PEAD on Korea Invest Insights

Crimson Desert 5M: The Sell-Side Consensus Gap

Sat, 18 Apr 2026 09:00:00 +0900

This post is part of the Pearl Abyss Crimson Desert Thesis series. For the franchise IP re-rating thesis, see the earlier post Pearl Abyss: Crimson Desert Conquers China — 86% Steam Rating & 4M Copies Sold. That post examined the China market breakthrough and production-level success confirmation. This post focuses on a different question: why does sell-side consensus still lag, and what does that gap mean analytically?

The Central Question: 5M Copies Confirmed — So Why Is the Stock Still Slow?

Pearl Abyss (263750.KQ) officially confirmed the 5 million copies sold milestone for Crimson Desert on April 15, 2026. This is not a rumor, projection, or beta metric — it is a direct company disclosure. The question that production success raises is: “Is the game viable?” That question is now answered, definitively, in the affirmative.

Yet the stock’s price action since the milestone has been notably subdued. As of the April 17 close, shares traded at 54,100 KRW, with volume at approximately 1,081,588 shares — roughly 40% of the trailing average of 2,691,280. Relative performance versus KOSPI on the same day was approximately -5 to -6 percentage points in a session where KOSPI gained over 2%.

One analytical framework for understanding this divergence is what can be called the sell-side consensus gap thesis: the lag is not caused by information asymmetry — the market knows about the 5M milestone — but by an interpretation delay driven by how sell-side models update and how institutional capital is positioned relative to quarterly earnings confirmation cycles.

This working thesis does not constitute a buy or sell recommendation. It is an examination of market mechanics around a specific information-processing dynamic.

The Consensus Snapshot: Where Sell-Side Currently Stands

The following table summarizes publicly available full-year 2026 unit sales estimates from Korean sell-side research houses, as of the date of this analysis. These are the figures that currently anchor institutional modeling.

Brokerage	2026 Full-Year Sales Estimate
SK Securities	4.95M copies
Mirae Asset Securities	5.00M copies
NH Investment Securities	5.26M copies
Samsung Securities	6.00M copies
DS Investment Securities	8.00M copies

Source: Individual brokerage research reports. Figures reflect estimates as of the time of publication and may have been updated subsequently.

The structural pattern is clear. Four of the five tracked brokerages have full-year estimates at or below 6 million copies. Three of those four sit at 5.26 million or below — meaning their full-year estimates are functionally already achieved or exceeded as of April 15, with approximately eight and a half months remaining in the calendar year.

Only DS Investment Securities carries an 8 million estimate, which better reflects a post-5M normalization trajectory. The consensus, by any weighted reading, remains anchored to a pre-milestone world.

This is the gap. Not a gap in disclosed information, but a gap between the fact on the ground and the model universe most sell-side analysts are still operating inside.

Why Sell-Side Models Lag: Three Structural Mechanisms

Understanding why this gap exists requires understanding how sell-side research actually operates — not as a continuous real-time feed, but as a periodic, model-driven system with specific update triggers.

Mechanism 1: The Quarterly Earnings Reset

The most important single mechanism is the earnings confirmation cycle. Institutional capital — particularly long-only funds and active managers with quarterly attribution accountability — does not re-rate stocks primarily on press releases or milestone announcements. It re-rates on audited, line-item-level P&L confirmation.

Knowing that 5M copies were sold is useful. But knowing how those 5M copies translate into Q1 2026 revenue recognition, gross margin, and operating income requires a published income statement. Until that income statement exists, the institutional base tends to hold rather than expand.

Q1 2026 earnings are scheduled for release on May 12, 2026. Under this lens, that date — not the April 15 sales announcement — is the true forced-recalibration point for sell-side models.

Mechanism 2: Model Inertia and Rebuild Cost

Updating a sell-side financial model is not trivial. Changing the top-line unit sales assumption requires propagating the change through ASP assumptions, platform mix (Steam vs. console vs. regional), royalty cost structures, and quarterly phasing. A Korean brokerage analyst running a Korean gaming company model has additional complexity: Crimson Desert is a packaged AAA title in a market where Korean analyst teams have historically modeled live-service mobile games, not one-time purchase PC/console titles with long-tail decay curves.

The practical result: even when a milestone announcement arrives, model updates tend to wait for the next regular report cycle, not the day of the announcement. This is not analyst negligence — it is the rational allocation of research bandwidth under time and validation constraints.

Mechanism 3: The Analytical Frame Has Shifted — But Not Everywhere

The most sophisticated reframe after the 5M milestone is this: the debate is no longer “will Crimson Desert succeed?” That is a resolved question. The active debate is now about deceleration rate and terminal unit count.

