Formula AI: Month 2 Review
Introduction
Formula AI is a project I began in June 2025 to experiment with investing using AI as my guide. You can read the previous posts in my series here:
July 1st: Formula AI: Month 1 Review
As I reflect on Month 2, I wanted to capture my observations during the month and my feelings on the performance of my investment. Moving into Month 3, which began on August 1st, I shared my portfolio performance, observations, and ideas with ChatGPT to allow it to guide me through the next set of investment decisions.
Month 2 in Review
When I start a new project, I tend to become emotionally invested in my second step. My first step is all about setting up a structure and seeing what happens. My second step is where I start to see my feelings and emotions take over - with questions constantly bombarding me, such as, "what if I'm doing this wrong?" or "how do I know what I should do next?". What grounds me is sticking to my plan and seeing things through. Trust the process, as they say!
Throughout Month 2, I kept last month's takeaways at the forefront of my mind. I didn't obsess over exchange rates, but instead I looked at the growth of the stock itself and the overall growth of the portfolio. I was able to feel my feelings, knowing that I was following a strategy to take action once per month. I was learning about what my instincts were pushing me towards by listening to those gut reactions.
As last month ended with a loss of 1.73%, I was happy to see that reverse almost instantly within the first week of July. I remained optimistic. I was gobsmacked to see that my portfolio reached +14.07% on July 17th. This peak was staggering, and I started to ask myself whether this was compounded solid growth that would stick around, or if it was just a peak in a volatile market. I didn't investigate because learning more would not change my strategy, so I waited. The month ended with me sitting at +3.60% since this project’s inception. Compared to last month, I was up 5.52%, yet down almost 10% from that peak in July.
I was grateful, confused, and annoyed. So, I tackled these feelings with ChatGPT and entered Month 3 with renewed vigour to improve my system!
What I Asked ChatGPT
This month felt like I had a lot to do. I had 5.92 EUR in my cash account, and so I was expecting that any change to my strategy would either require more cash inflow or selling some of my shares. I built this project with a fixed amount of money - 4,000 EUR - and I wanted to keep it that way. I wanted this project to be self-sustaining. To start, I shared a snapshot of my current portfolio and asked how I could improve my strategy to capture some of this high growth in a systematic manner. As I worked through this process, I documented my questions to build an improved strategy. The questions are summarised below - bear in mind that when I started talking to ChatGPT for this month, I didn't know these questions would come up. They are questions I asked ChatGPT based on recommendations it made during the discussion. I wanted to understand the theory behind the concepts it raised before I decided if those strategies would work for me. These are some of the questions I asked:
If there is big growth, how can I lock in my gains? What rules can I implement?
What key metrics should I monitor to consider mid-month transactions?
What is a good threshold at which to consider trimming stock?
What is a trailing stop, and how can I use it in my strategy?
How can I implement these rules on my platform of choice - IBKR?
Now that I have a strategy for August, what trades should I make today to begin following this new strategy?
While working through these questions, I analysed the movements over the month for each of my stocks. On the day I hit the +14.07% peak, I pulled out the stock prices for each one to see how they each contributed to it. The increases ranged from +5.66% to +37.37%. As you can see, there was a lot of earning potential here that I missed out on by not taking notice of individual spikes in price.
I didn't beat myself up. I used this as a learning opportunity. I reminded myself of the original goals I set out when beginning this project, and gave myself the permission to release my feelings in the moment, and let the system work for me. Entering a new month was my chance to update this strategy so it would take care of the worries and anxieties for me.
Month 2 Decisions
It took some time to work through all of ChatGPT's suggestions so that I could understand the point of view, the impact on my strategy, and how the suggestions interact with each other. Some of the suggestions included:
Trim Positions After a Threshold: Sell a portion of your holding after a threshold increase is reached, so you can lock in your gains
Use Trailing Stop Limits: Automatically sell your position if it falls by a specified amount/percentage
Monitor Key Metrics: Use certain metrics to signal times to trade
Rebalancing: Reorganising your holdings to match your portfolio allocation plan
I wanted to get real numbers behind these ideas, so I took real examples from my portfolio as use cases. For each of these suggestions, I quizzed ChatGPT in order to understand the boundaries, limitations, and scenarios that creating rules might lead me to. In the end, I created 4 rules - one for each suggestion - and a rationale to support that rule, which suited my needs.
Spike: IF monthly gain > +15% THEN trim ~20–25% of the position
Hybrid Trailing Stop: IF overall gain > +10% THEN set a trailing stop of 3-5%
Return: IF total gain > +25% AND < +10% in the last 2 weeks THEN sell half, and rebalance
Rebalance: WHEN cash >100 EUR
Some of the language I use is correct investment vernacular, and others are terms I use to make sense of my project. For my rules below, I am using this language for my benefit, as I am tying the words to rules I will follow for August. These rules form a decision tree I can follow to remove emotional responses from my trading activity. Building a system and sticking with it is how you grow! One final note to keep in mind is that while I don't publish my blog straight away, I do all of my work on the first day of the month and record it. When I talk about the start or end of a month, I am talking about both calendar and project months because they align.
