Track record
The live book and the model behind it.
Below is the full record: the live example portfolio since inception, and the 3.8-year walk-forward backtrained model that was tested before we charged anyone. Simulated results are labeled as such — they are not a realized track record.
Track record
Winners compound. We show both.
Our edge isn't one headline return — it's finding stocks that double and letting them run. The model was built and walk-forward tested on 3.8 years of trailing market data. We actively trade a live example portfolio to demonstrate the process. That is not a recommendation to copy us — we expect you to do your own research and build your own portfolio.
Total return
+250.39%
3.8-year trailing history · walk-forward validated · backtrained model
Picks that doubled
5
In the 3.8-year backtest
Out-of-sample
+67%
Outpicked S&P · Jul 2024 – Apr 2026
Inception Apr 01, 2026 · Day 146
Live return
—
On capital deployed into picks — cash held back to fund future buys isn't counted.
- 2× winners
- —
- Winning picks
- —
- Open picks
- —
Our portfolio, published for transparency — not what you should buy. Trades are logged manually with real fill prices and dates, not auto-synced from a brokerage feed. Position sizes are illustrative, so every return here comes from price movement, not dollar size.
Jun 15, 2022 — Apr 06, 2026 · 35W/18L of 53 closed picks · 132 trades
Total return
+250.39%
vs +83.34% S&P 500
2× winners
5
Point-in-time fundamentals with a 90-day filing lag. Past simulated performance is not indicative of future results.
Every name the model ended holding
Winners circle — picks that doubled
5 total · 35W / 18L of 53 closed picksHow the backtrained model and live example portfolio work
Backtrained on 3.8 years of history. We built the model on trailing market data, then validated it walk-forward — training on one period, testing on unseen data (Jul 2024 – Apr 2026) before going live. This is simulated performance, not a track record — see the badge above.
Winners run, losers get cut. Position sizing and risk guardrails keep the book disciplined, but we don't cap upside on names that pay back their cost basis. In the backtrained model 5 positions doubled, and we sold out of them across 8 separate exits — trimming as they ran rather than closing all at once.
Live example portfolio, not instruction. We place real trades on our own capital, marked at each day's closing price, and publish every entry and exit. We're not chasing the perfect entry or exit — our research is about finding valuable businesses and cost-basis averaging into them over a long horizon. Position sizes shown are illustrative, so nothing here implies a fixed account size.
Backtrained model winners · picks that doubled
YPF
+199.87%
2023-11-20 → 2025-01-06
TGS
+189.83%
2023-02-21 → 2026-01-05
BMA
+179.01%
2023-12-04 → 2024-09-16
IRS
+160.98%
2023-07-17 → 2025-06-02
Concentrated by design, not diversified: four of these five doubles came from a single macro theme (Argentine equities, 2023–2025). A repeat requires a comparable dislocation, not a repeatable process across unrelated names.