Based on Yi Fan (弈樊), Gold Stock Game, 4th Edition (金股博弈, 第四版) (2020)
Yi Fan approaches the Chinese stock market not as an exercise in valuation or technical analysis, but as a multi-player strategic game where understanding the motivations, constraints, and probable actions of other players is the primary edge. The 4th edition of Gold Stock Game reflects two decades of refinement, incorporating lessons from the 2015 crash, the rise of quantitative trading, and the increasing influence of foreign capital through Stock Connect.
The core argument rests on three pillars:
The stock market is a game of incomplete information. Unlike chess, where all pieces are visible, the stock market hides crucial information — who is buying, why they are buying, and what they plan to do next. Winning requires probabilistic reasoning about hidden information, not certainty.
There are structurally different player types. Retail investors, institutional funds, and market makers (主力/庄家) have fundamentally different information sets, capital sizes, time horizons, and constraints. Each player type has predictable behavior patterns that can be exploited by those who study them.
Strategy must account for other players' strategies. The price you see is not a reflection of "intrinsic value" — it is the equilibrium of competing strategies. To profit, you must understand not just what a stock is worth, but what other players believe, what they will do next, and how your actions interact with theirs.
Yi Fan writes from the perspective of a strategist who has observed the A-share market's unique game dynamics — a market where retail investors constitute 80% of trading volume, where major shareholders and market makers regularly manipulate price action, and where information asymmetry is extreme. The book provides a practical framework for the retail investor to survive and profit within this adversarial environment.
The operational goal: identify the game being played in any given stock, determine which player type is driving the action, anticipate their next move, and position accordingly — or stay out entirely when the game is unreadable.
Yi Fan frames the stock market as a non-cooperative game — players act in their own self-interest, and one player's gain often comes at another player's expense. This is distinct from the cooperative framing of "we're all in this together" that market cheerleaders promote.
Key game theory concepts applied:
The fundamental asymmetry of the A-share game:
| Player Type | Information Advantage | Information Disadvantage |
|---|---|---|
| Market Maker (主力) | Knows own position and plan; may have undisclosed corporate information; sees order flow | Must move large positions without alerting the market |
| Institutional Fund | Research team, management access, industry network | Constrained by mandates, compliance, reporting requirements |
| Foreign Capital (QFII/Connect) | Global perspective, fundamental analysis depth | Less understanding of local policy nuances, A-share sentiment |
| Retail Investor | Agility, no reporting requirements, can hold cash indefinitely | No information edge, emotional decision-making, small capital base |
Yi Fan constructs a simplified payoff matrix for the retail-vs-market-maker game:
| Market Maker Accumulating | Market Maker Distributing | |
|---|---|---|
| Retail Buys | Retail profits (aligned with MM) | Retail loses (absorbing MM's sell) |
| Retail Sells | Retail loses (gives up position too early) | Retail profits (avoids distribution) |
The core challenge: retail cannot directly observe whether the market maker is accumulating or distributing. The entire book is essentially about solving this inference problem.
Yi Fan provides a detailed profile of the market maker — the most powerful and most misunderstood player in the A-share game:
Capital: Typically 500 million to 5 billion yuan for a single stock operation. Must acquire 20-40% of the tradable float to control price action.
Time Horizon: A complete market maker cycle (accumulate → mark up → distribute) typically lasts 6-18 months. Patience is their greatest weapon.
Operational Phases:
Accumulation (吸筹): Quietly buy shares over weeks or months, suppressing price to avoid attracting attention. Techniques include: placing large sell orders to create the appearance of selling pressure while buying through multiple accounts; shaking out weak holders with sharp intraday drops; spreading negative rumors to depress sentiment.
Test (试盘): After accumulating a sufficient position, briefly push the price up 5-10% to test selling pressure. If heavy selling appears, return to accumulation. If selling is light (floating supply is absorbed), prepare for markup.
Markup (拉升): Aggressively push the price higher through sustained buying, often using limit-up boards to create momentum and attract retail followers. This is the phase where the market maker wants attention.
Distribution (出货): Sell the accumulated position to retail investors who are now chasing the stock's momentum. Techniques include: creating artificial consolidation patterns that look like "rest before the next leg up"; placing large buy orders that are cancelled before execution (spoofing); coordinating positive media coverage to maintain bullish sentiment.
Decline (下跌): Once distribution is complete, the market maker withdraws support. Without the buyer of last resort, the price collapses under its own weight as trapped retail investors sell at ever-lower prices.
Institutional players in A-shares operate under constraints that create predictable behavior:
Yi Fan's analysis of retail behavior is unflinching:
Yi Fan maps information flow through the market as a gradient from most to least informed:
Company insiders → Market maker → Connected institutions →
Sell-side analysts → Mutual funds → Financial media →
Stock forums → General retail
By the time information reaches the average retail investor, it has already been priced in by everyone above them in the gradient. This is why "buying on news" is consistently unprofitable for retail — you are the last to know.