A stock price driven by uncertainty about binary success will respond dramatically to success confirmation. A stock price where the market has moved on to debating the slope of the decay curve will respond more slowly and more episodically. Part of what makes post-5M stock behavior feel “slow” is that the most information-rich participants are already pricing a different distribution of outcomes — one anchored to deceleration assumptions rather than launch-risk uncertainty.

Velocity Analysis: Reverse-Engineering the Sales Trajectory

One of the cleaner analytical tools available here is a simple velocity calculation from the two official public datapoints.

Date	Confirmed Copies Sold	Incremental	Days Elapsed	Daily Run Rate
April 1, 2026	4,000,000	—	—	—
April 15, 2026	5,000,000	1,000,000	14	~71,400 / day

The April 1 to April 15 interval produced approximately 71,400 units per day — the most recent official velocity observation. This is an important baseline because it captures the game’s trajectory after initial launch enthusiasm had partially normalized, incorporating any Qingming Festival tailwinds.

From this baseline, one can construct scenario-conditioned estimates for the remainder of the year. The critical insight: significant deceleration is already embedded in every scenario below Bull. The scenarios are not extrapolations of peak velocity, but modeled decelerations.

Scenario Framework: Three Normalization Paths

The following table presents three independent scenario analyses for Crimson Desert’s full-year 2026 unit trajectory. These are not Pearl Abyss guidance figures, not sell-side consensus estimates, and not predictions. They are analytical scenarios constructed under explicit assumptions about deceleration rate.

Scenario	Full-Year 2026 Units	Incremental Needed	Required Daily Avg (≈260 days)	Key Assumption
Bear	7.5M	2.5M	~9,600 / day	Heavy deceleration; limited content; no new platforms
Base	8.5M	3.5M	~13,500 / day	Moderate deceleration (~81% below April velocity); regular updates
Bull	10.0M	5.0M	~19,200 / day	Mild deceleration; content expansion; possible new region or platform

A critical observation: the Bear scenario requires just ~9,600 units per day — roughly 87% below the observed April 1–15 velocity of 71,400. Even the most pessimistic modeled path produces a full-year outcome (7.5M) well above the majority sell-side position (≤5.26M for three of five brokerages).

The Base scenario at 8.5M implies ~13,500 units per day — an 81% deceleration from the recent observed rate. This is a conservative, not heroic, assumption.

This is the structural reason the consensus gap exists: if even a Bear scenario produces outcomes above most current full-year estimates, the current consensus is almost certainly anchored to outdated assumptions.

What the Market Is (Still) Getting Wrong

Under this analytical framework, the primary mispricing is a framing lag: a meaningful portion of the sell-side is still running models calibrated to answer “will the game succeed?” rather than “how slowly will it decelerate?”

These are fundamentally different questions. The first is a binary/bimodal distribution problem. The second is a continuous parameter estimation problem. Markets price these very differently.

When the key question is binary, a success confirmation triggers sharp repricing. When the question has already pivoted to deceleration slope, the market needs additional data points — specifically, the May 12 Q1 results and subsequent sales cadence disclosures — before it can anchor its deceleration estimate with statistical confidence.

The genuine observable opportunity in the consensus gap is the delta between where sell-side models currently sit (5M–6M majority) and where they will be forced to move after Q1 earnings require a model rebuild. That delta is quantifiable. The timing of its resolution — May 12 — is visible on the calendar.

Near-Term Catalysts

Three near-term events are identifiable under this framework as price-relevant:

1. May 12, 2026 — Q1 2026 Earnings Release The single highest-impact near-term event. If Q1 revenue recognition materially exceeds embedded consensus assumptions — which the velocity analysis suggests is plausible — sell-side models will update, price targets will revise upward, and institutional capital waiting for hard data confirmation will have the basis to act. This is the structural analog to Post-Earnings Announcement Drift (PEAD): stocks that beat consensus tend to continue drifting positively for weeks after the announcement as institutional repositioning completes.

2. 6 Million Copy Milestone Announcement The timing and velocity of the 6M announcement serves as a real-time update to the deceleration slope estimate. An announcement arriving quickly (before end of April) would strengthen the Base/Bull narrative. A longer wait shifts the probability distribution toward Bear.

3. Content Update Cadence Pearl Abyss’s ability to moderate deceleration depends on the content calendar — major patches, DLC, and potential new platform or regional availability. Each confirmed content event is a quantifiable deceleration-moderator with direct model implications.

Bull and Bear Case

Bull Case: Q1 earnings on May 12 reveal revenue materially above consensus; sell-side models rebuild upward; 6M announcement arrives quickly, validating Base/Bull velocity assumptions; deceleration proves shallower than Bear scenario; DS’s 8M estimate proves conservative; multiple expansion accompanies upward earnings revision.

Bear Case: Q1 revenue recognition disappoints on accounting phasing or cost surprises; deceleration accelerates faster than Bear scenario assumptions; content update calendar stalls; 6M milestone takes longer than expected, confirming a steeper decay curve; broader KOSDAQ multiple compression limits re-rating scope.

