73.33% win rate, and a 13x profit to loss ratio!
120 Day Breakout
System 3 follows the same logic as system 2, but with 120/90 parameters. I.e. we buy or short on the 120 day closing high or low, and close the trade on the subsequent 90 day reversal.
Whilst system 2 doesn't short equities or metals, system 3 does. It's the only system that shorts stocks, because at a 120 day low historically, stocks haven't been in good shape and the risk/reward ratio is skewed more positively for this system than the others.
Below you can see system 3's parameters applied to emerging market bonds, which took a short position in October long before current market and global instability.
Fixed Average True Range
Lastly, Systems 4 and 4.2 (short version) are the same as System 1 but with fixed rules universally applied across all 95 assets. My personal favourite, and a great medium term trend following model that has performed consistently over many years.
Below you can see a long/short fixed-atr signal applied to Crude Oil futures. You can see it's very similar to 1 and 1.2.