Fundamental asymmetry: The market maker may know about upcoming contracts, earnings surprises, or strategic changes before public disclosure. While illegal, this remains prevalent in A-shares.
Positional asymmetry: The market maker knows their own position size, average cost, and exit plan. Retail investors can only guess.
Intentional asymmetry: The market maker can create false signals — volume spikes, price patterns, news leaks — designed to mislead retail investors about the market maker's true intentions.
Structural asymmetry: Institutional investors have Bloomberg terminals, research teams, management meetings, and channel checks. Retail investors have a brokerage app and Baidu.
Yi Fan offers practical methods for retail investors to narrow the information gap:
Track the Dragon and Tiger List (龙虎榜). The exchange publishes the top 5 buying and selling brokerage seats for stocks that hit unusual price or volume thresholds. Recurring appearances of the same institutional seats indicate coordinated accumulation or distribution.
Analyze the shareholder registry. Quarterly reports disclose the top 10 shareholders. Tracking changes — new institutional entries, increasing positions, exits — reveals the positioning of informed players.
Monitor Block Trade Data (大宗交易). After-hours block trades at discounts to market price may indicate major shareholder selling or institutional repositioning.
Volume at Price Analysis. By analyzing volume distribution across price levels, estimate where the market maker's average cost likely sits. The market maker will not allow the stock to trade sustainably below their cost.
Margin Trading Data. The exchange publishes daily margin buying and short selling volumes for eligible stocks. Rising margin buying into an advancing stock signals leveraged retail participation — a late-stage indicator.
Yi Fan identifies the telltale signs of market maker accumulation:
The mirror image of accumulation:
Yi Fan describes the markup phase as the only period where retail and market maker interests align. Identifying the transition from accumulation to markup is the single most profitable skill:
Markup confirmation signals:
Risk during markup:
Yi Fan models institutional herding as a self-reinforcing loop:
Implication for retail: Stocks with extremely concentrated institutional ownership (top 10 holders = 70%+ of float) have high potential downside when institutional consensus shifts.
Retail investors can exploit institutional constraints that do not bind them:
Yi Fan's core recommendation for retail investors is to adopt an asymmetric strategy — accept that you are the weakest player in the game and design your approach accordingly:
Yi Fan provides a game-theory-aware stock selection process:
Step 1: Exclude unplayable games.
Step 2: Identify institutional interest.
Step 3: Detect accumulation phase.
Step 4: Wait for confirmation.
Before every trade, Yi Fan requires answering three questions:
Who is on the other side? If you are buying, someone is selling. Why? Is it a market maker distributing? An institution rebalancing? A forced seller? If you cannot identify a plausible reason the seller is wrong, do not buy.
What is the market maker's probable phase? Accumulation, test, markup, or distribution? Only enter during late accumulation or early markup. Entering during distribution is the retail investor's most common fatal error.
What is my exit condition? Define the exit before entry. The exit should be based on the game theory framework — exit when the market maker's behavior shifts from markup to distribution signals, not based on arbitrary price targets or stop-loss percentages.
Yi Fan adapts Howard Marks' concept of second-level thinking to the A-share context:
In the A-share game, second-level thinking extends further:
Yi Fan advocates a minimax strategy — minimize the maximum possible loss before attempting to maximize gain:
For important trading decisions, Yi Fan recommends constructing a simple game tree:
Current situation: Stock X in apparent accumulation, approaching breakout level
Branch A: Breakout is real (probability: 40%)
→ Markup phase: expected gain 30-50%
→ Expected value: 0.40 × 0.40 = +16%
Branch B: False breakout, return to range (probability: 35%)
→ Expected loss: 5-10%
→ Expected value: 0.35 × (-0.075) = -2.6%
Branch C: False breakout, new low (probability: 25%)
→ Expected loss: 15-25%
→ Expected value: 0.25 × (-0.20) = -5.0%
Total expected value: +16% - 2.6% - 5.0% = +8.4%
Decision: Positive expected value, but with significant downside risk.
Action: Take a partial position (50% of target) on breakout.
Add remaining 50% only if markup confirmation holds for 3 days.
Yi Fan classifies volume patterns by their game theory implications:
| Pattern | Description | Probable Game State |
|---|---|---|
| Low volume, flat price | Minimal trading activity | Quiet period — no major player acting |
| Rising volume, flat price | Increasing activity without price progress | Accumulation or distribution in progress |
| Volume spike on up day | Abnormally high volume on a rising day | Either breakout (bullish) or distribution into strength (bearish) — requires context |
| Volume spike on down day | Abnormally high volume on a falling day | Either panic capitulation (near bottom) or market maker exiting (more downside) |
| Volume declining during uptrend | Each advance has less volume support | Buyers exhausting — markup may be ending |
| Volume declining during downtrend | Selling pressure diminishing | Sellers exhausting — bottom may be forming |
In A-shares, the level-2 order book (买卖五档) provides rich game information:
Yi Fan identifies game-relevant time-of-day patterns specific to A-shares:
Yi Fan adapts the Kelly Criterion for position sizing but applies a significant safety factor given the information uncertainty in A-shares:
Kelly fraction = (Win probability × Average win) - (Loss probability × Average loss)
÷ Average win
Adjusted fraction = Kelly fraction × 0.3 (safety factor for A-share uncertainty)
Example:
Kelly = (0.55 × 0.25 - 0.45 × 0.12) / 0.25 = (0.1375 - 0.054) / 0.25 = 0.334
Adjusted = 0.334 × 0.3 = 10% of portfolio per position
Yi Fan advocates pyramid entry — a game-theoretic approach to position building:
Never invert the pyramid. Putting the largest position at the first entry (before confirmation) is the retail investor's most common sizing error.