1. Spike
A Spike, for me, is a short-term gain. I define this relative to the stock price at the start of the month. I have set +15% as the benchmark. I set this value based on the experience of my portfolio (seeing between -10% losses to +37% gains), and I may change this in future months.
I set alerts in IBKR so that I would get an email if one of my stocks hit a spike. If any alerts trigger, I will sell 20-25%, depending on my holding at the time.
IF monthly gain > +15% THEN trim ~20–25% of the position
2. Hybrid Trailing Stop
A trail is a type of sell activity where you specify the number of stocks to sell, and what level of wiggle room you want to play with. The broker tracks the stock price from the moment you issue this kind of transaction. The peak price that is achieved while this transaction is active becomes the new reference point. If the price drops by the specified amount or percentage (as decided by you) relative to the reference point, the sale triggers. This has two important features that I liked.
The first is that the trail tracks the new peak for you. Let's say the price of a share is currently 100 EUR - this is your reference point. If you specify a 5% trail, the platform will sell your shares when they hit 95 EUR. Over time, as the share price moves, your reference point changes with the increases above your previous reference point. If the price increases to 105, that becomes your new reference point, and now the sale triggers at 99.75 EUR, which is a 5% decrease from 105. You can see that this is protecting more money than your earlier reference range while also allowing the position to grow. You are still locking in your 95 EUR but also capturing some of the new growth as well.
The second is that you can specify the wiggle room you want, in amount or percentage. I like to use percentages because all of my rules are based on percentages of growth. If I set a trailing stop, I am effectively locking in a percentage gain for my holding. If my stock is up +15% and I set a trail for 5%, then I am locking in +10% whenever the trail triggers. This feature is great because I can give my portfolio room to breathe while I sit back and know that I'm protected from big downturns.
The important distinction for this, compared to a Spike, is that this is based on overall growth since I bought the position. I set alerts to be notified when a position reaches at least +10% growth, and if it triggers, I set up a trailing stop.
IF overall gain > +10% THEN set a trailing stop of 3-5%
3. Return
My final rule is for stocks that grow slowly. I may never notice a spike in the stock if it grows by less than 15% during a month, but over time, those small increases can compound. If the overall growth reaches +10%, then I will implement my Hybrid Trailing Stop rule. But what happens if it just keeps growing?
In this case, I set alerts for all stocks to notify me when the position reaches +25%. I want this rule to give me access to capital for rebalancing my portfolio. If a position grows significantly and remains stable in its growth, it can start to impact my portfolio allocation over time. In this case, I will sell half of my holding and rebalance my portfolio. This is my way of locking in bigger profits and bolstering stocks that may not be growing as fast but that have more earning potential.
The final addition to this rule was a caveat about the increases in the past 2 weeks. This protects from falling into Spike territory. If the gain in the last two weeks was more than 10%, then I was looking potentially at a spike or market movement, and not a slow, solid growth. This addition gives me safety in avoiding big market movements. This rule's aim is to capture solid growth and reallocate it. If the entire market is suddenly up 10% today, then I could be selling my position just to buy other stocks at inflated prices. The risk is that the market falls again, and those stocks fall much harder.
IF total gain > +25% AND < +10% in the last 2-weeks THEN sell half, and rebalance
4. Rebalance
In order to understand rebalancing, I want to revisit the idea of a portfolio allocation. When I started this project, I said I wanted to use the following profile: Core (60%), Growth (30%), and Speculation (10%). Each group contains different stocks of differing volatilities. During each monthly review, I check if my portfolio is still in line with this allocation. If all the stocks increase in the same way, then it should remain the same. But if one group grows a lot, and another goes down, then it becomes unbalanced.
Rebalancing is the act of buying and selling shares in order to reset your portfolio according to your portfolio allocation. To do this, I specified that I will only rebalance when I have cash over 100 EUR (which is 2.5% of my portfolio). I decided upon this number because I wanted to maintain a small cash reserve for transaction fees, but not so much that I am sitting on a cash-heavy portfolio.
WHEN cash >100 EUR
My Rules in Action
I spent time going back and forth with ChatGPT to make sure I understood each of the rules entirely. I wanted to be confident that I was comfortable with all of the conditions, all of the reactions, and all of the alerts and transactions within IBKR. Once I established the rules, I implemented them immediately using Month 2's ending figures as a live example.
ARKI | Core | +15.28% since purchase
Hybrid Trailing Stop: Set trailing stop loss of 3% so I can lock in +12% gains
RGTI | Growth | +13.58% since purchase
Hybrid Trailing Stop: Set trailing stop loss of 3% so I can lock in +10% gains
RGTI | Growth | +23.97% since last month
Spike: Sell 20% of the position now (11 shares @ 14.12 USD each)
Rebalance: Buy IONQ (also a Growth stock) using my cash (since cash > 100 EUR)
All:
Set +10% alerts for QUTM, IONQ, QBTS
Set Spike alerts for all using this month’s starting prices
Set +25% price alerts for all stocks
The transactions I made included:
Sell - 11x RGTI @ $14.12 = $155.32 (-$1.00 fee)
Buy - 4x IONQ @ $38.16 = $152.64 (+$1.00 fee)
The Cost Basis metric is used to tell you how much cash you have invested in a position at the time of purchase. It is the average of your costs divided by the number of shares. As you start making more transactions, the Cost Basis can change. Knowing an accurate Cost Basis is what allows you to calculate the percentage increases and decreases - luckily, IBKR tracks this for you.
