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

良爱洳 · 2021年03月13日

问一道题:NO.PZ2018122701000032 [ FRM II ]

问题如下:

Below are statements about verifying the accuracy of a VAR model by its failure rate. The least appropriate statements should be:

I. The frequency of exceptions in the model is related to its confidence level.

II. If the log-likelihood ratio is large than 3.84, we should reject the hypothesis that the model is correct (Kupiec, 1995).

III. A small confidence level VAR model is difficult to backtest, as the small number of exceptions barely provide useful information.

IV. There is a tradeoff between the probability of rejecting an accurate model and the probability of accepting an inaccurate model, when considering the number of exceptions

选项:

A.

I and IV

B.

II only

C.

III only

D.

II and IV

解释:

C is correct.

考点 Backtesting VaR

解析 回测high confidence level的VaR模型比较困难,III说反了。

为什么说回测high confidence level的模型比较困难?没在讲义中找到……
1 个答案
已采纳答案

小刘_品职助教 · 2021年03月13日

同学你好,

这题在基础班讲义第66页老师作为一个例题讲过,也比较好理解。首先C描述本身就是错误的,对于low confidence level的VAR来说,超出VaR的数值个数会更多一些,因此exception的个数就较多,对于high confidence level来说恰恰相反,exception的个数更少,因此在回测的时候能找到的意外值更少,所以比较困难。

具体老师的视频,在Describe The Basel Rules For Backtesting 1.5倍速9分钟左右,如果觉得我描述的不清楚,可以再去听听


410140980 · 2023年06月30日

老师我不理解low confidence level的VAR来说,例外的个数更多,回测的时候找到意外值的概率更高呢?

  • 1

    回答
  • 3

    关注
  • 578

    浏览
相关问题

NO.PZ2018122701000032 问题如下 Below are statements about verifying theaccuraof a Vmol its failure rate. The least appropriate statementsshoulbe: I. The frequenof exceptions inthe mol is relateto its confinlevel. II. If the log-likelihooratio islarge th3.84, we shoulrejethe hypothesis ththe mol is corre(Kupiec,1995).III. A small confinlevel VARmol is fficult to backtest, the small number of exceptions barelyprovi useful information. IV.There is a traoff between the probability of rejecting accurate mol anhe probability of accepting inaccurate mol, when consiring the numberof exceptions I anIV II only III only II anIV C is correct. 考点 Backtesting V 解析 回测high confinlevel的VaR模型比较困难,III说反了。 如题

2024-03-17 18:59 1 · 回答

NO.PZ2018122701000032 问题如下 Below are statements about verifying theaccuraof a Vmol its failure rate. The least appropriate statementsshoulbe: I. The frequenof exceptions inthe mol is relateto its confinlevel. II. If the log-likelihooratio islarge th3.84, we shoulrejethe hypothesis ththe mol is corre(Kupiec,1995).III. A small confinlevel VARmol is fficult to backtest, the small number of exceptions barelyprovi useful information. IV.There is a traoff between the probability of rejecting accurate mol anhe probability of accepting inaccurate mol, when consiring the numberof exceptions I anIV II only III only II anIV C is correct. 考点 Backtesting V 解析 回测high confinlevel的VaR模型比较困难,III说反了。 如题

2024-03-05 02:14 1 · 回答

NO.PZ2018122701000032问题如下 Below are statements about verifying theaccuraof a Vmol its failure rate. The least appropriate statementsshoulbe: I. The frequenof exceptions inthe mol is relateto its confinlevel. II. If the log-likelihooratio islarge th3.84, we shoulrejethe hypothesis ththe mol is corre(Kupiec,1995).III. A small confinlevel VARmol is fficult to backtest, the small number of exceptions barelyprovi useful information. IV.There is a traoff between the probability of rejecting accurate mol anhe probability of accepting inaccurate mol, when consiring the numberof exceptions I anIV II only III only II anIV C is correct. 考点 Backtesting V 解析 回测high confinlevel的VaR模型比较困难,III说反了。 A感觉应该是例外的个数,这里写的频率是不是有点问题

2024-02-24 22:25 1 · 回答

NO.PZ2018122701000032 4的tra off改如何理解?

2021-04-09 15:37 1 · 回答