Mobile Phone Contracts

[tatos]

I've been going through this past paper - When I was trying to build the model, I had wanted to use the time at which the switch is made in proposal 2 (third scenario) as a parameter (in case the user might want to change the parameter's value later on)

But this was taking forever, and I was just not using the right combination of vlookup and nested if statements apparently. I think it must be due to how I set up my model in general, which had (up until that time) worked great and provided the correct answers without too much effort. Then I decided to remove the parameterisation and insert simple formulae at the switching times which were mentioned in the instructions. I then mentioned what I had done in my audit trail and highlighted cells where I had done this with a text box note in the model. This provided the correct answer and it was very quick.

But I'm wondering a couple things now:

- Was it really necessary to even mention that I had essentially fixed the time at which the switch would take place (rather than used a parameter)?
- How do you know what should DEFINITELY be parameterised? Or is it more a question of doing what you are able to do and documenting it well so that the next user then knows how to modify it and check it?

It seems that the marks I would lose for not parameterising something like that are far outweighed by the time I would save (and more so, I guess, if I made note of this in the audit trail)... Anyone with any advice or experience to share?

 

[cjno1]

The biggest tip I would give is to read the question carefully and only do what is required.

If the question says that all switches will occur in month X, then just build the model assuming that this is fixed, I wouldn't waste time trying to make the model more complicated than you need to. However, if it says something like "switches ARE MOST LIKELY to occur in month X" or "switches might occur BETWEEN MONTHS X AND Y" or something like that, then you'll need the flexibility you talked about.

If you finish the model, the audit trail and the summary and you still have an hour or so left and you want to add in little fancy tweaks that you think might score extra credit, then by all means do so. But CA2 is usually time pressured enough already for most people without trying to go above and beyond what the question asks of you.

 

[didster]

I agree with cjno1 that you're usually under time pressure with CA2 so you should try to keep it simple and to just do what is necessary.

The whole point of setting up the model to use parameters is to

• make it easier to follow
• make it easier to adjust for different scenarios

for both you and someone checking/picking up where you left off.

If including the parameters as you go along is easy, then you should do so.
If it is slowing you down then you need to balance whether doing it with parameters would save you time later on. (Focus primarily on time saved in the exam. If you can do little things to save hypothetical time for someone modifying/checking you work ( as if it was a real life project) then these may get extra points but treat it as secondary)
If a parameter is causing incorrect results then don't.

One approach is to make sure small parts are working and then tidy it up later.

If you're considering setting something up as a parameter then you should include it in the audit trail.
It might not be absolutely necessary, but chances are that if you're considering it, someone else might. A simple sentence in the audit trail points you to what you need to do to expand the model, confirms the approach you used (eg assume a fixed date), etc.

 

[tatos]

Thanks cjno1 and didster! That really helped to give more perspective/ clarity on this..

 

[Tommy]

That's exactly the type of thing I checked with the examiner..."I'm going to assume X, Y and Z, does that sound reasonable?".