Our last piece on pondering inconsequential questions was a such a massive hit, that we have decided to follow it up with another. At this rate, in a week’s time we will be shooting to answer “Where are the Snowdens of yesteryear?”.
We all want to know how many questions does one need to attempt in order to get a certain percentile. In fact, we want to know just some number, some arbitrarily concocted number, just some number – and we want this number so badly that we care not about the veracity of said number. So, here goes –
The numbers indicate correct answers/attempts. For instance, in order to get 90th percentile in DI-LR, one should shoot to attempt 15 questions and get 12 of these correct.
Now, on to the assumptions
Without assumptions, what is the basis of all analysis? Those fungible, flexible, stretchable, delicious things called assumptions are the vital cog in the modern economy. We would not want a crucial table to be created without those assumptions, would we?
- We have assumed DI-LR would be slightly easier in 2016 than it was in 2015 and that Verbal would be slightly tougher. Why? We like to use the term “regression to mean”. But you could have easily said “because we can”
- In an ideal world, there is nothing called as Target Hit rate or target accuracy ( target hit rate should always be 100%) , but we suspended the idea of the ideal world a while ago. So, we have accounted that students factor this kind of accuracy variable in.
What are the real take-aways here? Enough of snark
If you want a higher percentile, make sure your error rate is close to zero. From 95th to 99th percentile, attempts go up very little, improved accuracy contributes most of the leap. Anyone who is ambitious enough to think of 99th percentile and beyond has to be absolutely pig-headed about accuracy. All this talk of 70% hit rate is humbug. Would you go in a bus if the driver had a 70% accident-free rate?
If someone says that he/she is confident of 70% accuracy, he/she is talking out of his/her hat. Which 70%? Why did you just not leave the other 30%? If you are clueless about which 70%, might it be lower? Could it go as low as 20%? What are you really sure of? are some of the good questions that can be asked of the student. If you are part of the brigade happy with a lowly 70% accuracy, ask yourself these questions, and then relentlessly chase 100% accuracy.
What about the overall percentile?
80th percentile in each section should see the overall percentile close to 90. 90 in each section should take you to 95 overall. 95 in all three will take you to 98-ish overall. 97-ish in all three will fetch an overall score of 99.xx. 99 in each section should see you in the 99.6+ range. Average of the three percentiles and a healthy boost to that would usually take you to the overall percentile.
Best wishes for CAT preparation. Bear in mind that these are just broad targets to give yourself an anchor prior to the exam. Do not take these two seriously. Approach the exam with an open mind.