· 方法论

自建数据管道发现$VLN异常,强调人工复核优于算法。

涉及标的:

中文翻译

谢谢,我自建了类似 Citadel 的数据管道,在研究机器人板块股票时发现了 $VLN。这可能是我迄今发现的最大异常,因此我不得不手动将财务报告与分析师/扫描器报告进行交叉核对,以验证其真实性。即使算法 + 大语言模型(LLM) 也搞错了这一分析,只有人工审查才能发现这种差异。

英文原文

Thanks, I built my own Citadel-like data pipeline and I was researching robotics sector stocks and came across $VLN. This was probably one of the biggest anomalies I've ever found so I had to manually cross-check financial reports with analyst/scanner reports to see if it were true.   Even algorithms + LLMs got this analysis wrong, so only manual review could spot this discrepancy.

在 X 上查看原推 ↗