Frequently Asked Questions

Q: Is this a buy or sell recommendation? No. This is analytical commentary on market information asymmetry — specifically, the mechanism by which sell-side consensus models lag confirmed factual developments. Nothing in this post constitutes investment advice. All investment decisions involve risks that each investor must assess independently.

Q: What does the 5M milestone actually confirm? The April 15, 2026 official disclosure confirms production-level commercial success for Crimson Desert on a global multi-platform basis. It validates that the title has crossed the threshold where its economic contribution to Pearl Abyss’s 2026 fiscal year is material and largely de-risked from a binary success/failure standpoint. It also confirms that the company’s first major AAA packaged title — a significant strategic pivot from its Black Desert Online live-service model — has achieved industry-relevant scale in its launch window.

Q: Why does sell-side lag buy-side in situations like this? Sell-side research operates on a periodic model update cycle anchored to earnings releases, not continuous news flow. Updating a financial model requires propagating a changed assumption through revenue, cost, and earnings estimates — a process requiring time, validation, and formal publication. Buy-side analysts, operating without publication obligations, can update internal models in real-time. The structural result: sell-side consensus reflects the world as it was before the last major data point, while sophisticated buy-side positioning may already reflect the updated view. This creates the observable “slow stock” phenomenon — the price has not been pulled to a new equilibrium because the consensus anchor has not yet moved.

Q: What is PEAD, and is this related? Post-Earnings Announcement Drift (PEAD) is a well-documented market anomaly in which stocks reporting earnings surprises — particularly positive surprises — continue to drift in the direction of the surprise for weeks to months after announcement. The leading explanation: institutional capital repositioning takes time, and sell-side model updates and target price revisions are episodic rather than instantaneous. Under this framework, if Pearl Abyss’s Q1 2026 earnings on May 12 materially beat sell-side embedded assumptions, the PEAD dynamic could create a prolonged drift as consensus catch-up extends across the following weeks.

Q: Are the 7.5M / 8.5M / 10M scenarios official Pearl Abyss guidance? No. These are independent analytical scenarios constructed for this post, based solely on publicly disclosed unit sales data and observed velocity calculations. They do not represent Pearl Abyss company guidance, brokerage price targets, or any official projection. Pearl Abyss has not issued explicit full-year unit guidance as of the date of this analysis. These scenarios are a structured framework for thinking about the range of plausible outcomes under different deceleration assumptions — not forecasts.

Conclusion: The Gap Is in the Interpretation, Not the Information

The central finding of this analytical framework is simple: Pearl Abyss’s stock is not slow because the market lacks information. It is slow because the market is transitioning its interpretive frame from a binary success/failure question to a continuous deceleration-rate question — and that transition requires confirmed quarterly earnings data, not just milestone announcements, to complete.

The consensus gap is quantifiable. The majority of sell-side full-year 2026 estimates sit at or below 5.26M copies. The confirmed fact as of April 15 is that 5M has already been reached. Even a conservatively pessimistic deceleration scenario (7.5M Bear) produces outcomes well above the current sell-side majority. The forced-recalibration event — Q1 earnings on May 12 — is visible on the calendar.

Alpha, in this type of situation, tends to emerge not from better information but from a more accurate mental model of where the consensus is anchored and where it will be forced to move.

Related reading in this series: Pearl Abyss: Crimson Desert Conquers China — 86% Steam Rating & 4M Copies Sold — franchise IP thesis and China market re-rating.

Disclaimer

This post is market interpretation commentary and does not constitute investment advice, a solicitation, or a recommendation to buy or sell any security. Pearl Abyss (263750.KQ) is referenced solely for analytical and informational purposes. Sell-side estimates cited in this post reflect figures from individual brokerage research reports as of their respective publication dates; these figures may have been updated or revised since publication. Past sales velocity is not indicative of future performance. All investment decisions involve risk, including the risk of total loss. Readers should conduct their own independent research and consult qualified financial advisors before making any investment decision.

Korean Stock Screener Top 5: Apr 16, 2026

Sat, 18 Apr 2026 09:00:00 +0900

Korean Stock Screener Top 5 (Week of Apr 16, 2026): Names Passing Multiple Independent Filters

Each week, our quantitative framework runs ten independent screeners across 459 uniquely identified Korean-listed securities. This report surfaces the five names that cleared the highest number of those screens simultaneously during the period April 9–16, 2026 — and explains why intersection, not single-screen, is the meaningful unit of analysis.

Why Intersection Screening Matters

Most quantitative stock screens are single-thesis tools: a mean-reversion screen finds names that have corrected deeply; a momentum screen finds names in price uptrends; a smart-money flow screen finds names where institutional positioning is building. Any one of these can produce false positives — a stock passes a mean-reversion filter merely because it has fallen, not because a genuine inflection is underway.

