Actuarial Outpost > SoA Can-do - Task 3 How do you obtain the optimal number of assays?
 Register Blogs Wiki FAQ Calendar Search Today's Posts Mark Forums Read
 FlashChat Actuarial Discussion Preliminary Exams CAS/SOA Exams Cyberchat Around the World Suggestions

#21
03-20-2011, 04:30 AM
 kouzhong23 Join Date: Aug 2010 Posts: 19

Quote:
 Originally Posted by elfueg07 The decision strategy is whatever you think it should be. Your point #1 is very good, and in real life I would imagine the 'decision strategy' would be developed with the client's input as to what is important to them. Unfortunately.

For Task 3, what are the advantages and disadvantages of conducting too many or too few assays in terms of ending cash flow. In my original answer, more assays means more costs but less deviation of ending cash flow. Now, if I do not make the two modifications as I mentioned above, more assays will still lead to more costs and less ending cash flow. However, I found it difficult to state the advantages of conducting more assays. General speaking, more assays can lead the estimated decay factor to approach the simultated decay factor according to the original model, so it might affect the ending cash flow if the minimum to mine value is set to be higher than a certain value, say, 0.92. But the effect of adding more assays on ending cash flows is not clear to state. In my model, for minimum to mine larger than 0.92, the Var(90) will decrease and then increase when number of assays increase, however, the Var(80) is decreasing all the way when number of assays rises. So, how should I express the advantages or effect of assay numbers on ending cash flow. Should I need to answer this question by the actual simulation results or just by general statement, for which the effect is not easy to state. Thanks!
#22
03-20-2011, 11:43 AM
 elfueg07 Member SOA Join Date: Nov 2010 College: University of Wisconsin Posts: 95

Let's say you conduct 10 assays. Now you want to examine the effect of instead running 11 assays.

You have 1000 simulations. Some simulations you open the mine, some you close the mine. Doing 11 assays can have one of the following three effects, on each individual simulation:
1) do not change the mine/no mine decision
2) change from mine to no mine
3) change from no mine to mine

For all those simulations that fall into category #1, all you are doing is adding the cost from the 11th assay. Probably over 75% (maybe even 90%) of the simulations fall into this category, so this is why your average ending cash probably decreases.

For simulations in category #2 and #3, you would hope that the change was to the advantage of the mine, so overall they would probably have a positive result (increase cash). I don't think there would be very many simulations in these categories to balance the first.

However, categories 2 and 3 would hopefully have the effect of reducing the very bad losses, and perhaps lowering CTE measures.

Everything I have said depends on the levels you have recommended for MTM and number of assays, so please keep that in mind.
#23
03-20-2011, 12:24 PM
 kouzhong23 Join Date: Aug 2010 Posts: 19

Quote:
 Originally Posted by elfueg07 Let's say you conduct 10 assays. Now you want to examine the effect of instead running 11 assays. You have 1000 simulations. Some simulations you open the mine, some you close the mine. Doing 11 assays can have one of the following three effects, on each individual simulation: 1) do not change the mine/no mine decision 2) change from mine to no mine 3) change from no mine to mine For all those simulations that fall into category #1, all you are doing is adding the cost from the 11th assay. Probably over 75% (maybe even 90%) of the simulations fall into this category, so this is why your average ending cash probably decreases. For simulations in category #2 and #3, you would hope that the change was to the advantage of the mine, so overall they would probably have a positive result (increase cash). I don't think there would be very many simulations in these categories to balance the first. However, categories 2 and 3 would hopefully have the effect of reducing the very bad losses, and perhaps lowering CTE measures. Everything I have said depends on the levels you have recommended for MTM and number of assays, so please keep that in mind.
I would probably understand how you answer the advantage and disadvantage of doing too much or too few assays. You recommend an optimal minimum to mine and number of assay first, and too much and too few are relative to the optimal point, right? So actually, it may be more approriate to answer this question first with general statment, and after we have done the analysis to seek the optimal value of assay number and MTM value, we might re-state the answer with specifc statistics.
#24
03-20-2011, 12:33 PM
 elfueg07 Member SOA Join Date: Nov 2010 College: University of Wisconsin Posts: 95

