What's new

# FRM Level 1 Nov 2012 Feedback

#### RiskNoob

##### Active Member
itisme

I think the question regarding Bayes' theorem was something like this:

Given

P (positive | fraud ) = 1
P (negative | not fraud) = 0.95 (hence P (positive | not fraud) = 0.05)
P ( fraud ) = 0.001

What is P (fraud | positive) ?

P (fraud | positive) = P(fraud) / P(positive) * P(positive | fraud) , by Bayes’ Theorem
= P (fraud) / [ P(positive | fraud)*P(fraud) + P(positive | not fraud)*P(not fraud) ] * P(positive | fraud)

Plug in the numbers to find the answer

#### mongooseny23

##### Member
The test was fair but seemed highly concentrated in a few areas and it's already only 100 questions. I underestimated the time aspect of it. I hope i don't have to take it again. Jan 2 cannot come soon enough.

#### Bryon

##### Member
I took the test in Paris. I must say that the administrators had no idea about GARP policies. They took away all the pencil sharpeners.

#### Maiyaa

##### New Member
Any idea why the Z table was given..i could not find a relevant question..may be im wrong..they could have saved some trees.

#### balajismz

##### New Member
Can someone remember the options given for the Baye's theorem (Fraud test) question?? I reckon I have guessed it but cant recollect the options.

There is one for question regarding the duration of a callable bond maturing in 15 years and callable after 10 years. This answer still remains 15 years as Callable bonds are never called throughout their term

Regards,

Balaji

#### RiskNoob

##### Active Member
Mayur G

There were at least couple of questions to normalize given random variable to find the probability using Z-table.

Furthermore, one question asked to find (two-sided) p-value - once t-statistic is calculated, we need to refer to Z-table to find the p-value (needs to multiply by 2 since it is two-sided.

RiskNoob

#### delta123

##### New Member
Can someone remember the options given for the Baye's theorem (Fraud test) question?? I reckon I have guessed it but cant recollect the options.

There is one for question regarding the duration of a callable bond maturing in 15 years and callable after 10 years. This answer still remains 15 years as Callable bonds are never called throughout their term

Regards,

Balaji
I would think the answer would be 'between 10 and 15 years' for this question. Pl see this related post http://www.bionicturtle.com/forum/t...-vs-interest-rate-volatility.5738/#post-16431

#### caramel

##### Member
Mayur G

There were at least couple of questions to normalize given random variable to find the probability using Z-table.

Furthermore, one question asked to find (two-sided) p-value - once t-statistic is calculated, we need to refer to Z-table to find the p-value (needs to multiply by 2 since it is two-sided.

RiskNoob

hey Risk noob I thought the T test is 2 sided
and I agree with delta on 10 to 15 years.

#### RiskNoob

##### Active Member
caramel

Yes, Z-table (for the region outside of computed t-statistics) only gives the one sided p-value, so multiply by 2 gives the desired answer (ignoring precision, I think answer was 9.1%*2 = 18.2%).

Yes, I also agree the duration would be somewhat between 10 to 15 years for the answer stated above.

RiskNoob

#### balajismz

##### New Member
caramel

Yes, Z-table (for the region outside of computed t-statistics) only gives the one sided p-value, so multiply by 2 gives the desired answer (ignoring precision, I think answer was 9.1%*2 = 18.2%).

Yes, I also agree the duration would be somewhat between 10 to 15 years for the answer stated above.

RiskNoob

Guys, check this attachement. This question is actually a 1998 FRM question in the similar lines and also refer to Garp's explaination provided below

Cheers

Balaji

#### Attachments

• callable zero coupon bond.docx
203.4 KB · Views: 43

#### balajismz

##### New Member
In theory I would agree with you guys, but the question is practical and I totally agree with Garp's explaination that the Duration should infact be 15 years. The attachment I have provided in my above comments has Garp's explaination for the same.

Cheers,

Balaji

#### balajismz

##### New Member
By the way at around 40-60 problems I saw a spree of Ds, around 5-7 on a trot. Hope I didnt go wrong there

Balaji

#### balajismz

##### New Member
I remember 1 more question on Coherent risk measures comparing both Var and ES.

The options were deceptive as they mentioned Var (A+B)<= Var (A)+ Var (B) among the options.

Clearly at the first glance we would go for this toption, however it is also mentioned to select the option that satisfies in all the cases.

The sub additivity rule only works for Var when the distribution of retuns is normally distributed, so this options seemed incorrect.

Cheers
Balaji

#### caramel

##### Member
I too got many D's in the latter half and very few As

What do you suppose the cut off score will be . My guess it cannot be 70%. its got to be lower. Keeping my fingers and toes crossed as I had t guess in a few places

#### balajismz

##### New Member
well....around 60 is what i am expecting.

Cheers,

Balaji

#### Bryon

##### Member
You guys can remember lots of question! I can't recall much now.
The Z-table were used in, like, 2-3 questions.

I also don't think 70% is the cut-0ff point. Can't be that high.
Imo, the individual questions are on par with level 2 CFA in term of difficulty for the more involved ones. Passing rate for FRM 1 is higher than CFA 2 last time I checked.
I think I should be ok at 60% cut-off then.

#### RiskNoob

##### Active Member
Guys, check this attachement. This question is actually a 1998 FRM question in the similar lines and also refer to Garp's explaination provided below

Cheers

Balaji

balajismz

Good posting! I also googled on this topic and there was a nice discussion from Hend and David:

And I made an educated guess and this particular quote seem to support why the duration (from the answer above for 2012 Nov exam) is not 15 years:

"...duration of the callable bond is monotonically increasing in yield ..."

RiskNoob

#### balajismz

##### New Member
Ohh no.....I actually thought that would be the case but I just went with Garp's explaination around the same.

Balaji

Replies
19
Views
2K
Replies
3
Views
8K
Replies
7
Views
890
Replies
39
Views
8K
Replies
141
Views
26K