Intersection screening applies a different logic. When a security simultaneously clears screeners built on opposing assumptions — for example, both a post-correction rebound model and an earnings-surprise momentum model — the overlap is structurally informative. The two frameworks were not designed to agree; their simultaneous activation suggests the name carries multiple independent tailwinds rather than a single transient factor. In academic terms, the intersection reduces factor-specific noise and amplifies the signal-to-noise ratio of each individual screen.

Our ten screeners target distinct market regimes: Mean Reversion (deep-correction snapbacks), PEAD (Post-Earnings Announcement Drift — price momentum following earnings surprises), Quality Compounder (ROE/ROIC/earnings-growth compounders), Quiet Accumulation (stealth institutional accumulation before broader market awareness), three Smart Money Flow variants, Value Quality (cheap-on-fundamentals with quality filter), and Cycle (sector rotation tied to macro regimes). Because the design assumptions of these screeners conflict, a name appearing in the top tier of five or more is — by construction — not a single-factor artifact.

The five names below represent those with the highest intersection counts from a scan of 459 Korean-listed companies across up to three daily snapshots during the coverage period. They span five distinct regimes (mean reversion, deep value turnaround, breakout momentum, holding-company re-rating, and smart-money biotech), which provides natural regime diversification from a portfolio construction standpoint.

Important framing: Signal strength scores used throughout this report reflect screener intersection counts and factor alignment — they are not buy, sell, or hold ratings. See the FAQ and Disclaimer sections.

#1 — Sepong Electric (세명전기, 017510.KQ)

Signal Strength: Maximum | Screener Intersection Count: 6/10

Screener Intersection Summary

Sepong Electric is the only name in the full 459-company universe that simultaneously cleared both the Mean Reversion and PEAD screeners — frameworks built on mutually exclusive assumptions (one requires a price decline; the other requires an earnings-driven price advance). It also cleared Quality Compounder, two Smart Money variants, and Value Quality, for a total intersection of six.

Financial Metrics

Metric	Value
Current Price	₩10,200
52-Week High Discount	−16.9%
Operating Profit YoY	+393%
Net Income YoY	+172%
Forward PER	12.2×
PBR	1.73×
Operating Margin	45.8%
ROE	15.1%
Foreign Ownership Change (20-day)	+1.56 pp, net buy ₩3.1bn
RSI	58.6
MACD	Positive crossover

Illustrative Regime-Diversification Matrix

Rank	Name (Ticker)	Screener Count	Primary Regime	Forward PER	Foreign Flow (20d)
1	Sepong Electric (017510)	6	Mean Reversion + PEAD	12.2×	+1.56 pp
2	Woowon Development (046940)	5	Quiet Accumulation + Deep Value	1.9×	+0.83 pp
3	SK Hynix (000660)	5	Breakout Momentum	5.0×	+₩2.55tr (5d reversal)
4	SK Square (402340)	7	All-Regime (Holding Re-rating)	4.0×	−₩223.6bn
5	Pharmicell (005690)	4	Smart Money Flow	26.7×	+2.69 pp

Notable Exclusions

Excluded	Primary Reason
Ubiqus (264450)	Foreign ownership −2.08 pp over 20 days — adverse fund flow
Microcontact Sol (098120)	+100.8% over 60 days + RSI 67.6 + Bollinger upper breach = overheated
April (278470)	PBR 34× — premium too extreme despite strong fundamentals
Company K (307930)	ROE 9.2% + foreign ownership 1.36% — quality and liquidity below threshold

FAQ

Q1: What is intersection screening? Intersection screening is a quantitative methodology that identifies securities simultaneously in the top-ranked cohort of multiple independent screens. Rather than selecting stocks that pass any single filter, it requires passage of several filters designed under different assumptions. The overlap is treated as higher-confidence evidence of multi-factor alignment.

Q2: Why run 10 screeners rather than one comprehensive model? A single composite model can mask factor cancellation — a high score on one factor might compensate for a low score on another. Running 10 separate models that each optimize for a distinct regime forces the stock to demonstrate independent strength across multiple dimensions. The intersection is a form of ensemble validation.

Q3: What does PEAD mean? PEAD stands for Post-Earnings Announcement Drift. It is a documented market anomaly — identified in academic literature (Ball & Brown, 1968; Bernard & Thomas, 1989) — in which stock prices continue to drift in the direction of an earnings surprise for weeks or months after the announcement date, rather than adjusting fully on announcement. A PEAD screen identifies stocks that recently reported an earnings surprise and whose price has not yet fully incorporated the information.

