Automated Trading Strategies: Q2 2023 Portfolio Update

13 min readJul 19, 2023

In case you didn’t know, we’re on the hunt for the holy grail of automated trade strategy. I define the holy grail of trade strategy as having the following attributes:

Profit factor (gross profit/gross loss) greater than 3
Annual max drawdown less than 3%
Annual return greater than 500%
Maximum daily low of -$1,000
Avg Daily profit greater than $1,000
Less than 5,000 trades annually
Greater than 253 trades annually

As leader of the hunt, every few months I like to climb the highest tree in search of an eagle’s view. In particular, I like to review:

where we came from,
where we are now; and,
where we’re going from here.

We’re making a big change this quarter by shifting the focus to strategy variations with a higher profit factor, even if the trade count is low. I’ll tell you what information in the forward and backtest led to that decision. I’ll also look at how some of our top strategies are performing and update you on key findings, highlights, takeaways and what’s in the pipeline for Q3.

Important: There is no guarantee that these strategies will have the same performance in the future. I use backtests to compare historical strategy performance. Backtests are based on historical data, not real-time data so the results we share are hypothetical, not real. There are no guarantees that this performance will continue in the future. Trading futures is extremely risky. If you trade futures live, be prepared to lose your entire account. I recommend using these strategies in simulated trading until you/we find the holy grail of trade strategy.

Q2 In Review:

The goal of this review is to update you on any progress made since the last update. You can view past updates below:

When we first started, I published 10 strategies that barely had a profit factor (gross profit/gross loss) greater than 1.05. Today, we have 70+ strategies, and I only document those strategy variations with a minimum profit factor of 1.50.

When we first started, in a kind of mental masturbation, we optimized every parameter; today, after being impaled by a few over-fitted unicorns, we only optimize the time series, including day of week.

When we first started, I thought the simulation had no limitations. Today, I know that using the simulation as a tool to find viable strategies must carry with it an understanding that accuracy holds an inverse relationship with complexity.

When we first started this, I didn’t think it was possible to find the holy grail. Now, I have a kind of blind faith that it is. Like the alchemists and technicians of history, it has become an obsession, just shy of a religion.

Where do we go from here…

Embrace The Challenges

First, I want to apologize for the delay in getting this update out. My computer got hacked again and I can’t write when I’m agitated. It happens once a month. This time it did more damage than usual, but it’s something that happens with such regularity that I’ve learned to be an expert at redundant systems.

Second, while waiting for my system to be fixed/updated I decided to go on a digital quest for my next favorite TV show. I found a show called Lodge 49. It’s another “hero’s journey’“ about a young man that finds a home in a fraternal order dedicated to finding the holy grail for alchemists: the formula for gold. This is a hunt that seems to have transfixed humans across time and space as well.

I couldn’t help but liken the alchemaic hunt for gold to our hunt. In particular the part about all the challenges I’ve been facing. A hunt like this is characterized by challenge, but I had no idea it would come from multiple angles. Many alchemists (Knights Templar) were persecuted by the Catholic church as heretics so at least it hasn’t gotten that bad, but this brings me to my third point.

Third, aside from the challenge of technology, and the sheer difficulty of the task itself (finding the holy grail of automated trade strategy), my work has been heavily scrutinized as of late. It is too easy to some and too complex to others. I’ve been admonished for being ridiculous, ignorant and naive in this pursuit. One of my best friend’s said I was on a “fool’s errand”.

The hardest part about digesting these comments is that they might be true. I can’t be serious about attainment of this goal and not be open to criticism. I mean here we are, over two years in, and no holy grail.

But here’s the thing, if you’re open, criticism is like water to a seed.

When I look at the strategies we published in January 2021 compared to where we are now, I can honestly say that while we may not have reached the holy grail quite yet, we are so much closer to it. I also hold dear the emails/comments I get from those of you that have benefited from this newsletter — I can’t tell you how much they inspire this effort. While the primary goal was to find the holy grail of automated trading strategies, the secondary goal was to introduce others to the world of trading and automated trading in particular.

Trading is not something that’s taught in schools, it’s passed down, and is generally relegated to witchcraft by fundamental analysts, but every bank has a trading floor and it’s a skill-set that has been around — in documented form — for hundreds of years. I recently published Strategy 69 which walks through some of the trading systems developed throughout history. I also had the pleasure of reading a great article about trading legend and professional dancer Nicolas Darvas. In particular, all the hurdles he faced in his holy grail search. From Jesse Livermore to Darvas, every professional trader has a story to tell. It generally goes like this: lose money, break-even, make money, lose money, break-even, make more money, lose money, break-even, make even more money. It’s a process that has broken many.

