Automated Trading Strategy #78: Two NQ Scalping Strategies
When modeled, 78v1 makes $8K on 1 contract, $80K on 10 contracts annually. 78v2 makes $100K on 1 contract, $1M on 10 contracts annually. I'll be putting both in the Q1 forward test.
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 shared 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.
I’m a big fan of the AI’s ChatGPT and Perplexity. In particular, their ability to assimilate large amounts of information and find patterns. Some say these generative AI programs can offer nothing new, but such a claim can only be made by the one observing a thing.
What I am most struck by is the ability to grasp concepts like god. If it can grasp concepts like god, why not the holy grail of automated trade strategy?
A good example of this is something I like to refer to as the “final bunny”. Joe Rogan coined the phrase after showing a series of pictures ChatGPT had created to depict an increasingly happier bunny.
The bunny starts out as just a regular happy bunny, but with each request for a ‘happier’ depiction, the bunny grows more and more psychedelic in nature until it turns into what is ostensibly a fractal. One could argue that the final bunny is at least one AIs depiction of god.
If you haven’t seen the depiction here’s an overview; it starts with the first picture in the upper left and flows to the right, then repeats on the next row and the next.
As a lifelong student of ancient text, the final bunny looks like a mandala to me:
According to AsiaSociety.org mandalas are:
…Buddhist devotional images often deemed a diagram or symbol of an ideal universe.
So there seems to be some cross section of belief between AI and what has become known in our culture as ancient wisdom, at least when it comes to a depiction of nirvana.
Is There A Final Bunny of Trading?
So I’ve been obsessing over this term ‘final bunny’. I’m constantly trying to work it into a conversation—as I’ve managed to do in this post. It was all I could do to not make it the title. Of course it wasn’t long before I wondered if I could apply the concept of a final bunny to our pursuit.
While I was able to glean several ways to improve profitability, neither ChatGPT or Perplexity really helped (shocker). I did, however, go down a fascinating rabbit hole with regard to a connection between fractal patterns and frequencies that might provide insights on the relationship between trading systems and the frequency of bar formation used for the data series (minute, tick, range, volume), but I’ll leave that to a future discussion.
Still, the thought experiment helped to articulate the final bunny in trading—not what it is, but what it could do. The final bunny of trading, if articulated in a way that all would agree on, is the guaranteed trade; a trade that is always profitable. I know that’s not a revelation for most, but it helps me to anchor the hunt.
Before going any further, I want to take a moment to say that anyone that tells you that a guaranteed trade exists is either a fool or hopes they’re talking to one. The hope is that by trying to articulate what a ‘final bunny’ is, like the concept of infinity, it gives us a placeholder—something to conceptualize as a goal. We might be able to get close to some form or rendering of it, but the actual thing itself cannot exist.
In the same way that most disciplines cycle between application and theory, market theorists do as well (from M&M to Keynes, I have yet to read a seminal economics paper that was based in reality). If you’re a trader you know that in “theory” the holy grail is most likely a scalping/pivot strategy of some sort. Theoretically, I can articulate what the strategy is, but the application is the final bunny.
We’ve covered pivots, most recently in Strategy 58, but what is scalping?
Here are a few definitions I found on the web:
Corporate Finance Institute - A day trading technique where an investor buys and sells an individual stock multiple times throughout the same day.
Investopedia - Scalping is a trading style that specializes in profiting off of small price changes and making a fast profit off reselling.
This is a bit like saying that a rainbow is a bunch of pretty colors.
Then you get definitions like this one from TD Bank:
To their credit, this is only an excerpt of a larger piece, but it makes no sense. The absolute worst time to scalp is when you’re expecting a big market event because if the price goes against you, it will go against you hard. You don’t want event driven volatility, you want a market that is predictably volatile at certain times.
So what are we looking for?
Like the definitions above suggest, there’s a vast chasm that exists between theory and application when it comes to scalping. In theory, the best scalping strategy is a strategy that runs on a 1-minute chart taking small wins consistently throughout the day. In practice, I can’t figure out a way to do this with any consistency, though I’m still developing a few ideas that I hope to share with you in the coming months. If you have one and would like to share, please let me know.
Theoretically, a scalping strategy is one quick, guaranteed trade after another. The size and speed of the trade makes the task of catching that final bunny easier than a long range strategy. Practically, a scalping strategy takes small wins and rarely has some small losses. Small is where I tend to get stuck. What can we do?
We can Frankenstein a scalping strategy with a Martingale (large stop loss and tight take profit), but it requires the ability to sustain a large loss or two.
We can also focus in on rare, but highly predictable, trade opportunities.
We’ve been focused on the latter for the last two months, but not specifically with the intent of creating a scalping strategy. So I asked two of the students I’m working with to give it a try. One came up with something that was great in theory, but he could never pull it off in the application. Welcome to the hunt guy! The other developed a strategy that resulted in decent results and she agreed to let me share that strategy with you. These are the results:
It made 82 trades on a profit factor of 99. The net income is small, but that’s to be expected with a scalping strategy. The beauty of scalping strategies is in the ability to scale. For example, if I increase the number of contracts to 10, we go from a net profit of $8K to a net profit of $80K:
This only contains one side of the trade, so it might also be possible to increase net income by adding the long leg, which I’m going to do in a moment, but using a very different approach.
The beauty of scalping trades from a psychological perspective is that you’re in and out of trades quickly (average length of each trade is 4.20 minutes), but it also requires fast trades so it’s the kind of strategy that may not backtest well.
This is a great start, but my concern is that the data series this runs on is highly sensitive to changes in market conditions. Thankfully, we have a way to test that sensitivity, which means we can also improve it.
What’s the best way to test a strategy’s sensitivity to changes in market conditions? By measuring its sensitivity to changes in the frequency of bar formation in the backtest. So the question naturally evolves to ‘how do you build strength into a strategy that, when backtested, is highly sensitive to changes in the data series?’
Answer: Find a ‘hot spot’ or multiple data series to forward test on. In particular, we’re looking for a strategy with a high profit factor regardless of how you change the data series or the parameters (within a specific range). So you want your hot spot to be in this target range.
That’s what I created with the long leg of Strategy 78 (78v2). It is slightly different from Strategy 78v1, but uses the same logic.
Before we get into how to create and/or download the strategies, I want to give a quick update on what to expect from ATS for the rest of January:
The rest of the month will be used to prepare for the live test with a start date of January 24. Look for a Kickoff memo around the 14th. It will cover:
the strategy selection process
the risk management system
performance update process
The Q1 2024 Backtest portfolio was published at the end of December. I used the Q1 2024 Backtest portfolio to create the Q1 Master Forward Test. Forward test results are published in the Mudder Report. To learn more about how to create your own forward test, click here.
The first Mudder Report of 2024 will be published around January 28th. It will include forward test updates on several portfolios. 2024 has already produced a few new ‘leaders of the pack’. Not all strategies make it the forward test, but I’ll definitely be adding Strategies 78v1 and 78v2.
January is a big month. If you have any questions, please let me know or comment below. If I haven’t responded to you, please send your email to the following address: AutomatedTradingStrategies@protonmail.com.
Now, let’s get into how you can recreate Strategy 78 for yourself.