开发者:上海品职教育科技有限公司 隐私政策详情

应用版本:4.2.11(IOS)|3.2.5(安卓)APP下载

K · 2023年02月21日

怎么判断这里想问“造成什么问题”而非“做对了什么”

NO.PZ2021101401000020

问题如下:

Galic is surprised to see that some of the backtest results are unfavorable. He asks, “Why has GWP not considered strategies that perform better in backtesting?” Galic recently met with Fastlane Wealth Managers, who showed much better performance results. The portfolio manager at Fastlane told Galic that the company selects the top-performing strategies after performing thousands of backtests.

The approach used by Fastlane Wealth Managers most likely incorporates:

选项:

A.

risk parity.

B.

data snooping.

C.

cross-validation.

解释:

B is correct. The fact that the two firms’ investment performance results differ over similar time horizons using the same data and factors may be the result of selection bias. Data snooping is a type of selection bias. Fastlane Wealth Managers is most likely selecting the best-performing modeling approach and publishing its results (i.e., data snooping).

A is incorrect because risk parity is a portfolio construction technique that accounts for the volatility of each factor and the correlations of returns among all factors to be combined in the portfolio. It is not regarded as selection bias.

C is incorrect because cross-validation is a technique used in the machine learning field, as well as in backtesting investment strategies, to partition data for model training and testing. It is not considered selection bias.

题干同时提到F公司管理的组合表现好,容易让人误以为题目想考的是他们通过cross validation的方式规避了data snooping

1 个答案
已采纳答案

星星_品职助教 · 2023年02月21日

同学你好,

题干的描述“... the company selects the top-performing strategies after performing thousands of backtests”是明显的错误和考点,即这个方法本身就是错误的。所以,“most likely"包括的就是错误的原因。

不太可能在一个严重错误的方法里去找做对了什么。此外,这种错误的方法无论是否做交叉验证,都规避不了data snooping。


  • 1

    回答
  • 0

    关注
  • 414

    浏览
相关问题

NO.PZ2021101401000020 问题如下 Galic is surpriseto see thsome of the backtest results are unfavorable. He asks, “Why hGWP not consirestrategies thperform better in backtesting?” Galic recently met with Fastlane Wealth Managers, who showemubetter performanresults. The portfolio manager Fastlane tolGalic ththe company selects the top-performing strategies after performing thousan of backtests.The approauseFastlane Wealth Managers most likely incorporates: A.risk parity. B.ta snooping. C.cross-valition. B is correct. The faththe two firms’ investment performanresults ffer over similtime horizons using the same ta anfactors mthe result of selection bias. ta snooping is a type of selection bias. Fastlane Wealth Managers is most likely selecting the best-performing moling approaanpublishing its results (i.e., ta snooping).A is incorrebecause risk parity is a portfolio construction technique thaccounts for the volatility of eafactor anthe correlations of returns among all factors to combinein the portfolio. It is not regarselection bias.C is incorrebecause cross-valition is a technique usein the machine learning fiel well in backtesting investment strategies, to partition ta for mol training antesting. It is not consireselection bias. 老师,我听了课后题老师的讲解,还是不太明白。The portfolio manager Fastlane tolGalic ththe company selects the top-performing strategies after performing thousan of backtests.感觉选择的对呀,不就是选择表现好的策略么?不这么选择,应该怎么选择呢?

2022-09-07 00:10 1 · 回答

NO.PZ2021101401000020问题如下 Galic is surpriseto see thsome of the backtest results are unfavorable. He asks, “Why hGWP not consirestrategies thperform better in backtesting?” Galic recently met with Fastlane Wealth Managers, who showemubetter performanresults. The portfolio manager Fastlane tolGalic ththe company selects the top-performing strategies after performing thousan of backtests.The approauseFastlane Wealth Managers most likely incorporates: A.risk parity. B.ta snooping. C.cross-valition. B is correct. The faththe two firms’ investment performanresults ffer over similtime horizons using the same ta anfactors mthe result of selection bias. ta snooping is a type of selection bias. Fastlane Wealth Managers is most likely selecting the best-performing moling approaanpublishing its results (i.e., ta snooping).A is incorrebecause risk parity is a portfolio construction technique thaccounts for the volatility of eafactor anthe correlations of returns among all factors to combinein the portfolio. It is not regarselection bias.C is incorrebecause cross-valition is a technique usein the machine learning fiel well in backtesting investment strategies, to partition ta for mol training antesting. It is not consireselection bias. 谢谢

2022-04-16 16:48 1 · 回答