At the moment, ATS is composed of many different types:

Some have never traded before, but have a programming background. Their goal is to use these strategies to learn about the art of trading.
Some subscribers are traders, but don’t know any programming. Their goal is to use these strategies to learn how to automate their own trading system.
Some subscribers are traders with a programming background and/or experience with NT8, but they’re stuck where I was some two and half years ago. Their goal is to learn which techniques we use to raise our average backtested portfolio profit factor from 1.05 to 1.97.

I hope everyone’s goal is to use this newsletter and its 70+ strategies as a way to leapfrog over the mistakes we’ve made. I can’t think of a better way to make up for all the challenges I’ve personally encountered. Finding the holy grail is one thing, but helping others to do the same is truly golden.

Q2 Performance Chart: Where Are We Today?

The current backtested portfolio made $2 million (down from $5.2 million) over the time period July 2022 to July 2023. In total, the portfolio made:

8k trades (~32 trades per day and ~$257 per trade); compared to,
24k trades (~94 trades per day and ~$219 per trade) in Q1.

What’s happening?

We’re getting smaller and leaner in favor of strategies that only produce 30 to 80 trades per year, if the profit factor is over 1.50. The result is a portfolio with a lower trade count and a 17% higher dollar per trade. Except for Strategies 1 and 5, there are no strategy variations with a profit factor less than 1.50 included in this performance report.

There was a time when I viewed strategy variations that produced under 100 trades as being unreliable — the logic being, the more trades you have, the more you know how a strategy performs. Over time however, I’ve noticed that there’s another side to this coin. It is also possible for low trade strategies to be indicative of higher quality as long as the backtest is accurate.

So what?

So instead of a portfolio with a profit factor of 1.48, we have a portfolio with a profit factor 1.97. Instead of relying on a high trade count to push net income (which generally lowers profit factor), we can use scale. As a rough estimate, if we used 3 contracts to trade the portfolio instead of the 1 being used in the model now, our trade count would go back to what it was in Q1, but the net profit would be ~$7 million rather than $5.2 million, and you maintain a higher profit factor. In addition:

All strategies on the updated Q2 performance chart have a net profit of at least $10K for the year and a profit factor of 1.50 (except for Strategies 1 & 5).
Strategies 44b, 47, 54 and 65 all have a win rate higher than 90%. I will note that Strategy 47 has an unlimited stop loss and no take profit; the win rate is organic. Those of you that subscribe to the newsletter know what that means. In the past, I’d overlooked this strategy due to the low trade count so I’m looking forward to testing it in the forward test.
Strategies 45 and 51 have the highest profit per trade at $3,588 and $2,210, respectively.

For Q3 of 2023, I’ll be adding the following strategies to the forward test: Strategies 7, 30, 36, 37, 39, 40, 41a(f), 51, 59, 65, 69. All have a profit factor over 3. I’ll also be adding 44b, 47, 54, 65, 45 and 51 if they haven’t already been added.

The hard part is developing a framework for selecting which strategy variation to trade on multiple accounts. Even though each strategy has a low trade count, we run the risk of two trades cancelling out because they’re both on the same instrument. I have a few ideas that I’ll share in the Mudder Report.

The Forward Test

If you’ve been following ATS for a while, you know that we’ve had some ups and downs, in particular with regard to backtest accuracy and alpha decay. Since we’ve come up with ways to improve backtest accuracy, we can now turn more attention to alpha decay, which is playing out in our forward test as well.

The purpose of this update is two fold,

it’s a way to document and share our progress over time
it gives us more strategies for the forward test.

The forward test is tracked in the Mudder Report, which is generally published to paid subscribers every two weeks or so. The goal of the Mudder Report is to track our best backtested strategies based on real-time trade data. In other words, the trades are being made on a simulated account, so the results are still hypothetical, but the data driving the simulation is live, which gets us closer to how these strategies will perform once traded on a live account.

This is where we are after 22 weeks:

Let’s walk through the chart data:

The first column is the Strategy, the second column is the approximate start date for each strategy. The earliest date is 1/2/23, the latest is 6/19/23.
Only ~35% of our 70 strategies have been forward tested so far. Among those that have been forward tested, I see room for improvement even in the strategies that are performing well. Over the next year or so I’ll be transitioning from an 80/20 focus on strategy development and forward testing to a 20/80 focus. Which is to say my priority will be transitioning from a focus on strategy development to forward testing. Best case scenario, I’d like to post the results of each strategy in real-time. If you have any ideas on the best way to do this, please let me know.
Moving from left to right on the chart above, Max MAE, ETD and MFE are useful for understanding how the strategy moves in fluid (raw) form as well as the volatility (risk) associated with the strategy.
MFE/MAE gives us the ratio of a favorable move to an unfavorable move. The higher the ratio, the more favorable the average movement is for the strategy. It’s also a way to measure the effectiveness/efficiency of the entry command.
MFE/ETD tells us how much you’re leaving on the table. The smaller the ETD the better, so you want a high MFE/ETD ratio. It also tells us that the strategy’s performance could improve with a better exit strategy.
The 10th column is net profit followed by profit factor. Strategy 44 has the highest net profit at $50K, followed by Strategy 2 — but take a look at Strategy 2’s Max MAE — it’s the highest on the chart at $8,295, which explains why the strategy’s performance is so volatile. Strategy 57 has the highest profit factor, which confirms the power of a high MFE/MAE ratio. It also suggests that Strategy 57 has the potential to command an even higher net profit if we can improve the exit strategy. I think Strategy 26 also has a lot of potential for improvement — it has the third highest profit factor with a relatively low MFE/MAE and MFE/ETD ratio.

