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 real time 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 out consistent alpha in the dangerous world of micro-cap stocks? Lets find out.
Overview
Week 7 was great! Both the market and the portfolio pushed to new highs, with the portfolio briefly touching $126.94 on Thursday. Over the same five-day stretch, the S&P 500 gained 0.94%.
The standout move came on Wednesday, when ChatGPT finally put its idle cash to work. It deployed $40.72 (32% of total capital) into aTyr Pharma Inc. (ATYR), the largest single allocation shift in weeks. This was a clear conviction play, targeting the company’s Phase 3 readout for Efzofitimod in pulmonary sarcoidosis, expected in mid-September. With prior trial data showing strong reductions in steroid use, the upside case is massive if results are positive.
Of course, this is a binary bet, either the trial succeeds and ATYR spikes, or it fails and the position collapses. By sizing it at roughly a third of total capital, ChatGPT is signaling it’s willing to embrace that volatility as a potential portfolio-maker.
Performance Graph
Metrics:
[ Risk & Return ]
Max Drawdown: -7.11% on 2025-07-11
Sharpe Ratio (period): 1.0897
Sharpe Ratio (annualized): 2.9812
Sortino Ratio (period): 2.0442
Sortino Ratio (annualized): 5.5924
Note: Due to smaller sample size, total ratios are more accurate than annualized.
[ Capital Asset Pricing Model (CAPM) vs Benchmarks ]
Beta (daily) vs S&P 500: 2.0598
Alpha (annualized) vs S&P 500: 158.26%
R² (fit quality): 0.148 Window: 33
Note: Short sample and low R², alpha/beta are unstable.
Portfolio Review
To see the full report: Click Here
Here was this week’s revaluation summary:
Trade Execution Plan (Orders for Week of Aug 18, 2025)
Sell Orders
Actuate Therapeutics (ACTU) – Aug 18, 2025
- Place a limit sell for all 6 shares at $8.00 (GTC/DAY).
- Locks in ~40% gain while ensuring no sale below recent price ($8.22 last close).
- Rationale: Taking profit and freeing capital for better opportunities.
Esperion (ESPR) – Aug 18, 2025
- Place a limit sell for all 2 shares at $1.60 (DAY).
- Around stop level and just below ~$1.63 pre-market price (post-earnings) to ensure execution.
- Rationale: Redirect funds to nearer-term catalysts; minimal portfolio impact.
Buy Orders
Inspira Tech (IINN) – Aug 18, 2025
- Place a limit buy for 10 shares at $1.25 (DAY).
- If filled, set stop-loss at $1.00 (~20% downside).
- Rationale: Positioning for upcoming contract catalysts with defined risk.
Axogen (AXGN) – Aug 18, 2025
- Place a limit buy for 2 shares at $15.10 (DAY).
- If filled, set stop-loss at $12.00 (~20% drawdown protection).
- Rationale: Positioning ahead of Sept 5 FDA decision; risk managed with stop.
Adjustments
Abeona (ABEO) – Stop-loss raised
- Raise stop-loss from $4.90 → $6.00 on 4 shares.
- Rationale: Lock in no-loss floor after run to $7+, while allowing early launch volatility.
Execution Notes
- All orders to be placed at market open, Aug 18, 2025 (Monday).
- Sell orders (ACTU, ESPR) should free funds before buys trigger.
- If sells fail (e.g., gap down below limits and no recovery), buys won’t execute due to insufficient cash.
- In that case: reassess (adjust limit or skip purchase).
Updated Portfolio Thesis Summary (Post-Trade)
ABEO (Abeona Therapeutics) – Gene Therapy Launch
- Thesis: First-and-only one-time treatment for RDEB; FDA-approved and launching Q3’25. $155M PRV sale funds launch; strong early demand and payer coverage.
- Catalysts: First patient treated (Q3), initial sales reported Q4’25.
- Risk: Launch execution.
- Position: Hold (Stop raised to $6.00).
IINN (Inspira Tech) – Respiratory Device Turnaround
- Thesis: $22.5M ART100 order validates product; more contracts (including gov’t) in pipeline. Transitioning to revenue phase.
- Catalysts:
- New deal announcements
- 2H’25 product deliveries
- Risk: Possible dilution (shelf filed) or order delays.
- Position: Buy new position (Stop $1.00).
AXGN (Axogen) – Nerve Repair FDA Bet
- Thesis: Leader in nerve repair with growing sales; Avance graft FDA decision (Sep 5, 2025) could grant 12-yr exclusivity and drive growth.
- Catalyst: PDUFA Sep 5, 2025 (likely approval).
- Risk: Binary FDA outcome.
- Position: Buy & hold through PDUFA (Stop $12.00).
ATYR (aTyr Pharma) – Pulmonary Sarcoidosis Phase 3
- Thesis: Efzofitimod Phase 3 readout mid-Sep 2025; prior data positive in reducing steroid use. Huge upside in high-need disease.
- Catalyst: Top-line results ~Sep 15, 2025.
- Risk: Binary trial outcome.
- Position: Hold small position (Stop $4.20).“
Full chat log from 8-1 to now: here
My Thoughts
I changed up the prompts, and I’m glad it’s finally producing detailed trade orders instead of just evaluating companies. Before, I had to extract orders after deep research, which meant the model was basically guessing on pricing. Now it makes immediate, actionable decisions, after I forced the issue by rejecting a buy when it asked for permission.
On rebalancing: I’ve been waiting weeks for it to cut ACTU due to fading momentum, and I’m glad it finally pulled the trigger. The portfolio feels a lot more active and conviction-driven now. Although, I am very disappointed that it wants to buy IINN (Inspira Technologies) AGAIN. I was hoping the model would diversify into a fresh catalyst instead of buying in on a losing thesis.
I’ve been getting kind of frustrated with the constant attempt at rebalancing just to end up making any decisions, so I’m excited to see such a drastic shift in behavior. Given the catalysts ahead, I expect heavy volatility going forward, so stay tuned!
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 Frontend help would be appreciated :)
If you have any suggestions or advice, my Gmail is : [email protected]
Thanks for reading! Subscribe to get an update every week and support my work.