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

qiqi1996 · 2020年03月01日

问一道题:NO.PZ201809170400000706

* 问题详情,请 查看题干

问题如下:

Nowacki’s calculated price/book ratios (P/Bs) and price/earnings ratios (P/Es), in his back-testing of the new strategy, are a problem because of:

选项:

A.

data mining.

B.

look-ahead bias.

C.

survivorship bias.

解释:

B is correct. Look-ahead bias results from using information that was unknown or unavailable at the time the investment decision was made. An example of this bias is using financial accounting data for a company at a point before the data were actually released by the company. Nowacki computed historical P/Bs and P/Es using calendar year-end (31 December) stock prices and companies’ financial statement data for the same calendar year, even though the financial statement data for that calendar year were likely unavailable at year-end.

Data mining refers to automated computational procedures for discovering patterns in large datasets, which can introduce a bias known as overfitting. Survivorship bias occurs when back-testing uses companies that are in business today but ignores companies that have left the investment universe.

look ahead偏差该怎么理解呢

1 个答案
已采纳答案

maggie_品职助教 · 2020年03月02日

嗨,爱思考的PZer你好:


Look-ahead bias:前视偏差。用将来的数据进行判断导致判断不可行,因为站在现在这个时间点,将来的数据你是得不到的,那么这种判断规则就是不可用的。比如说我用P/E选股,通过过去的数据我发现P/E越低,将来的股价往往会上涨。比如我现在站在2020.1.1的时间点想根据P/E来进行选股,“E”代表的是去年一整年的每股净收益,但是2019年的年报还没有披露,所以现在根本得不到这个指标,这个P/E是在将来才会得到的,所以用将来才能得到的指标作为一个选股策略是不可行的。


-------------------------------
努力的时光都是限量版,加油!