Recap
Hey guys! If you’re new here, I am running a 6 month long experiment to see if a Large Language Model (like ChatGPT) can be a skilled micro-cap portfolio manager. I give it daily closing data at the end of every trading day and it has full control over its assets. Also, once every week it gets to use Deep Research to completely reevaluate it’s account. Can ChatGPT carve consistent alpha in the dangerous world of micro-cap stocks? Lets find out.
Overview
This week the portfolio hit another all-time high, up 32.6% overall, spearheaded by aTyr Pharma (ATYR). I accidentally placed an order for 4 additional shares instead of 3, but ChatGPT decided to keep the extra share. As a result, ATYR now makes up 50% of the entire account. Despite the insane risk (and a minor drawdown on Thursday), the portfolio managed to push higher into the close of the week.
Performance Graph
Current Portfolio:
- ATYR (aTyr Pharma) – 12 shares at an average cost of $5.21 (total cost basis: $62.48). Stop loss set at $4.22. This is the core holding, making up roughly half the account.
- ABEO (Abeona Therapeutics) – 4 shares at $5.77 each (cost basis: $23.08). Stop loss at $6.00. A second backbone position alongside ATYR.
- AXGN (Axogen Inc.) – 2 shares at $14.96 (cost basis: $29.92). Stop loss at $13.00. Pending exit order as catalyst was delayed.
- FBIO (Fortress Biotech) – 1 share at $2.85 (cost basis: $2.85). Stop loss at $2.00. Currently a small tactical position, with an order in to add more.
Metrics:
Max Drawdown: -7.11% on 2025-07-11
Sharpe Ratio (period): 1.3428
Sharpe Ratio (annualized): 2.9669
Sortino Ratio (period): 2.3796
Sortino Ratio (annualized): 5.2579
[ CAPM vs Benchmarks ]
Beta (daily) vs ^GSPC: 1.5980
Alpha (annualized) vs ^GSPC: 182.96%
R² (fit quality): 0.109 Obs: 47
Note: Short sample and/or low R² — alpha/beta may be unstable.
Portfolio Review
To see the full report: Click Here
Here was this week’s revaluation summary:
Trade Summary for 2025-09-08
Sell AXGN (Axogen Inc.)
- Action: Exit full position (2 shares) at $15.50 limit.
- Reason: FDA decision delayed until December; better to redeploy cash (~$31).
- Note: Will sell at open if price ≥ $15.50, partial fills allowed.
Buy FBIO (Fortress Biotech)
- Action: Add 2 shares (total now 3) at $3.80 limit.
- Reason: FDA catalyst Sept 30; position size small but event-driven.
- Risk Control: Existing stop at $2.00 applies to all 3 shares. No chasing above $3.80.
Buy FDMT (4D Molecular Therapeutics)
- Action: New 2-share position at $7.50 limit.
- Reason: Positive trial data; potential rebound toward long-term $25 target.
- Risk Control: Initial stop at $5.75 (~20% downside), just below support.
- Note: If gap >3% above limit, no entry — will re-evaluate.
In summary, our portfolio is positioning for strong catalyst-driven growth while tightening risk controls. We have exited Axogen to redeploy into more immediate opportunities, namely Fortress Biotech (upcoming FDA decision) and 4D Molecular Therapeutics (post-positive data momentum). Our core holdings Abeona and aTyr remain high-conviction to outperform the market – Abeona is transitioning to commercial stage with a funded runway , and aTyr is on the cusp of Phase 3 results that could be transformative.
These two form the backbone of our thesis, supported by solid fundamentals and news flow. Fortress and 4DMT add tactical spice: one a near-term binary event play, the other a beaten-down innovator with new validation. Overall, the portfolio is up ~32% in 9 weeks, far ahead of the S&P 500’s ~4.5% in the same period, and we aim to build on that lead. The coming week (Week 10) will be about catalyst anticipation and risk management – ensuring we have the right exposure levels as we head into late September’s event calendar. We will review the thesis on each name continuously and remain ready to adjust if the story changes.”
My Thoughts
ChatGPT has started to develop tunnel vision around low-success, orphan-drug catalyst plays rather than balancing with momentum or value-driven names. Obviously, there’s nothing inherently wrong with catalyst exposure, the issue is concentration. With only four holdings, leaning too heavily into binary outcomes turns the portfolio into more of a lottery ticket than a repeatable strategy.
Will ChatGPT’s high-stakes risks pay off, or will the portfolio come crashing back to Earth? Find out next week!
This project is purely educational and research-focused. Nothing here should be taken as financial advice. Full disclaimer: Here
GitHub Page and Email:
To see past deep research reports and summaries: Click Here
Have a question? Check out: Q&A
If you’re curious about the code I’m using to automate results and graph, the GitHub page is: ChatGPT-Micro-Cap-Experiment.
If you have any suggestions or advice, Gmail is : [email protected]