Yi Fan catalogs the most common traps set by market makers for retail investors:
Trap 1: The Bull Trap (多头陷阱) The price breaks above a significant resistance level on apparently strong volume, attracting breakout buyers. Within 1-3 days, the price reverses and falls below the breakout level. Retail investors who bought the breakout are now trapped with an immediate loss, and many will hold hoping for recovery — allowing the market maker to continue accumulating at lower prices.
Detection: Check if the volume on the breakout day is dominated by small retail orders or by large institutional blocks. Small-order-dominated volume on a breakout suggests retail chasing, not institutional conviction.
Trap 2: The Bear Trap (空头陷阱) The price breaks below a significant support level, triggering stop-losses and panic selling. Within 1-3 days, the price recovers above the breakdown level. Retail investors who sold at the bottom have provided the market maker with cheap shares.
Detection: Watch the Dragon and Tiger list. If institutional buyer seats appear on the breakdown day, it is likely a bear trap.
Trap 3: The Dividend Trap (分红陷阱) A stock with an attractive dividend yield draws income-focused retail investors. However, in A-shares, the stock price is adjusted downward on the ex-dividend date by the dividend amount. If the stock does not fill the gap, the investor has merely received their own capital back minus taxes. Market makers may pump the stock before the ex-dividend date and distribute to dividend seekers.
Trap 4: The Restructuring Rumor Trap (重组传闻陷阱) Rumors of a corporate restructuring or asset injection circulate, often through stock forums and WeChat groups. Retail investors buy hoping for a transformative deal. The rumor may be planted by the market maker to facilitate distribution. Even if true, the deal may fall through after the market maker has exited.
Trap 5: The "National Team" Trap (国家队陷阱) During market crashes, reports that "the National Team is buying" create a false floor. Retail investors buy believing the government has guaranteed the price. In reality, the National Team's buying is temporary and concentrated in index heavyweights — they are not protecting every stock. Small-caps may continue declining even as large-caps stabilize.
Yi Fan walks through a complete game cycle using a composite example based on real A-share patterns:
Setup: Stock "Z" — mid-cap manufacturing company, 8 billion yuan market cap.
Month 1-3: Accumulation Phase
Month 4: Test Phase
Month 5: Markup Initiation
Month 5-6: Markup Phase
Month 7: Distribution Signals Appear
Month 7-8: Distribution Continues
Month 8: Game Over Signal
Result:
"The stock market is not a game between you and a stock. It is a game between you and every other player who is also trading that stock. The moment you forget that other players exist, you have already lost."
"The market maker's greatest weapon is not capital — it is patience. They will accumulate for months while retail investors, bored by sideways action, move on to the next exciting chart. Then, when retail has forgotten the stock entirely, the markup begins."
"Information in the A-share market flows downhill. By the time a tip reaches your WeChat group, it has already passed through the hands of insiders, market makers, connected institutions, and their friends. You are not the smart money in that trade — you are the exit liquidity."
"The correct question is never 'should I buy this stock?' The correct question is 'who is selling this stock to me, and why?' If you cannot answer the second question, you have no business answering the first."
"Retail investors lose not because they are stupid, but because they play a game designed for them to lose. The A-share market is structured so that information, capital, and rules all favor the larger player. Survival requires acknowledging this asymmetry and designing a strategy that accounts for it."
"The best trade is no trade. A retail investor who sits in cash for six months waiting for a clear setup will outperform the retail investor who trades every day for the same six months. Patience is not passive — it is the most aggressive form of edge management."
"Do not confuse a rising stock with a winning game. A stock rising during distribution is a trap. A stock falling during accumulation is an opportunity. Price direction without game context is meaningless information."
"The pyramid position is the retail investor's shield. Small entry when uncertain, larger addition when confirmed, final addition when validated. Never invert this — the largest position at the point of greatest uncertainty is the fastest path to ruin."
"Every market maker operation has an Achilles heel: they must eventually sell what they accumulated. This selling phase produces detectable signals — high volume without price progress, positive news with fading reaction, institutional seller seats on the Dragon and Tiger list. Learn to read these signals and you learn to exit before the collapse."
"Game theory teaches that in a game of incomplete information, the player who reduces their information disadvantage the most wins. You cannot become an insider. But you can study the footprints that insiders leave — in volume, in shareholder registries, in block trades, in order flow. The footprints are there for those willing to look."