Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →A股自相关/序列相关性/收益率自相关结构分析。当用户说"自相关"、"autocorrelation"、"序列相关"、"收益率预测性"、"动量还是反转"、"自相关系数"、"ACF"、"PACF"、"Ljung-Box"、"收益率是否可预测"、"随机游走检验"时触发。MUST USE when user asks about return autocorrelation, serial correlation tests, or whether a stock's returns are predictable. 量化分析收益率的自相关结构（ACF/PACF、Ljung-Box检验、随机游走检验）。支持formal和brief风格。
.claude/skills/aifinlab-a-share-autocorrelation/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-16 | ✗→✓ | ▲ Improved | 0% | 0% |
| case-03 | ✓→✗ | ▼ Worse | -77% | 0% |
| case-10 | ✓→✗ | ▼ Worse | -52% | 0% |
| case-12 | ✓→✗ | ▼ Worse | -28% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | -12% | 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]
lag 1-20的自相关系数
检验序列是否存在显著自相关
| 维度 | formal | brief | |------|--------|-------| | ACF/PACF | 完整图表 | 关键lag | | 检验 | LB统计量+p值 | 有无自相关 | | 含义 | 动量/反转判断 | 交易含义 | 默认风格：brief。
python# 调用 skill result = run_skill({ "param1": "value1", "param2": "value2" })
bashpython scripts/run_skill.py --input data.json
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 22,811 | 16,272 | -29% | 1 | 1 | 0% | 3,671 | 1,357 | -63% | 0 | 0 | — |
case-02 | fail→fail | 28,738 | 24,968 | -13% | 1 | 1 | 0% | 5,337 | 1,729 | -68% | 0 | 0 | — |
case-03 | pass→fail | 18,816 | 7,661 | -59% | 1 | 1 | 0% | 3,903 | 879 | -77% | 0 | 0 | — |
case-04 | pass→pass | 19,620 | 18,616 | -5% | 1 | 1 | 0% | 3,878 | 3,430 | -12% | 0 | 0 | — |
case-05 | fail→fail | 26,004 | 9,190 | -65% | 1 | 1 | 0% | 3,953 | 1,189 | -70% | 0 | 0 | — |
case-06 | fail→fail | 18,935 | 13,223 | -30% | 1 | 1 | 0% | 3,015 | 1,447 | -52% | 0 | 0 | — |
case-07 | fail→fail | 16,213 | 4,511 | -72% | 1 | 1 | 0% | 2,612 | 1,327 | -49% | 0 | 0 | — |
case-13 | fail→fail | 36,765 | 14,532 | -60% | 1 | 1 | 0% | 5,015 | 871 | -83% | 0 | 0 | — |
case-08 | pass→pass | 15,806 | 15,513 | -2% | 1 | 1 | 0% | 2,524 | 2,677 | +6% | 0 | 0 | — |
case-09 | pass→pass | 13,506 | 15,594 | +15% | 1 | 1 | 0% | 2,182 | 2,656 | +22% | 0 | 0 | — |
case-10 | pass→fail | 21,525 | 13,228 | -39% | 1 | 1 | 0% | 3,041 | 1,446 | -52% | 0 | 0 | — |
case-11 | pass→pass | 22,483 | 24,524 | +9% | 1 | 1 | 0% | 3,006 | 2,931 | -2% | 0 | 0 | — |
case-12 | pass→fail | 10,746 | 35,270 | +228% | 1 | 1 | 0% | 1,669 | 1,208 | -28% | 0 | 0 | — |
case-14 | fail→fail | 16,430 | 16,383 | -0% | 1 | 1 | 0% | 2,732 | 3,089 | +13% | 0 | 0 | — |
case-15 | pass→pass | 14,524 | 11,072 | -24% | 1 | 1 | 0% | 1,882 | 2,203 | +17% | 0 | 0 | — |
case-16 | fail→pass | 16,537 | 13,199 | -20% | 1 | 1 | 0% | 2,511 | 2,502 | -0% | 0 | 0 | — |
case-17 | pass→pass | 20,012 | 18,135 | -9% | 1 | 1 | 0% | 2,631 | 3,166 | +20% | 0 | 0 | — |
case-18 | fail→fail | 9,481 | 4,792 | -49% | 1 | 1 | 0% | 1,614 | 1,240 | -23% | 0 | 0 | — |
case-19 | fail→fail | 19,129 | 17,139 | -10% | 1 | 1 | 0% | 3,118 | 3,303 | +6% | 0 | 0 | — |
case-20 | fail→fail | 15,673 | 4,935 | -69% | 1 | 1 | 0% | 2,888 | 1,370 | -53% | 0 | 0 | — |
case-21 | pass→pass | 18,496 | 17,211 | -7% | 1 | 1 | 0% | 2,711 | 3,067 | +13% | 0 | 0 | — |
case-22 | pass→pass | 21,004 | 17,077 | -19% | 1 | 1 | 0% | 2,680 | 2,964 | +11% | 0 | 0 | — |
case-23 | pass→pass | 19,690 | 24,039 | +22% | 1 | 1 | 0% | 2,865 | 3,065 | +7% | 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. 23 cases were attempted, and 15 counted toward the lift figure. The other 8 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 -9 percentage points is the difference between those two pass rates over the 15 comparable cases. 8 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 | +18% |
Other measured skills in the registry, with their headline benchmark lift.