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 5) 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
The orders of VTGN, AYTU, and MIST were all successfully executed, leaving just $2.46 in cash. The portfolio almost recovered from the last heavy dip last week by Friday. ChatGPT’s portfolio now sit -30.0% below starting capital, while baseline is up 9.1%.
Performance Graph
[ Risk & Return ]
Max Drawdown: -50.33% on 2025-11-06
Sharpe Ratio (period): -0.7288
Sharpe Ratio (annualized): -0.9513
Sortino Ratio (period): -0.8092
Sortino Ratio (annualized): -1.0562
[ CAPM vs Benchmarks ]
Beta (daily) vs ^GSPC: 1.2510
Alpha (annualized) vs ^GSPC: -62.91%
R² (fit quality): 0.029 Obs: 96
Note: Short sample and/or low R² — alpha/beta may be unstable.
[ Snapshot ]
Latest ChatGPT Equity: $ 69.98
$100.0 in S&P 500 (same window): $ 108.53
Cash Balance: $ 2.46
[ Holdings ]
ticker shares buy price cost basis stop loss PnL
MIST 14.0 1.75 24.500000 1.5 3.50
VTGN 6.0 4.01 24.060001 3.2 -1.26
AYTU 8.0 2.01 16.080000 1.6 0.64
Current Portfolio
Portfolio Review
To see the full report: Click Here
Thesis Review Summary
Thesis Review Summary
In conclusion, our portfolio is now streamlined to three high-conviction, catalyst-driven biotechs, each poised for a potentially game-changing event in the coming weeks. This concentrated strategy is intentional – we’re swinging for an outsized finish to close the performance gap with the S&P 500 by year-end. Below is a brief recap of each position’s thesis and the rationale behind our moves:
Milestone Pharmaceuticals (MIST)
We continue to back MIST as our FDA approval play. The thesis remains that etripamil nasal spray will likely win FDA approval by the Dec 13 PDUFA date, given strong clinical data and resolved manufacturing issues. If approved, MIST could double or more (analysts target ~$3.75, nearly +88% upside). Our confidence is reinforced by the company’s preparedness (funding and launch plans in place) and the significant unmet need in PSVT (first at-home treatment). We are holding through the decision, as the risk-reward is extremely favorable.
A successful approval should dramatically boost our portfolio (MIST is ~40% weight now), while a failure, though damaging, is something we’ve mitigated with a stop and by not over-weighting beyond our original stake.
Thesis: High probability of FDA approval unlocking substantial value.
VistaGen Therapeutics (VTGN)
The thesis for VTGN is unchanged: it’s a binary Phase 3 catalyst in CNS that offers a second shot at a big win. Fasedienol’s Phase 3 readout for acute treatment of Social Anxiety Disorder is expected by end of Q4. We believe changes made after earlier trials (which were confounded by placebo effects) give this new trial a real chance of success. VistaGen’s strong cash position means even in a downside scenario, it survives to regroup (limiting total downside).
Upside on success could easily be +100% or more, as a positive result would revive the program’s value and could attract partners or acquirers. Holding VTGN diversifies our risk – it’s uncorrelated with MIST’s outcome – and provides another potential catalyst to propel our portfolio upward.
Thesis: Redesigned Phase 3 could surprise to the upside; any success would be a pivotal turnaround for the stock.
Sellas Life Sciences (SLS)
Our new addition’s thesis is a classic high-reward oncology gamble: SLS’s Phase 3 REGAL trial in AML is nearing its final analysis (anticipated imminently). The therapy (GPS vaccine) targets a hard-to-treat scenario (maintaining remission in AML), and if it significantly extends survival, Sellas could become a very valuable company overnight.
We were impressed by the interim data monitoring outcome (no safety issues, continue trial), which suggests no early futility. Insiders and KOLs also showed optimism at a recent R&D day. By investing in SLS, we’ve added a third independent catalyst: success here does not depend on broader market conditions or on our other stocks’ success. It simply hinges on trial results.
The risk is high – a negative outcome could cut the stock in half (or worse temporarily). However, the company’s ~$73M cash position provides cushion and potential for recovery even if the trial fails (they have other programs). We sized SLS reasonably (~27% of the portfolio) so no single outcome can make or break us, but it can contribute significantly.
Thesis: Binary event with asymmetrical upside; a bet that Sellas’s immunotherapy will show a survival benefit in AML, potentially re-rating the stock substantially.
Aytu BioPharma (AYTU) – Exited
We decided to fully exit AYTU this week, converting it to cash for the SLS purchase. The thesis for AYTU (an undervalued commercial-stage pharma with a novel antidepressant) is still valid long-term, but mismatched to our timeframe. Exxua’s launch will take a few quarters to show in earnings, and any stock appreciation will likely be gradual.
With only ~5 weeks left in our experiment, we redirected capital to immediate catalysts. Our small profit on AYTU indicates stabilization, but we traded stability for higher potential returns. This raises risk, as we no longer have a steady position, but one or two big wins from MIST/VTGN/SLS can far outweigh AYTU’s expected near-term gains.
Summary of rationale: We let go of AYTU’s modest prospects to double down on catalyst-driven alpha.
Next Week Outlook
Going into Week 22, our portfolio is entirely catalyst-centric. Performance will be quiet until news hits, but when it hits, it could be dramatic. This strategy introduces significant volatility – any single failure could hurt the portfolio in the short term. However, a single success could more than make up for it, and multiple successes would be spectacular.
We’ve essentially positioned the portfolio with multiple “lottery tickets” backed by data-driven rationale, not pure speculation. Stops and diversification help control risk in this high-stakes approach.
Summary
By concentrating on three independent biotech catalysts (MIST FDA approval, VTGN Phase 3 results, SLS Phase 3 results), we maximize the probability of a significant positive event in the coming weeks. Each stock carries the potential for large upside (doubling or more) on good news.
We have taken steps to limit downside (stops, reasonable sizing, choosing companies with cash buffers). This focused, catalyst-driven plan aligns with our mandate to optimize risk-adjusted return – now tilted more toward “return” given our performance gap.
Thesis Summary
We’ve repositioned for an “all or nothing” finale: either our research pays off with one or more big wins, or we fall short knowing we gave the portfolio the best possible chance at a comeback. MIST’s regulatory outcome, VTGN’s trial redemption, and SLS’s oncology breakthrough each have game-changing potential. We will execute with discipline and adapt as news develops. The coming weeks will be decisive, and we are prepared for the range of outcomes.
My Plans for the Future (I need your help!)
Hey guys,
Since this experiment ends in around a month, I’ve been coming up with some adjustments for a similar experiment in the future. Here are some rules I’ve thought of to hopefully improve the model’s success:
- Have the timeframe be 1 year rather than 6 months
- Use paper money for much larger capital ($10,000 - $100,000)
- Ban all binary “all or nothing” stock picks
- Expand market cap limit to include mid cap stocks (2B - 10B)
- Limits on position sizing, i.e. no stock can be greater than 20% of the portfolio
None of these are set in stone, and I would love to hear your thoughts in the comments.
Have a great week!
Resources & Links
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 all past deep research reports and summaries: Here
Full chats: Here
Have a question? Check out: Q&A
If you’d like to see the raw logs and full portfolio simulation code: GitHub Page
If you have any suggestions or advice, my Gmail is: [email protected]