什么是 Alpha？｜What is Alpha?
超额收益、基准与因子验证的研究边界。Excess returns, benchmarks and factor validation.
编辑更新 / Reviewed: 2026-10-06 · ORAKT editorial
定义 / Definition
Alpha 是相对于指定风险模型或基准、不能被该模型解释的收益部分。它不是总收益，也不是上涨概率。Alpha describes returns unexplained by a specified benchmark or risk model; it is not total return or a probability of gains.
为什么需要挖掘？ / Why research it?
新闻、财报和价格中可能存在与收益有关的信号，但相关性可能来自行业行情、风险暴露或偶然。研究需要提出机制，再检验它是否超出这些解释。Signals in news, filings and prices may reflect common market moves or chance. A mechanism must survive competing explanations.
如何核验 / Validation
ORAKT 将候选信号与已验证因子区分。预先固定样本、披露滞后与成本，比较行业和大盘，保留失败实验，并使用独立样本、置信区间与未来验证。历史重建报告不是当时真实发出的预测。Candidates are not validated factors. Use disclosure lags, costs, sector and market benchmarks, untouched samples and uncertainty. Retrospective reports are not historical live predictions.
来源与进一步阅读 / Sources
来源支持对应背景，不代表验证所有传导或投资结论。Sources support context, not every inference or investment conclusion.