引導式測驗 · 中等 等級
交易策略回測與過擬合測驗
解釋回溯測試如何過度擬合歷史數據,以及哪些假設、成本和驗證檢查屬於可信評估。
問題 1 的 8
What does a backtest measure?
預覽每個問題8 問題
當您參加測驗時,就會出現答案和解釋。
- What does a backtest measure?
- A hypothetical replay of a strategy on historical data under assumptions
- Actual live returns earned by every investor
- A guarantee of future returns
- The legal risk of an issuer
- Why can trying many strategy variants create overfitting?
- It ensures all strategies have equal returns
- It removes trading costs
- It guarantees independent evidence
- Selecting the best historical result can capitalize on noise
- Which bias occurs when a backtest uses information that was unavailable at the simulated decision time?
- Survivorship bias only
- Look-ahead bias
- Currency risk
- Diversification
- Which costs should a realistic backtest consider?
- Commissions alone while assuming every quote is filled
- Only subscriptions for historical price data
- Fees, slippage and liquidity or execution limits
- Borrowing costs alone without market impact or fills
- Why can one untouched holdout be inadequate after many strategy searches?
- It always prevents overfitting
- It can be reused repeatedly and become part of the selection process
- It guarantees live returns
- It makes prices independent
- What should a researcher record before reporting a selected strategy?
- Only the highest-return run after tuning
- Only the final parameter values
- All strategy variants tried and the selection process
- Only the chosen holdout score without prior trials
- Does a strong backtest guarantee future profits?
- Yes, if it covers ten years
- Only when it has no drawdown
- Yes, if a model uses AI
- No; historical results do not guarantee future performance
- What may the SEC marketing rule require when an adviser advertises hypothetical performance?
- Information about assumptions and conditions for the intended audience
- Nothing if a chart is included
- A guarantee the strategy will work
- Only a gross return figure