问题如下图：
选项：
A. 
B. 
C. 
D. 
解释：
怎么区别是overestimate还是underestimate？
品职答疑小助手雍 · 2019年09月10日
同学你好，解析的第一部分就是在假设检验8次exception是不是太高了，结果是太高了99%的confidence双尾的检验最高exception 的次数是2.5+2.58*根号下（250*0.01*0.99）小于8.
但是exception实打实的出现了8次，那证明现在的var估计的数值太小了，总损失被超过。也就是之前的var是被低估的。
moon · 2019年09月11日
第一步Z score检验的原假设是exception出现大于等于8次这样吗？然后检验出来3.5〉2.58，所以拒绝原假设？2.58是什么值吖？
品职答疑小助手雍 · 2019年09月11日
原假设是出现8次exception的情况下的var值是合理的，检验结果是拒绝的。2.58对应的是双尾99%的z值。
moon · 2019年09月18日
2.33不是99%的Z值吗？ 我好像有点记混乱
品职答疑小助手雍 · 2019年09月19日
单尾是2.33比如算var的时候，这里是双尾的检验。单边其实是0.5%不是1%。
NO.PZ2016070202000006问题如下 Tycoon Bank announceththere were eight ys in the previous yefor whilosses exceethe ily 99% VAR. a result, concerns emergeabout the accuraof the Vimplementation. Assuming ththere are 250 ys in the year, whiof the following statements is/are correct?I. Using a two-taile99% confinlevel z-score test, the current Vimplementation unrstates the acturisk in the bank's portfolio.II. Using a two-taile99% confinlevel z-score test, the current Vimplementation overstates the acturisk in the bank's portfolio.III. The bank's exception rates for Vminaccurate if the bank's portfolio changes incorporate the returns from low-risk but highly profitable intray market making activities.IV. If these eight exceptions all happenein the previous month, the mol shoulreexaminefor faulty assumptions aninvaliparameters. A.I anIIIB.I, III, anIVC.Ill onlyI, II, anIVThe z-score gives 8−2.5250×0.01×0.99=3.5\frac{8-2.5}{\sqrt{250\times0.01\times0.99}}=3.5250×0.01×0.998−2.5=3.5 This is too high (greater th2.58), whilea to rejection of the null ththe Vmol is well calibrate Hence, Vis too low anstatement I. is correct. Statement II. is incorrect. However, this me to intray trang, so III. is correct, too. Finally, if all eight exceptions occurrein the last month, there is bunching, anthe mol shoulreexamine so IV. is correct.老师，III的意思是说，因为日内交易使头寸中少了很多风险小、收益高的资产，所以剩下头寸的亏损情况就变了，最终会出现很多exception么？
NO.PZ2016070202000006 截图中的公式是怎么算的，原理是怎样？
NO.PZ2016070202000006 能一下Ⅲ吗？有点没太明白
NO.PZ2016070202000006 当V天数越大，那算出来p-value 大 月会就拒绝，怎么是 overestimate ,明明是unrestimate ma
NO.PZ2016070202000006 I, III, anIV Ill only I, II, anIV The z-score gives 8−2.5250×0.01×0.99=3.5\frac{8-2.5}{\sqrt{250\times0.01\times0.99}}=3.5250×0.01×0.99 8−2.5=3.5 This is too high (greater th2.58), whilea to rejection of the null ththe Vmol is well calibrate Hence, Vis too low anstatement I. is correct. Statement II. is incorrect. However, this me to intray trang, so III. is correct, too. Finally, if all eight exceptions occurrein the last month, there is bunching, anthe mol shoulreexamine so IV. is correct.请问这里的VAR是如何判断出是被低估了呢？