Q4: How do you validate that screener intersections are structural signals rather than coincidental factor overlap? This is the central methodological question. Our current validation approach: (a) check factor correlation among intersecting screeners to ensure they are not measuring identical variables (QC, VQ, and SM-Q all share ROE sensitivity — users should treat those three as a partial cluster, not three fully independent votes); (b) examine out-of-sample hit rates for prior intersection cohorts; (c) analyze 12-month forward return distributions for intersection sizes of 4+, 5+, 6+, and 7+. A full backtest on the MR+PEAD simultaneous configuration (as seen in Sepong Electric this week) is planned.

Q5: Is this investment advice? No — educational analysis only. This report does not constitute investment advice, a solicitation to buy or sell any security, or a recommendation to take any financial action. All signal strength scores, scenario reference levels, and technical observations are descriptive and informational in nature.

Conclusion: Why Intersection Is the Meaningful Unit of Analysis

Single-screen investing carries a specific failure mode: every screen has a false positive rate, and in a universe of hundreds of names, even a 5% false positive rate generates dozens of misleading candidates. The solution is not to build a better single screen — it is to require that a candidate survive multiple independent stress tests simultaneously.

When Sepong Electric clears both a mean-reversion screen (requiring a price decline) and a PEAD screen (requiring earnings-surprise momentum), it is not a coincidence. It means the stock passed a filter expecting price weakness and a filter expecting price strength simultaneously — because it offers both: a reset valuation and an accelerating earnings base. That is a more durable thesis than either signal alone.

The five names profiled this week span five distinct investment regimes with low mutual correlation. This diversification is itself a product of the methodology: when you screen for multi-factor validation across regime-diverse screens, you naturally surface a regime-diverse output set. The open questions — whether intersecting screens are truly independent, and how intersection count correlates with forward returns in Korean market conditions — remain items for ongoing backtesting, and we flag them explicitly as open agenda items alongside every weekly publication.

Disclaimer

This report is produced for educational and informational purposes only. It does not constitute investment advice, a solicitation, an offer to buy or sell any security, or a recommendation to take any financial action. The information herein is based on publicly available data and quantitative screener outputs; it may contain errors or omissions, and no representation is made as to its accuracy or completeness.

Individual investors should conduct their own independent research and consult qualified financial advisors before making any investment decision. Every investor’s risk profile, portfolio context, tax situation, investment horizon, and regulatory environment is different.

Liquidity risk notice: Sepong Electric (017510.KQ) and Woowon Development (046940.KQ) are small-to-mid-cap KOSDAQ securities with materially lower daily trading liquidity than KOSPI mainboard large-caps. Investors in illiquid securities face wider bid-ask spreads, greater price impact from position changes, and potentially impaired ability to exit at desired prices.

Past screener performance and historical factor returns do not guarantee future results. Korean equity markets are subject to regulatory, currency, geopolitical, and macroeconomic risks that may not be fully captured by any quantitative framework.

Published: April 18, 2026 | Coverage period: April 9–16, 2026 | Universe: 459 Korean-listed securities | Screeners: 10

Korean Stocks: 8 Names That Beat 5+ Screeners

Thu, 16 Apr 2026 09:00:00 +0900

Why “Screener Intersection” Is a Different Kind of Signal

Most quantitative screens answer one question at a time: Is this stock cheap? Is momentum turning? Are insiders accumulating? The answers are useful in isolation, but individually noisy. A stock can look cheap because it is genuinely overlooked — or because earnings are about to collapse. A momentum breakout can reflect real fundamental improvement or a one-day squeeze with no follow-through.

The intersection approach flips the logic. Instead of asking whether a stock passes a filter, it asks how many independent filters — each designed around a different market regime and a different underlying assumption — return the same name. A stock that surfaces in a mean-reversion model, a post-earnings-drift model, a quality-compounder model, and a quiet-accumulation model at the same time is not just cheap, or just revising, or just being accumulated. It is all of those things simultaneously. That convergence across uncorrelated frameworks is harder to explain away.

This report covers the week of April 9–16, 2026. We ran 10 independent screeners across 459 unique Korean-listed names over up to three daily snapshots each, totaling over 1,300 data points. The 15 names in the leaderboard below appeared in the most screeners with the highest combined hit counts. The top 5 are analyzed in depth.

The Ten Screeners at a Glance

Screener	What It Targets
Mean Reversion (MR)	Deep-corrected stocks showing early reversal conditions
PEAD	Price momentum following a positive earnings surprise
Quality Compounder (QC)	Stocks combining strong ROE, ROIC, and earnings growth
Quiet Accumulation (QA)	Gradual institutional or foreign buying with below-average volume
Smart Money — Earnings (SM-E)	Flow signal anchored to earnings quality
Smart Money — Volatility (SM-V)	Flow signal anchored to risk-adjusted returns
Smart Money — Quality (SM-Q)	Flow signal anchored to balance sheet quality
Value Quality (VQ)	Low-valuation stocks that also clear a quality hurdle
Cycle	Stocks at or near cyclical inflection from trough to expansion

All ten screeners ran without errors for the review period. The Smart Money files timestamped April 13 (16:08:44, 16:08:55, and 06:06) used a slightly different naming convention but produced valid output. The Mean Reversion snapshot updated live in the early hours of April 16, completing the dataset.

