Methodology

Goal

향후 5년 내 시가 10배(텐베거) 가능성이 높은 종목을 규칙 기반으로 스크리닝합니다. Score v2는 피터 린치 GARP(저평가 성장주)와 Yartseva(2025) 멀티배거 연구를 반영해 소형·FCF·지속 가능 성장·역발상 진입을 강조합니다.

Screen for stocks with high potential for a 10x price move within five years. Score v2 blends Peter Lynch GARP heuristics with Yartseva (2025) multibagger factor research: size, cash yield, sustainable growth, and contrarian entry timing.

Market rotation

The universe is refreshed daily from full exchange listings. Only symbols recommended in the last 30 days are excluded before screening.

Pre-filter

Low analyst coverage receives a small size bonus.

Pick vs none

So far, 0 of 112 market days had no pick. The lowest published score is 76.2. The cutoff has not filtered anything yet: in practice the top-scoring candidate is published every day.

Simple baselines

The Does the pick beat simple baselines? table on the performance page is not the benchmark excess percent. Each number is the published pick's H20 return minus the baseline's H20 return on the same date. Both sides subtract that date's benchmark, so the benchmark cancels.

The baselines are a random 10 names in the same market, random names whose market cap sits between half and double the pick, and same-day score ranks 2 through 5. H20 means 20 trading days. The table recalculates once a week, on the Friday as-of date.

The day count is small. Consecutive 20-trading-day windows overlap. On a day with a snapshot the random sample is that day's listing. Names delisted before 2026-10-09, and days without a snapshot, still use today's listing, so survivorship bias remains.

Excess-return gate

The performance page publishes the headline Excess figure only when a separate gate passes. Cumulative portfolio and benchmark curves may still render, but the Excess % stays hidden until all of the following hold (today:not published). This gate applies only to that Excess percent. The simple-baseline table is a same-date difference with the benchmark cancelled, so the gate does not cover it.

• Sample — ≥ 20 pick-complete rows on the cumulative horizon
• Benchmark — every pick-complete row on that horizon also has a complete benchmark return
• Horizon tier — presentation horizons only (1M / 3M / 6M / 1Y); engineering H20 / H60 excluded
• Price basis — #119 price-basis validation complete (adjusted_preferred or adjusted_auto)

Data & schedule

Score v3 candidate factors join the public methodology only after measurement gates pass. Score v2 remains live today.

Top-N observational tags

당일 Top-5 후보 행에만, 점수·선정과 분리된 관찰용 라벨을 붙일 수 있습니다. 비즈니스 요약(longBusinessSummary)과 TypeSafe API 키가 있을 때만 JEVsystem_one Choice로 분류하며, 선정·순위·Score v2에는 영향을 주지 않습니다.

Optional observational labels on Top-5 candidate rows only — separate from scores and the published pick. Classification runs via TypeSafe JEVsystem_one Choice only when a business summary and API key exist; it never changes selection, rank, or Score v2.

• Risk (#147) — one salient business-risk category
• Moat (#149) — none, narrow, orwide
• Megatrend (#149) — none,ai_infrastructure, power_grid, orrobotics

API 키 없음·요약 없음·오류·타임아웃 시 JSON에는 fallback provenance만 남기고 화면 라벨은 숨깁니다. 과거 일별 JSON(필드 없음)은 그대로 유효합니다.

On missing API key, missing summary, errors, or timeouts, entries keep concisefallback provenance in JSON but hide UI labels. Older daily JSON without these fields remains valid.