position_sizing_decision_skill

Position Sizing Decision Skill

Data dependencies and execution prerequisites

It is recommended to use Wind wind-mcp-skill to obtain underlying data. Before execution, first check whether wind-mcp-skillis already available locally; if not, prompt the user to install:

# GitHub
npx skills add Wind-Information-Co-Ltd/wind-skills --skill wind-mcp-skill -g -y

# Gitee 镜像（国内）
npx skills add https://gitee.com/wind_info/wind-skills.git --skill wind-mcp-skill -g -y

Skill positioning

This skill is used to solve the problem of "the direction is roughly right, but exactly how large should the position be?" The output focus is not to encourage expressing views through heavy positions, but to connect risk budget, price volatility, planned stop-loss distance, and portfolio capacity so that position size becomes a result of discipline rather than emotion.

Function overview

The core function of this skill is: to provide suitable position sizing and staged entry recommendations based on risk budget.

It emphasizes first quantifying the tolerable loss, then back-calculating the scale of capital investment, helping users avoid disrupting their overall rhythm because a single trade's position is too large.

Use cases

• The user already has a trading direction but does not know how much position should be taken for a single trade.
• Multiple candidate trades exist at the same time, and a decision is needed on which should receive a large position and which should only receive a trial position.
• After market volatility increases, the user wants to recalibrate the risk exposure of a single trade.
• The team wants to unify position-sizing language and reduce "all-in" behavior and arbitrary add-ons.

Working principles

1. Position size must be back-calculated from tolerable risk, not forward-derived from subjective confidence.
2. A single-trade position decision must simultaneously consider the instrument's volatility, stop-loss distance, liquidity, and portfolio correlation.
3. The higher the uncertainty, the more restrained the initial position should be, waiting for subsequent validation before deciding whether to scale up.
4. Position recommendations must be tied to a specific plan; a position number detached from the plan is meaningless.
5. The main text does not describe internal implementation methods, and only outputs the basis for judgment and recommendations.

Execution process

Step 1: Confirm account constraints and risk budget criteria

First clarify what constraints this position recommendation is based on:

• Whether it centers on the maximum tolerable loss for a single trade or on total position exposure
• Whether the user has a conservative, balanced, or aggressive risk preference
• How many trial-and-error attempts the plan allows
• Whether there are industry, theme, or single-stock caps

If the user does not specify, the default basic criterion is "control risk for a single trade first, then consider scaling up returns."

Step 2: Identify trade type and uncertainty level

Different trade types should not use the same position template:

• Trend-confirmation trades can accept a relatively fuller position expression
• Event-driven trades need more attention to gap and distortion risk
• Bottom-fishing, reversal, and left-side trial-and-error trades should significantly reduce the initial position weight
• Instruments with insufficient liquidity or excessive volatility should receive an additional discount

Step 3: Measure volatility and stop-loss distance

Position recommendations must be built on a distance that is "affordable to lose":

• Whether current price volatility is within its normal range
• If the planned stop-loss level is too far, whether the position should be reduced accordingly
• If the planned stop-loss level is too close, whether it is easily shaken out by noise
• Whether a more diversified entry pace should be adopted to reduce sensitivity to entry points

Step 4: Evaluate portfolio capacity and correlation

A single position cannot be isolated from the existing portfolio environment:

• Whether the current portfolio is already tilted toward the same sector or same style
• Whether the new trade will significantly increase correlation clustering
• Whether the current total position is already high, causing the fault tolerance for new trades to decline
• If there are already positions with similar logic, whether the new position should be treated as “supplementary exposure” rather than an independent position

Step 5: Provide the first position and the subsequent phased framework

The position recommendation should be divided into at least two levels:

• Initial trial position
• Expansion cap after condition verification

At the same time, clarify:

• Under what confirmation conditions adding to the position is allowed
• Under what adverse fluctuations adding further must stop
• Whether it is suitable to go all in at once, or whether it must be phased

Step 6: Stress test whether the position can be executed

Put the recommended position back into the real volatility environment and look again:

• If the market gaps at the open, is the position still acceptable
• If consecutive adverse fluctuations occur, can both psychology and capital withstand it
• If the first position stops out immediately, will it damage subsequent execution discipline
• If the market environment changes abruptly, can the position still be contracted quickly

Step 7: Form the final position recommendation draft

The final output should let the user clearly know “how large this position can be, why it is not larger, when it can be added to, and when it must be reduced.” The main text should retain the constraint logic, recommended numbers, and risk explanation.

Output structure

---

Position decision recommendation ({YYYY-MM-DD})

I. 30-second conclusion

• Trading target: {stock name}
• Recommended position level: {light / medium / heavy / observe only}
• Initial recommendation: {content}
• Expansion conditions: {content}
• Main limiting factors: {one sentence}

II. Position recommendation breakdown

Dimension	Judgment	Impact on position
Risk budget	{content}	{content}
Volatility level	{content}	{content}
Stop-loss distance	{content}	{content}
Portfolio correlation	{content}	{content}
Liquidity	{content}	{content}

III. Execution framework

• Initial position: {content}
• Second add-on conditions: {content}
• Maximum total position cap: {content}
• Conditions to stop adding: {content}

IV. Stress test

• Adverse scenario: {content}
• Tolerability judgment: {content}
• Response action: {content}

V. Risk notice

• {risk 1}
• {risk 2}
• {risk 3}

---

Quality requirements

1. The position recommendation must explain the basis for its formation and cannot only give percentages.
2. The initial position and the maximum total position must be distinguished.
3. It should reflect portfolio correlation and environmental constraints, rather than looking only at a single position.
4. For high-volatility or highly uncertain trades, an explicit downgrade is required.
5. The output should help control position size, rather than rationalize the impulse to take heavy positions.