Cross-Screener Leaderboard — Top 15 (Week of Apr 16, 2026)

Rank	Ticker	Name	Screeners	Total Hits	Combination
1	402340	SK Square	7	18	Cycle + PEAD + QC + SM×3 + VQ
2	264450	Ubiqus	6	11	Cycle + PEAD + SM×3 + VQ
3	017510	Semyung Electric	6	6	MR + PEAD + QC + SM×2 + VQ
4	000660	SK Hynix	5	13	Cycle + PEAD + QC + SM×2
5	098120	Microcontact Solutions	5	11	PEAD + QC + QA + SM×2
6	307930	Company K	5	9	Cycle + PEAD + QC + SM×2
7	046940	Woowon Development	5	9	PEAD + QA + SM×3
8	330730	Stonebridge Ventures	5	7	PEAD + QA + SM×3
9	005690	Pharmicell	4	9	PEAD + QC + SM×2
10	278470	APR	4	9	PEAD + QC + SM×2
11	078070	Ubiqus Holdings	4	8	Cycle + PEAD + SM×2
12	021080	Atinum Investment	4	8	MR + PEAD + SM×2
13	101490	S&S Tech	4	6	PEAD + QC + SM×2
14	005440	Hyundai GF Holdings	4	4	Cycle + QA + SM + VQ
15	425420	TFE	3	8	PEAD + QC + SM

MR = Mean Reversion · QA = Quiet Accumulation · QC = Quality Compounder · VQ = Value Quality · SM = Smart Money variant

Top 5 Deep Dives

1. SK Square (KOSPI: 402340) — 7 Screeners, 18 Hits

SK Square is the listed investment holding company spun out of SK Telecom in 2021. Its primary asset is a roughly 20% stake in SK Hynix, making its financials unusually sensitive to semiconductor cycles.

Fundamentals (FY2025, consolidated):

Metric	Value
Revenue	₩10.5 trillion
Operating Profit	₩8.8 trillion
Net Income	₩8.8 trillion
Revenue YoY	+77.8%
Operating Profit YoY	+124%
Net Income YoY	+142%
ROE	37.8%
Operating Margin	84.1%
Trailing PER	10.0×
Forward PER	4.0×
PBR	3.17×
VQ Score	0.79
Quality Score	0.96
Cycle Score	0.96

The 84% operating margin reflects the holding-company structure: most reported profit is mark-to-market or dividend income from subsidiaries rather than operational cash flow. The forward PER of 4.0× against a Quality Score of 0.96 is the combination that triggered six additional screeners beyond Cycle.

Price action: At ₩690,000, the stock is at a 52-week high, up +54.5% over 60 days. RSI stands at 66.6 and price has broken above the upper Bollinger Band (at 103% of the band). These are early-stage overheating conditions.

Flow breakdown:

5-day foreign net buy: +₩1.4 billion; institutional net buy: +₩181.5 billion
20-day foreign net buy: −₩223.6 billion (profit-taking); institutional net buy: +₩130.8 billion (offsetting)
Foreign ownership change: −0.26 percentage points over 20 days

The divergence is notable: institutions are accumulating aggressively while foreigners who entered earlier are trimming. This is a rotation pattern, not a clean accumulation setup. The 7-screener score is the strongest in the dataset, but the near-term technical picture suggests a 1–2 week consolidation is more likely than immediate continuation.

2. Semyung Electric (KOSPI: 017510) — 6 Screeners, MR + PEAD Dual Signal

Semyung Electric manufactures distribution transformers and power grid equipment. It is a direct beneficiary of Korea’s grid modernization investment cycle and the broader global power-infrastructure buildout driven by data-center electricity demand.

Fundamentals (FY2025):

Metric	Value
Revenue YoY	+124%
Operating Profit YoY	+393%
Operating Margin	45.8%
ROE	15.1%
Trailing PER	12.2×
PBR	1.73×

An operating margin of 45.8% in a manufacturing business is unusually high and suggests significant operating leverage materializing as revenue scales rapidly. A PER of 12.2× against +393% operating profit growth implies the market has not yet priced the earnings step-change.