We will continue to add\develop strategies throughout the year as discovered through backtests. The goal is to use the backtest to find strategies to test in the forward test and then use the forward test to find a few true gems by year end.

Q3: Putting It All Together — Refining Edge & Next Steps

The scene below is a depiction of Mansa Musa, ruler of the Mali Empire in the 14th century.

https://commons.wikimedia.org/wiki/File:Mansa_Musa.jpg

Musa is shown holding a gold nugget. After becoming a big fan of the TV Show Lodge 49, I couldn’t help but liken the alchemaic hunt for gold to our hunt. I wonder if we can learn anything; is it possible for us to somehow use their hunt to leapfrog over the challenges and obstacles that may come our way?

After some research, it dawned on me that the current hunt for gold really boils down to the refinement process. We know that in the 10th and 11th centuries Europeans purified gold through cupellation, a process in which lead is mixed with gold laced with impurities, and then heated in a furnace until the droplets of purer gold can be skimmed off. We also know that West Africans used glass to refine gold. Gold is inert, and doesn’t fully dissolve into melted glass, while impurities and other materials do.

The question isn’t how to find the gold, but how to get at it. More to the point: what technology can be used to refine or pull the purest gold out of a crude deposit? Likewise, what techniques can we use to refine the edge already discovered with these strategies?

That’s what I want to turn this hunt into — I want to transition from a focus on finding edge to a focus on refining it. Practically speaking, that means going back over what will be 100 strategies by this time next year to see what we’ve learned. For example, in the last Mudder Report I suggested that we might be able to Frankenstein our way to the holy grail by cutting and pasting the entry command for the strategy with the best MFE/MAE stats with the exit command from the strategy with the best MFE/ETD stats. To learn more about MFE, MAE and ETD, click here.

Strategy 57 has the highest MFE/MAE. It also has a high ETD, which means there’s a good opportunity to improve the exit strategy.
Strategy 41a has the second highest MFE/MAE, and the MFE/ETD is relatively high as well.
Strategy 68 has the highest MFE/ETD followed by Strategy 43.

Opportunities abound. If anyone comes up with a good combination, and would like to share, please let me know. Meanwhile, we’re going to use Q3 to:

Continue to explore the high win rate portfolio strategy.
Continue tracking the forward test performance of certain strategies and portfolio theories in real-time.
Continue correlation studies — This is something we started in Q2 with the introduction of Strategy 67. While interesting, the correlation indicator proved tedious. We’re currently looking at other ways create a correlation indicator now (thanks for the tip Kevin).
Continue on AI track — I think we’ve made a great deal of headway here. It’s also an area of focus in the working group. Our next strategy will investigate NT8’s AI tool AI Generate, which uses generative AI over brute force to find optimized indicator combinations. I don’t know what the next step is from here, but I’d like to get to the point where we can ask some AI entity to create its own indicator instead of a combination of man-made calculations. Based on the limitations of ChatGPT, I’m thinking that’s about 2 years off.
ATS Working Group — The purpose of the group is to provide a safe space for ATS traders to share strategy ideas. I’m looking for a mathematician (or someone with a really good understanding of probabilities) and a project manager with 15 to 20 hours per week to dedicate to what may evolve into a business venture. If you’re interested in joining the group, please let me know how you can contribute and/or send a quick overview of your idea. I would also urge anyone with a new strategy idea to post it in the comments section for that strategy. That’s the best way for others to see it. You can also get my attention with a screenshot of the performance results.

In other news, all subscribers as of August 1 will receive 1 month free as I will be traveling.

Upcoming posts:

Strategy 70: pushed to 7/22–7/25
ATS Working Group: We’ve got an interesting post coming up on AI by someone who is very knowledgeable about ML technology.
Contract updates for all strategies: 7/30
Update to election year event study: 9/30; 2024 is an election year which is known to be a bull year for the market, especially stocks.

Needless to say, I’m really looking forward to Q3.

If you have any questions, comments or recommendations, please reach out by responding to this post or emailing at AutomatedTradingStrategies@protonmail.com.

Click here for strategy links.

Click here for FAQs.

Click here to subscribe and receive 2 months free added to a monthly or annual subscription until July 29.