Reflections on My 2015 ModelOff “Season” (Pt 2)

by Stevenson Yu
 
Find Stevenson on LinkedIn:

 
Part 1 is available here.
 
How did I approach Round 2?
First of all, realistically speaking, there is no chance of qualifying for the Finals. Estimating a total of 8,000 ModelOff participants this year (probably there’s more), there would be a full 1,600 who are ranked at least as good as me, and 800 who are definitely better. My goal was just to test myself (and get some free test materials for my Finance classes, mum’s the word).

Second, something was being cooked up by the Exam Team. Why are they extending the exam time by 15 minutes? Why is the Theory portion only 5 questions long, instead of the usual 20? Not that I wasn’t grateful for the break, but this had to mean that the cases were going to be harder than usual.

Third, the case titles seem to be more informative than usual:

  • “More Money Please” could only mean some debt modeling was in the cards. To prepare, I brushed up the theory on bond amortization, and the Excel functions to do so.
  • “All About Performance” implied that there was going to be some kind of standard that will be compared to some kind of actual performance. There wasn’t much to do to prepare for it, since I was already familiar with the VLOOKUP function, and its reverse equivalent, INDEX+MATCH.
  • “The Price is Right” brought the game show to mind. I expected this to be the 2015 equivalent of last year’s “Snakes and Ladders”, except that we were going to be modeling spins on a wheel. I expected the type of questions to focus on each player’s betting patterns, and probably simulate their chances of winning using Data Tables (my lack of programming knowledge is perhaps my biggest weakness in this type of contest).
  • So, how did I find Round 2 2015? These thoughts are written down immediately after the exam. As such, I can only speculate about how I did, but here goes:

    • Section 1 wasn’t as difficult for me as Round 1 was. The topics tested were vanilla accounting and finance (annuity due vs ordinary annuity and payback period!), and I think I’d get at least 4 out of 5.
    • Section 2 was just as I expected – somewhat-precise debt modeling. However, there was a rub. The first question outright disagreed with my model. Usually, I could answer the first few questions in ModelOff cases. This time, no matter how I tried it, I couldn’t even get the first question – I had nice round numbers while all the choices didn’t. After a fruitless 5 minutes trying to figure it out, I gave up – the first question was a preliminary input – if I can’t get it right, there was no point in continuing because everything else would be wrong. For the second time in two weeks, I found myself guessing on six multiple choice answers. I know my probabilities as well as the next guy; with random guessing, the probability of getting at least 1 question right was 52% [=1-BINOM.DIST(1,10,1/6,TRUE)], but drops to 22% for 2 questions and 7% for 3 questions. I don’t expect more than 2 points for this section, honestly!
    • Section 3 was again as expected, and I actually found it to be quite a lot of fun. I think the tricky part here is to associate the deduction for normal calls with the appropriate band, but an approximate VLOOKUP does the job easily. The main problem I encountered here was the NPV portion. No matter how hard I tried, I couldn’t get close to the choices. Now that I have time to think, I’m sure I erred when I compounded the interest on a quarterly basis (=1.025^1) instead of following instructions and compounding the interest on a fourth-year basis (=1.1^0.25). Assuming I have no other mistakes, I should get at least 7 out of 10 here.
    • Section 4 was completely different from what I expected. This was another financial forecast, with a fair bit of optimization thrown in. From the get-go, I made a huge mistake of not reading all three pages, and instead soldiering on with the first two pages. Needless to say, this made the job of optimizing quite difficult, because I didn’t realize until after the exam that we were supposed to zero-out the short-term debt together with the cash balance. I got the first few questions, but had to guess on the remaining ones that had optimization. My estimated score here should be around 35% to 50%.

    Overall, I found ModelOff 2015 Round 2 quite challenging. Despite the extra 15 minutes, I could not answer anything in Section 2 with absolute confidence (and I still can’t). I’m sure the top modelers did much better, and I congratulate you as you all model off in London next month. Happy Halloween, fellow Excel masochists, and I’ll be back for more punishment next year!
     

    © Copyright 2017 - Vumero Institute