Verdict Criteria — The 5 Levels and What Confidence Means

jini's verdict is not a fixed opinion — it's a function of your risk tolerance. Adjust your safety margin and the verdict recalculates instantly. Miss a quality threshold and the system automatically downgrades the rating. The verdict is yours to dial in.

Fair value table (4 methods × scenarios), final verdict, and margin-of-safety inputs (top right)

The 5-Level Verdict

Your upside (fair value ÷ price − 1) maps directly to five categories:

Verdict	Upside (default)	Meaning
Strong Undervalued	≥30%	Worth close study — compelling price
Undervalued	≥15%	Reasonable price — consider a position
Fairly Valued	−15% to +15%	Market has it right — limited margin of safety
Overvalued	≤−15%	High expectations baked in — avoid buying
Significantly Overvalued	≤−30%	Dangerous valuation — strong avoid

Upside is simple math: if fair value is $30 and price is $20, upside is 50%.

Safety Margin — Your Threshold, Your Verdict

The 15%/30% numbers above are just defaults. Every stock page has a slider. Move it and the verdict recalculates in real time:

Adjust the margin:
  Undervalued: 15% → 20%
  Strong Undervalued: 30% → 40%

Result: a stock with 18% upside shifts from 
"Undervalued" to "Fairly Valued"

This is core to jini's design. Your verdict is not set in stone — it's a live function of how much upside you require before you'll call something cheap.

What margin means

• 15% margin: A true risk-taker; 15% upside is enough to be "undervalued"
• 30% margin: Conservative; only 30%+ upside counts as "undervalued"
• 50% margin: Very cautious; only extreme bargains count

Set your margin and jini recalculates. This is not a company rating — it's your personal valuation template.

Confidence: How Sure Are We?

Two "undervalued" verdicts aren't created equal. One might rest on solid consensus across all four valuation methods. The other might be driven by a single outlier. That's what confidence levels capture:

Level	Meaning
HIGH	3 or more methods agree (dispersion ≤30%)
MEDIUM	2+ methods mostly align (dispersion ≤60%)
LOW	Serious disagreement among methods (dispersion >60%)

LOW confidence is a red flag. Example:

Stock: 40% upside (undervalued), but confidence is LOW

Breakdown:
  Multiples: Fair value $20 (−50% vs price)
  DCF: Fair value $35 (neutral)
  Peer: Fair value $30 (−20%)
  Historical: Fair value $28 (−10%)

Diagnosis: Only multiples method sees steep undervaluation.
The other three disagree. Use the verdict cautiously.

The page shows the actual dispersion number (e.g., "45% variance") so you can see why confidence is low.

Five Safeguards Against False Positives (Automatic Downgrades)

Just because a stock looks cheap by one metric doesn't mean it's undervalued. jini checks five gates before confirming an "undervalued" verdict. Fail any gate and the verdict steps down one level (max 2 steps):

Gate 1: Quality Score Must Exceed 45

"Cheap stocks are cheap for a reason." If the business is deteriorating (quality score <45), it's not undervalued — it's a trap.

Example: P/E of 8 (looks cheap) + Quality Score 30 (declining business) → Downgrade to "Fairly Valued"

Gate 2: Financial Health Must Exceed 40

High debt + weak liquidity = a cheap price hiding serious risk. A stock with 50% upside loses that appeal if the company can't service its debt.

Example: 50% upside + Debt/Equity 300% → Downgraded for financial risk

Gate 3: Type D/E High Confidence Triggers Automatic Downgrade

We'll cover this in the company types guide. Value Trap (Type D) and Cyclical Trap (Type E) classifications, when high-confidence, override an otherwise bullish valuation.

Example: AI flags Type D (value trap) with HIGH confidence → Downgrade even if metrics say undervalued

Gate 4: Bear Case Must Not Exceed −30% Downside

If even in a reasonable downside scenario the stock could fall 30%+, the downside risk outweighs the upside. The margin of safety vanishes.

Example: 30% base upside, but −40% in bear case → Downgraded for skewed risk/reward

Gate 5: Methods Must Substantially Agree

This is the confidence gate. If methods are in serious disagreement (confidence LOW), the verdict is downgraded.

Example: Half the methods bullish, half bearish → Downgrade to reflect uncertainty

Reading a Downgraded Verdict

On the screen:

🟡 Verdict: Undervalued (downgraded: low quality)

Raw rating: Strong Undervalued
Downgrade reason: Quality Score 40 < 45
Final verdict: Undervalued

Translation: "By the numbers it's strong undervalued, but the business is weak, so we're toning it down to undervalued."

"Withheld" Verdicts

Sometimes jini can't issue a verdict at all:

• Fewer than 2 valid methods: jini refuses single-method verdicts
• Current price is missing: Can't calculate upside

When this happens, you see the calculation results but no final verdict. It's honesty: "Not enough confidence to judge."

The Verdict as a Formula

Verdict = f(upside, quality, financial health, type, method agreement)

Every input is visible. You control the upside threshold (safety margin). The system checks quality, debt, and type automatically. And method agreement is shown as a dispersion number.

When you slide your safety margin from 15% to 25%, that's the only change in the formula — and the verdict recalculates to reflect it. That's why jini calls itself decision material, not a recommendation. The verdict is yours to dial.

Next: Company Types and Traps

Some downgrade triggers depend on what kind of company it is. A temporary earnings miss (Type B) is different from a structural value trap (Type D). Understanding the company's undervaluation type is half the judgment. Read Company Types next.