The Numbers
As I am holding EUR and USD stocks, I track the exchange rates used on each day I transact. These are used in my calculations for conversion to EUR, and in updating my figures and percentage changes. For reference, these rates are:
EUR-USD FX Rate (3rd June): 1.14
EUR-USD FX Rate (1st July): 1.18
EUR-USD FX Rate (1st August): 1.16
My portfolio over time is shared below, with the initial investments per stock (Cost Basis), last month's values, and this month's values. Percentage changes are also included based on the monthly movement and overall changes since Day 1. The "Next Month" columns show changes to the cost basis and value as I begin next month. I added this since my transactions change the cost basis, and this is valuable to know, especially as I start to rebalance the portfolio.
| Category | Ticker | Cost Basis | Last Month - Value | This Month - Value | % Change (Month) | % Change (Total) | Next Month - Cost Basis | Next Month - Value |
|---|---|---|---|---|---|---|---|---|
| Core | QUTM | 1,197.60 | 1,187.87 | 1,190.90 | +0.26% | -0.56% | 1,197.60 | 1,190.90 |
| Core | ARKI | 1,184.09 | 1,244.60 | 1,338.60 | +7.55% | +13.05% | 1,184.09 | 1,338.60 |
| Growth | IONQ | 596.97 | 611.44 | 565.84 | -7.46% | -5.21% | 750.15 | 692.55 |
| Growth | RGTI | 581.92 | 543.09 | 672.33 | +26.11% | +13.32% | 473.56 | 535.97 |
| Speculative | QBTS | 391.25 | 338.00 | 374.20 | +10.71% | -4.36% | 391.25 | 374.20 |
| Cash | EUR | 52.01 | 5.92 | 5.92 | 0.00% | 0.00% | 4.19 | 4.19 |
| Cash | USD | 0.00 | 0.00 | 0.00 | 0.00% | 0.00% | 2.30 | 2.30 |
| Total | 4,003.84 | 3,930.92 | 4,147.79 | +5.52% | +3.60% | 4,003.14 | 4,138.71 |
Notes:
All amounts are in EUR.
"This Month" refers to the values at the end of Month 2, and the "Next Month" columns refer to values after executing transactions when entering Month 3.
Due to currency exchanges, rounding, and transaction timing, there may be minor discrepancies between "This Month - Value" and "Next Month - Value".
Percentage changes indicate overall growth (stock changes and currency changes)
Changes in "Cost Basis" noted here are due to transactions that took place
USD instruments had their cost basis converted to EUR at the time of original purchase
What I’m Watching Next
I will be watching my new alerts for +10% and +25% thresholds on each of my stocks. I expect to make some trades based on my Spike rules, and perhaps the Hybrid Trailing Stops if there are some significant market movements. I am aware that I have yet to start building a watchlist of other stocks that I can use for rebalancing if any of my transactions get triggered. I will start building a watchlist during August so that I can reinvest my money rather than let it sit in idle cash.
I am also aware that as transactions become more frequent, it becomes harder to rely only on my spreadsheet tracking. I will try to rely on IBKR's reports instead to simplify my tracking next month.
Takeaways
As I navigate this process, here are some key highlights I would like to share from this journey.
Trusting the Process: My original goal was to build a system that I could use to run my investment and remove emotional ties. I noticed that there were some significant growth spikes during July, but I resisted transacting and going against my rules of the last month. It was difficult to sit back and watch the market; however, since I followed my rules, I was able to craft an updated system for August to capture anomalies like this.
Forgetting about "What if...": I realised that I could track prices based on leaving everything alone, but I don't want to start engaging in that. Seeing that my new strategy results in a gain over the old one would be validating, but seeing it result in less growth would be disheartening. Instead, I am building a system that allows me to tackle the "what ifs" I missed, and not the "what ifs" that are just speculation and could-have-beens.
Learning: I was pleasantly surprised to learn about new terms and strategies as I ended Month 2. Having some practical case studies to examine was helpful for my knowledge, as I could see things playing out in practice. I am hoping that as I move into future months, I will learn even more.
Taxes: As I move into Month 3, I am already aware that I will need to report on gains and losses because I sold 11 shares of RGTI. IBKR should provide me with tax reports at the end of the year to help with this. Of course, in the meantime, I have it on my to-do list to check my local obligations. I don’t want to be surprised at the end of the year about the paperwork I will need to submit.
Watchlist: Now that I am entering the realm of selling stocks, I want to build a watchlist of holdings I can invest in when any of my sales execute. I don't like holding a lot of cash in this portfolio, as that is not my objective.
Stay tuned for next month!
Take care!
Kiki <3