Technical setup:

Price: ₩10,200 — trading above the 5-day, 20-day, and 60-day moving averages
Distance from 52-week high: −16.9% (deep correction completed)
RSI: 58.6 (neutral-to-recovering)
MACD: positive crossover (early-stage reversal confirmation)
Price has touched the upper Bollinger Band: early breakout signal

Flow:

20-day foreign net buy: +₩3.1 billion
Foreign ownership change: +1.56 percentage points over 20 days

This is the only stock in the entire 459-name universe that simultaneously cleared both Mean Reversion (price corrected more than 16% from peak, technical reversal beginning) and PEAD (strong post-earnings drift conditions following the +393% operating profit surprise). The combination is significant: MR signals that the psychological and price reset has been completed, while PEAD signals that the fundamental revision cycle is still early. Foreign investors adding 1.56 percentage points of ownership during the correction adds a third independent confirmation layer.

3. SK Hynix (KOSPI: 000660) — 5 Screeners, 13 Hits

SK Hynix is Korea’s largest memory semiconductor company and the world’s second-largest DRAM producer. It is a primary supplier of High-Bandwidth Memory (HBM) to AI accelerator manufacturers.

Fundamentals (FY2025):

Metric	Value
Revenue YoY	+46.8%
Operating Profit YoY	+101%
ROE	44.1%
Operating Margin	48.6%
Forward PER	5.0×

Price action: At ₩1,136,000, the stock is at a 52-week high, up +51.7% over 60 days.

Flow — the key development this week:

Horizon	Foreign	Institutional
5-day	+₩250 billion	n/a
10-day	+₩290 billion	(continued accumulation)
20-day	−₩2.2 trillion	+₩1.95 trillion

The 20-day foreign figure shows heavy profit-taking, but the 5-day and 10-day data show that foreign investors have reversed to net buying within the past two weeks. Combined with institutional accumulation of nearly ₩2 trillion over 20 days, the flow picture has shifted. Foreign ownership sits at 53.08%.

A Forward PER of 5.0× for a company generating 48.6% operating margins with AI-driven HBM demand continuing to expand is the central investment thesis captured by the Cycle, PEAD, and Quality Compounder screeners.

4. Microcontact Solutions (KOSDAQ: 098120) — 5 Screeners including Quiet Accumulation

Microcontact Solutions manufactures semiconductor test contactors (socket hardware used in IC test handlers). It is a niche supplier in the back-end semiconductor equipment chain.

Fundamentals (FY2025):

Metric	Value
Revenue YoY	+44%
Operating Profit YoY	+73%
Net Income YoY	+59%
ROE	23.1%
Operating Margin	17.7%
Trailing PER	17.0×
PBR	3.54×

Price action: ₩34,150 — within 1.3% of the 52-week high, up +100.8% over 60 days. RSI at 67.6 and Bollinger Band position at 102% indicate the stock is in overheating territory.

Flow: 20-day foreign net buy +₩3.7 billion; foreign ownership +1.47 percentage points — consistent, below-radar accumulation across the period.

The interest here is methodological: it is unusual for a stock to clear Quiet Accumulation (steady buying, low volume spikes), PEAD (strong post-earnings drift), Smart Money Earnings and Volatility variants, and Quality Compounder in the same week. This means every regime-type the system monitors — contrarian, momentum, fundamental, and technical — flagged the same name. The practical constraint is the 60-day return of +100.8%, which substantially raises the risk of entering at current prices.

5. Woowon Development (KOSPI: 046940) — 5 Screeners, Extreme Turnaround

Woowon Development is a Korean civil engineering and construction company focused on infrastructure and residential projects.

Fundamentals (FY2025):

Metric	Value
Revenue YoY	+24%
Operating Profit YoY	+2,300%
Net Income YoY	+6,757%
ROE	36.8%
Operating Margin	17.9%
Debt Ratio	68.5%
Trailing PER	1.9×
PBR	0.58×

Price: ₩5,060, within 2.5% of 52-week high. RSI at 61.0, MACD positive.

Flow: Foreign ownership at 9.37%, with +0.83 percentage points added over 20 days.

The operating profit expanding 23-fold and net income expanding 68-fold on a 24% revenue increase implies the company passed a profitability inflection point — likely the resolution of project cost overruns or completion of low-margin legacy contracts. At PER 1.9× and PBR 0.58×, the valuation has not yet adjusted. The Quiet Accumulation flag suggests this re-rating is happening slowly and below institutional radar. The construction sector is inherently cyclical and carries execution risk, which explains the depressed multiples even after the earnings surge.

Investment Implications: Why Intersection Screening Matters

The aggregate signal from this week’s data points to three structural themes in Korean equities:

1. Power infrastructure is re-rating. Semyung Electric’s appearance across six screeners — including the rare MR+PEAD dual signal — reflects a broader market recognition that Korea’s grid equipment manufacturers are operating in a multi-year demand upcycle driven by AI data center power requirements and government grid modernization mandates.

2. The semiconductor cycle has entered institutional accumulation mode. Both SK Hynix and SK Square cleared multiple screeners, and institutional flow data shows sustained buying measured in the hundreds of billions to trillions of won. The Forward PER profiles (5× for Hynix, 4× for SK Square) suggest the market is pricing a earnings peak scenario that the underlying AI infrastructure demand trajectory may not support.