Yeah, I think that would do a good job of answering the question.
#25
05-27-2011, 11:55 AM
 Mr F Member SOA Join Date: Jun 2009 Posts: 4,829

Quote:
 Originally Posted by kouzhong23 2) I don't think the "decay from assays" in G12 is only used to compared to "minimum to mine" factor to judge whether or not to open the mine. It should take the place of G11 to be used to "G16:G95" to be the refined decay rate because more assays will not only take costs to the company, but also benefit! And the benefit, if only, should be a decreased variability of decay rate (from 0.02 to 0.01 when assay numbers approaching to 20 in cells "G12") and so a decreased variability of gold yield rate in "G16 to G95"
FYI, my passing submission contained the second revision outlined here.
#26
05-31-2011, 07:47 PM
 LifeTilt Member Join Date: Jul 2008 Posts: 184

Quote:
 Originally Posted by Mr F FYI, my passing submission contained the second revision outlined here.
Not to pick a fight, but this post seems a little reckless IMO. Just because you did one of the methods here and passed, doesn't mean that you were correct. You may have been 100% wrong, but done well on all the other parts. It's not that you aren't possibly right, I just think the qualifier that "my passing submission..." is misleading and people should take it with a grain of salt.
#27
05-31-2011, 10:34 PM
 Mr F Member SOA Join Date: Jun 2009 Posts: 4,829

There are posts that suggest that the line "Never change assumptions..." means if you change anything other than that one sheet then it is an autofail. I disagree. I believe that it refers to the "Assumptions" tab, since that also says to not change value on that sheet. Run through the macro, if you change assumptions there then the macro breaks.

The results I got back before making the change didn't fit in the context of the assignment. I actually had something to write about after the change.

Last edited by Mr F; 05-31-2011 at 11:10 PM..
#28
06-01-2011, 11:35 AM
 LifeTilt Member Join Date: Jul 2008 Posts: 184

Quote:
 Originally Posted by Mr F There are posts that suggest that the line "Never change assumptions..." means if you change anything other than that one sheet then it is an autofail. I disagree. I believe that it refers to the "Assumptions" tab, since that also says to not change value on that sheet. Run through the macro, if you change assumptions there then the macro breaks. The results I got back before making the change didn't fit in the context of the assignment. I actually had something to write about after the change.
Maybe you can explain what you did? No offense to KuoZhong, but his writing skills were questionable at best, so I could barely understand what was going on in this entire thread. I think this will help future takers, if you care to elaborate. You've been pretty helpful throughout so maybe you can add a little more.
#29
06-01-2011, 01:08 PM
 Mr F Member SOA Join Date: Jun 2009 Posts: 4,829

Quote:
 Originally Posted by LifeTilt Maybe you can explain what you did? No offense to KuoZhong, but his writing skills were questionable at best, so I could barely understand what was going on in this entire thread. I think this will help future takers, if you care to elaborate. You've been pretty helpful throughout so maybe you can add a little more.
Sure. Busy couple days at work. I'll elaborate later.
#30
06-01-2011, 01:31 PM
 malaprop Member Join Date: May 2011 Posts: 72

Quote:
 Originally Posted by LifeTilt Maybe you can explain what you did? No offense to KuoZhong, but his writing skills were questionable at best, so I could barely understand what was going on in this entire thread. I think this will help future takers, if you care to elaborate. You've been pretty helpful throughout so maybe you can add a little more.
I have read through the Can-Do threads. Most of the advice given is not good advice. Some of it has been downright dirigible, but I will refrain. The only person I recall who seemed to have a decent grasp on the assignment was elfueg07.

As an example, the spreadsheet adjustment in question is not appropriate. The "true" decay rate is in G11. The assays do not change the decay rate. The assays are based on an estimate which has a different, lower standard deviation than the one used to develop the "true" decay rate scenarios. Each assay removes 5% of the difference between the estimate and the "true" decay rate.

Mr. F's post is vindictive of the fact that the graders do not expect the project to be understood completely. This is probably because, unlike a real work assignment, you cannot ask questions of people actually involved in the project. Take this to heart, and do your best to defend whatever approach you took. The approach taken seems to matter a lot less.

 Tags can-do, final assessment