The Strategy
A plain-English description of exactly how the system picks, sizes, and sells stocks. This is the live algorithm — the same logic that runs the Alpaca paper account.
In One Sentence
Buy high-growth companies whose analyst estimates are being revised upward, hold them while the thesis stays intact, let winners run with conviction adds, and continuously cut the weakest names — all ranked sector-by-sector so we compare apples to apples.
1. How Every Stock Is Scored
Each stock is graded on five factors. A factor score is a percentile rank within the stock's own sector (we never compare a software company to an oil company). Percentiles convert to letter grades (A+ down to F), and the five factors combine into a single Quant Rating from 1.0 to 5.0.
| Factor | Weight | What it measures |
|---|---|---|
| Growth | 35% | Revenue & earnings growth, both trailing and forward. |
| Revisions | 30% | Direction of analyst estimate changes — the strongest leading signal. |
| Profitability | 15% | Margins, return on equity/assets/capital, cash-flow quality. |
| Momentum | 15% | 1/3/6/12-month price returns, weighted toward recent performance. |
| Valuation | 5% | P/E, PEG, EV/EBITDA, price/sales. Weighted low on purpose — cheap is often cheap for a reason. |
After the weighted blend, the composite score gets a few adjustments:
- Earnings surprises: a beat adds up to +8 points; a miss subtracts up to −8 (post-earnings drift).
- Insider buying: net insider purchases over the last 90 days add up to +5 points.
- Absolute-momentum brake: any stock down over the trailing 12 months is penalized −20 points, even if it looks good otherwise.
2. What's Even Eligible
- Market cap of $300M+ and share price of $5+ (no micro-caps or penny stocks).
- ETFs are excluded — individual companies only.
- A sector needs at least 15 names before we rank within it (so percentiles are meaningful).
- Bankruptcy filter: any stock with an Altman Z-Score below 1.8 is thrown out entirely, regardless of score.
3. When We Buy
A stock has to clear every one of these gates to be buyable:
- Overall Quant Rating of 4.0 or higher
- Revisions grade B+ or better
- Growth grade B or better
- Profitability grade D or better (must have real data)
- Valuation grade C- or better (not wildly overpriced)
Each evaluation we add the single best qualifying name, building the book gradually rather than buying a basket all at once.
4. Sizing & Letting Winners Run
- Up to 50 positions, each new buy sized at about 3.5% of the portfolio.
- Conviction adds: if a holding is up 30%+ and still top-rated, we buy more of it.
- Taking our money off the table: when a position is up 60%+, we sell just enough shares to recoup what we originally paid for it.
- House money: the shares that remain are pure profit, and they ride uncapped — we never trim a winner again until the fundamentals say sell.
- Sector cap: no more than 30% of positions in any single sector.
5. When We Sell
No fixed holding periods and no trailing stops (they cut multi-baggers short). We exit only when the fundamentals say so:
- Strong sell: Quant Rating drops below 1.5 → full exit, even if it's a winner.
- Hold removal: Quant Rating drops below 2.5 → full exit (winners will have already banked their cost basis, so what's sold is pure profit).
- Underwater stop: a position that's been underwater for 150+ days and has a Quant Rating under 3.0 is force-sold — it clears out stale losers the rating alone is too slow to catch.
6. Portfolio-Level Risk Brake
A drawdown circuit breaker halts all new buying once the portfolio falls 15% from its peak, and only resumes buying after it recovers to within 10% of that peak. Sells still happen while the brake is on — we stop adding risk, we don't stop reducing it.
7. The Schedule
- Daily (after market close): refresh prices and rescore.
- Weekly (Saturday): full refresh of fundamentals, analyst estimates, and revisions.
- Every other week: run the full evaluation — sells, cost-basis recovery on new winners, and the next buy.
- All orders execute through Alpaca paper trading.
Informational only — not financial advice. Past performance does not guarantee future results.