Portfolio — how to blend multiple stocks

Once you have a verdict on a single stock, it's time to blend multiple stocks. The portfolio feature does three things: gather stocks in one table, compare verdicts and scores side by side, and propose how to weight them.

Step 1: Register stocks (up to 20)

Click the ★ star icon on a stock page, or type the name/ticker in the portfolio input box.

• US stocks: ticker (AAPL, MSFT)
• Korean stocks: ticker (005930.KS, 035720.KS) or name in Korean

Note: stocks with missing data or that can't be analyzed won't register. Check the "Not Qualified / Insufficient Data" tab at the bottom.

Step 2: Read the comparison table — stocks side by side

All registered stocks appear in one table. What each column means:

Column	Meaning	How to read
Current Price	Stock price now (5-min cache)	Reference only — real trades go through your broker
Composite Score	Aggregate of 8 dimensions (max 100)	Higher = better company, but separate from verdict
Quality	Business fundamentals: growth, profit, cash flow, financial health	Higher = more stable
Growth	Revenue and earnings expansion	Higher = faster expansion ahead
Valuation	Low P/E, cheap vs. history and peers	Higher = cheaper
Risk Score	Volatility and financial metrics	Higher = safer (lower volatility)
Price Verdict	Strongly Undervalued ~ Significantly Overvalued	The "buy/skip" criterion
Expected Upside	If price reaches fair value?	Scenario, not forecast

With few stocks, even one weak dimension can fit. For example, 2–3 mature, slow-growth, cheap stocks mixed with 2–3 high-growth, expensive young companies can lower portfolio risk.

When to switch tabs

• Candidates: conditions passed — worth analyzing ✓
• Verifying: first screen passed, AI analysis pending — not final yet
• Excluded: looks cheap but deteriorating (Value Trap) — skip

Step 3: Compute allocation — rule-based weighting

Enter an amount to invest, and the system proposes each stock's weight using fully disclosed rules.

The formula

Composite score × price verdict × confidence × risk × correlation

These five factors are multiplied to see how much each stock "pulls" on capital. Example:

• Stock A: high composite score (+), undervalued verdict (+), HIGH confidence (+) → larger weight
• Stock B: medium composite, overvalued verdict (-), MEDIUM confidence (-) → smaller weight
• If correlation is high (they move together), we lower one stock's weight → diversification

Capping rules applied

Base: no single stock over 30%, no sector over 40%, positions under 3% are dropped.

Why caps? Concentration in one or two stocks means one misjudgment destroys the whole portfolio. No verdict is 100% right. Five to seven stocks mixed means even if two are wrong, the others prop up returns.

With few stocks (e.g., only 3 registered)

Caps scale proportionally. A 3-stock portfolio might let one stock reach 40–50% weight. The system will note: "cap automatically adjusted for portfolio size."

The "why" behind each weight

Each allocation comes with reasoning. You'll see "Stock X is 12% because: composite score 7.2 × verdict (undervalued) × HIGH confidence × risk 9 = normalized to 12%." The logic is transparent. If you disagree, edit it.

Step 4: Understand the proposal; you don't have to follow it

The allocation is automated guidance only. Keep this in mind:

It is not a buy recommendation

It does not mean "buy in this weight." It means "if you own these stocks, this weight is statistically sensible." Real buying and selling are entirely your call.

The rules aren't perfect

• Your risk appetite may differ — the system defaults to diversification, but you might prefer to concentrate
• Fees and taxes aren't built in — real trading has costs
• It doesn't predict the future — past correlation doesn't guarantee future correlation

How to use it

1. Treat it as guidance, but adjust if you disagree
2. Honor the cap rules — the 30% cap alone cuts portfolio risk sharply
3. Triage by confidence — HIGH confidence moves fast, LOW confidence moves slow
4. Rebalance regularly — verdicts change; recalculate monthly or quarterly

Step 5: Execute (your job)

The proposal is a calculation. You do the buying:

• ✓ Enter the proposed weights into your broker
• ✓ Scale in over time (buy all at once = market impact risk; buy too slow = miss gains)
• ✓ Log trades on the tracking page — then watch whether your verdicts actually worked

Common questions

Q: Is this allocation really optimal?

A: "Statistically reasonable" is the best it can claim. Reality has variables you don't know (policy, FX, special cases), and rules are built on backward-looking data.

Q: What if I add or remove a stock?

A: Weights recompute instantly. Caps adjust for portfolio size.

Q: Can I ignore this and weight differently?

A: Absolutely. The portfolio page is a tool to audit your strategy, not a prophet to dictate it. Trust your judgment.

Q: Can I tweak the calculation rules?

A: The rules and coefficients are public but not configurable in the UI. Feedback is welcome if you have ideas for better ones.