Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →A股仓位管理/凯利公式/仓位计算。当用户说"仓位"、"position sizing"、"该买多少"、"仓位管理"、"凯利公式"、"Kelly"、"加仓"、"减仓"时触发。基于 cn-stock-data 获取数据，量化计算最优仓位。支持研报风格（formal）和快速分析风格（brief）。
.claude/skills/aifinlab-a-share-position-sizing/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-17 | ✗→✓ | ▲ Improved | -49% | 0% |
| case-18 | ✗→✓ | ▲ Improved | -47% | 0% |
| case-09 | ✓→✗ | ▼ Worse | -57% | 0% |
| case-04 | ✓→✗ | ▼ Worse | -66% | 0% |
| case-05 | ✓→✗ | ▼ Worse | -69% | 0% |
bashSCRIPTS="$SKILLS_ROOT/cn-stock-data/scripts" python "$SCRIPTS/cn_stock_data.py" kline --code [CODE] --freq daily --start [日期] python "$SCRIPTS/cn_stock_data.py" quote --code [CODE] python "$SCRIPTS/cn_stock_data.py" finance --code [CODE]
获取标的K线 + 波动率 + 基本面数据。
| 维度 | formal | brief | |------|--------|-------| | 仓位计算 | 多方法对比 | 建议仓位 | | 风险分析 | 波动率+VaR | 风险等级 | | 加减仓计划 | 分批建仓方案 | 单次建议 |
默认风格：brief。
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-20 | fail→fail | 20,241 | 18,962 | -6% | 1 | 1 | 0% | 2,960 | 3,681 | +24% | 0 | 0 | — |
case-08 | fail→fail | 11,101 | 8,947 | -19% | 1 | 1 | 0% | 1,646 | 986 | -40% | 0 | 0 | — |
case-01 | fail→fail | 20,414 | 13,306 | -35% | 1 | 1 | 0% | 2,894 | 1,131 | -61% | 0 | 0 | — |
case-02 | fail→fail | 30,629 | 8,281 | -73% | 1 | 1 | 0% | 5,859 | 1,007 | -83% | 0 | 0 | — |
case-09 | pass→fail | 18,545 | 9,073 | -51% | 1 | 1 | 0% | 2,601 | 1,109 | -57% | 0 | 0 | — |
case-03 | fail→fail | 22,885 | 40,018 | +75% | 1 | 1 | 0% | 3,375 | 1,034 | -69% | 0 | 0 | — |
case-04 | pass→fail | 18,519 | 6,359 | -66% | 1 | 1 | 0% | 2,687 | 909 | -66% | 0 | 0 | — |
case-05 | pass→fail | 18,281 | 8,178 | -55% | 1 | 1 | 0% | 2,929 | 921 | -69% | 0 | 0 | — |
case-06 | pass→pass | 12,923 | 11,300 | -13% | 1 | 1 | 0% | 2,125 | 2,564 | +21% | 0 | 0 | — |
case-07 | pass→fail | 19,256 | 18,795 | -2% | 1 | 1 | 0% | 3,186 | 3,295 | +3% | 0 | 0 | — |
case-10 | fail→fail | 15,664 | 16,564 | +6% | 1 | 1 | 0% | 2,692 | 1,103 | -59% | 0 | 0 | — |
case-11 | pass→pass | 14,937 | 7,430 | -50% | 1 | 1 | 0% | 2,074 | 1,781 | -14% | 0 | 0 | — |
case-12 | pass→pass | 14,003 | 6,032 | -57% | 1 | 1 | 0% | 2,298 | 1,559 | -32% | 0 | 0 | — |
case-13 | pass→pass | 19,323 | 6,406 | -67% | 1 | 1 | 0% | 2,705 | 1,596 | -41% | 0 | 0 | — |
case-14 | fail→fail | 8,454 | 8,246 | -2% | 1 | 1 | 0% | 1,618 | 1,030 | -36% | 0 | 0 | — |
case-15 | pass→pass | 6,027 | 16,271 | +170% | 1 | 1 | 0% | 1,205 | 2,273 | +89% | 0 | 0 | — |
case-16 | pass→pass | 6,227 | 6,742 | +8% | 1 | 1 | 0% | 1,331 | 1,701 | +28% | 0 | 0 | — |
case-17 | fail→pass | 11,283 | 3,881 | -66% | 1 | 1 | 0% | 2,090 | 1,070 | -49% | 0 | 0 | — |
case-18 | fail→pass | 17,457 | 7,487 | -57% | 1 | 1 | 0% | 3,204 | 1,705 | -47% | 0 | 0 | — |
case-19 | fail→fail | 12,980 | 11,313 | -13% | 1 | 1 | 0% | 2,246 | 2,402 | +7% | 0 | 0 | — |
case-21 | fail→fail | 5,297 | 5,832 | +10% | 1 | 1 | 0% | 731 | 929 | +27% | 0 | 0 | — |
case-22 | pass→fail | 39,088 | 48,606 | +24% | 1 | 1 | 0% | 6,700 | 950 | -86% | 0 | 0 | — |
case-23 | pass→fail | 14,758 | 9,623 | -35% | 1 | 1 | 0% | 2,434 | 1,019 | -58% | 0 | 0 | — |
case-24 | pass→pass | 7,548 | 8,597 | +14% | 1 | 1 | 0% | 1,504 | 2,106 | +40% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 24 cases were attempted, and 12 counted toward the lift figure. The other 12 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of -17 percentage points is the difference between those two pass rates over the 12 comparable cases. 13 cases got worse with the skill loaded, and they are included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 8/28/2026 | -9% |
Other measured skills in the registry, with their headline benchmark lift.