> ## Documentation Index
> Fetch the complete documentation index at: https://raymondchang.me/llms.txt
> Use this file to discover all available pages before exploring further.

# Overfitting controls

> Deflated Sharpe、PBO、walk-forward 與 purged CV 等防止過度擬合的方法。

> [!NOTE]
>
> **Status of this page：** 大綱（skeleton）。內容尚未撰寫，未經來源驗證。

## 大綱

1. 過度擬合的來源：多次試驗、參數搜尋、資料窺探
2. Deflated Sharpe ratio 與試驗次數的修正
3. PBO（probability of backtest overfitting）與 CSCV 的概念
4. Walk-forward 與 purged / embargoed K-fold
5. 實務檢查表：參數敏感度、鄰近參數穩定性、經濟合理性

## 面試常問

* 測了 100 個策略挑出最佳者，Sharpe 該怎麼折減？
* 為什麼金融時間序列的 K-fold 需要 purge？

## Related pages

* [Estimation and hypothesis testing](https://raymondchang.me/quantitative-research/foundations/estimation-testing.md)
* [Bias taxonomy](https://raymondchang.me/quantitative-research/backtesting/bias-taxonomy.md)


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.