3. Micro-cap turnarounds with foreign accumulation are the highest-dispersion opportunity. Woowon Development (PER 1.9×, PBR 0.58×, OP +2,300%) and Semyung Electric (PER 12.2×, OP +393%) represent situations where the valuation has not yet caught up with an earnings step-change that has already occurred. Foreign investors quietly adding ownership in both names — before a re-rating — is historically the pattern that precedes meaningful price discovery.

A methodological caution: QC, VQ, and SM-Quality all weight ROE heavily. Stocks like SK Square that appear in all three may be generating three “independent” flags that actually reflect the same underlying factor. Genuine independence requires regime diversity — the MR+PEAD combination in Semyung Electric, or the QA+PEAD+SM combination in Woowon Development, are structurally more independent because they draw from different data sources (price history, earnings revisions, ownership changes).

Bull and Bear Considerations by Name

Ticker	Bull Case	Bear Case
017510 Semyung Electric	Grid capex super-cycle sustains 40%+ margins; PER expands to 20–25× on visibility	Project concentration risk; margin normalization as competition enters
046940 Woowon Development	PER/PBR normalization alone implies 3–5× upside; foreign accumulation accelerating	Construction cycle turns; high debt ratio amplifies downside; earnings quality uncertain
000660 SK Hynix	HBM pricing power + AI server buildout sustains HBM3E/HBM4 margin premium	DRAM commodity oversupply returns; US export controls tighten on advanced memory
402340 SK Square	Holding discount narrows; SK Hynix re-rating flows through; buybacks	Holding discount widens; Hynix correction disproportionately hurts NAV
098120 Microcontact	Backend semiconductor intensity rises with AI chip complexity; niche pricing power	60-day +100% has pulled forward much of the re-rating; customer concentration

FAQ

What is Mean Reversion screening in the context of Korean stocks? Mean Reversion (MR) screening identifies stocks that have declined significantly from recent peaks — typically 15–25% — and show early technical signs of stabilization or reversal. The logic is that short-term price moves often overshoot fair value, creating asymmetric entry opportunities. In this week’s scan, Semyung Electric’s −16.9% drawdown from its 52-week high, combined with MACD turning positive and price recovering above key moving averages, met the MR criteria.

Why run 10 screeners instead of one comprehensive model? Different market regimes reward different factor combinations. A single model optimized for mean reversion will systematically miss earnings-momentum opportunities, and vice versa. Running 10 independent models — each designed for a specific regime — and then finding names that appear across multiple models reduces the probability that any single name’s appearance is a false positive driven by data quirks or factor-specific noise. The intersection acts as a natural filter.

What does “Quiet Accumulation” mean, and why is it significant? Quiet Accumulation (QA) flags stocks where foreign or institutional ownership is rising steadily but total trading volume remains below average — suggesting informed buyers are building positions without moving the price. It is significant because it implies the re-rating has not yet been publicly recognized or broadly discussed, which historically precedes price discovery phases. Woowon Development, Microcontact Solutions, and Stonebridge Ventures all cleared QA this week.

How should international investors interpret Korean ownership data? Korean exchanges (KOSPI and KOSDAQ) require daily disclosure of foreign and institutional ownership changes. “Foreign” typically means non-Korean investors including global funds, hedge funds, and asset managers. “Institutional” means domestic Korean funds, insurance companies, and brokerages. When both are accumulating simultaneously — as with SK Hynix over 5–10 days — it suggests broad conviction across investor types. When they diverge — as with SK Square, where foreigners are selling and institutions are buying — it signals a rotation between investor groups rather than net new demand.

What is PEAD, and how is it applied here? Post-Earnings Announcement Drift (PEAD) is a well-documented market anomaly: stocks that report earnings significantly above consensus tend to continue drifting upward for weeks to months following the announcement, as the full magnitude of the earnings surprise is gradually absorbed by the market. The PEAD screener identifies stocks where this drift condition is active — typically within 30–60 days of a large positive earnings revision. Of the 459 names scanned, PEAD appeared in the screener combinations of 11 of the top 15.

Disclaimer

This report is produced for informational and educational purposes only. It does not constitute investment advice, a solicitation to buy or sell any security, or a recommendation of any specific investment strategy. All data referenced — including financial metrics, price levels, ownership percentages, and screener scores — reflects the analysis period of April 9–16, 2026, and may have changed materially since publication. Korean-listed equities involve market, currency, liquidity, and regulatory risks that may differ significantly from investors’ home markets. Past screener performance and historical data patterns do not guarantee future results. Readers should conduct their own due diligence and consult a licensed financial advisor before making any investment decisions. The author and publisher hold no positions in the securities discussed at the time of publication.