Smart Guesswork

The two-minute CAT maxing strategy.

First of all, be warned: nobody can "max" the CAT. Not the brainy-heads. Not the frazzle-heads. Not you either. The test has a time limit, and the limit is real, no matter how good you are.

The test last year had 123 questions, "humanly impossible" to do in the two hours given, in the words of Sandeep Manudhane, chairman and managing director, PT Education, a CAT coach. That has always been the case. The surprise last year was differential marking.

Some questions were worth half a mark, some 1 and some 2. Complicates things, doesn't it? Without a strategy, you can kiss your as... er, astronomical starting salary goodbye.

As luck would have it, Napolean did us a big favour more than 200 years ago (okay, in a roundabout way). He formulated a thumbrule to avoid paralysis-by-analysis: even if you have just two-thirds information, go ahead. You'll emerge victorious.

You don't exactly have to dedicate yourself to that proposition. But at crunch moments, smart guesswork could pay handsomely. "If you look at the reality of life and business," says Gautam Puri, managing director, Career Launcher India Ltd, another CAT coach, "we're forced to take decisions with incomplete information, and CAT simulates reality as faced in life and business."

The operative word there is "smart". Here it means knowing basic probability like the back of your hand. Now, CAT is a multiple-choice test with four options to choose from by way of an answer to each question.

If you fill in "Answer A" for all the questions, you could reasonably assume that one-fourth are going to turn out correct. But you obviously can't expect to make it to GE's corner office doing that, right? ... right? Right?

Heck, anyhow. To get back. Guesswork is part of the economy. It isn't free. If a correct answer gets you one point, an incorrect answer imposes a penalty of one-third "" as estimated by leading CAT coaches (it's an estimate because CAT gives out scorecards but not the actual negative-marking fraction).

Now, what if you go random guessing the entire test sheet? Over a series of questions you attempt this way, the law of expected value should kick in. Your expected gain is the probability of success (one-fourth) multiplied by the gained value (total potential score points), and your expected loss is the probability of failure (three-fourths) multiplied by the loss (one-third the total score). The difference is what you end up with. Do the math here: it's bad news.

What's more, as Manudhane says, "You never know, CAT may actually follow progressive negative marking, which could mean a penalty of X on the first ten wrong answers and a penalty of X plus delta X on the next 10, and so on." Uh-oh.

So nobody guesses foolishly. The trick lies in gauging the test well. To score well, you must do most of it by working out the actual answer. You first do the stuff you know for sure, and then do some good guessing. If you guess smart, the equation turns in your favour.

In theory, if you can narrow your choice of answers down to three, you can scrape into the positive with guessing. This is still risky, though. It could go wrong, since the number of guessed answers has to be sufficient (the sample) before the rules of probability actually kick in.

The solution: narrow your choices even more. What coaches such as Puri and Manudhane recommend is the striking off of two options, effectively reducing your failure likelihood to half, before making a guess. "We call it zeroing in on the likely answer using the elimination of options technique," says Manudhane.

It's a huge time saver, adds Puri, which is all the more important in a test that's unfinishable the proper way. So go ahead, if a question is taking too much time, strike off two, and toss your choice.

"This can be done quite easily within around 20 seconds for a quantitative analysis or data interpretation question," says Puri, "which saves a lot of time, since solving it could take maybe two minutes or so." Some eliminations are fairly simple. For instance, a question where the answer simply has to be a multiple of four at a glance.

The technique can be used all through, regardless of how many marks the question is worth (variable marking alters the what-to-attempt aspect of your strategy, not how-to-attempt).

"In English too, in critical reasoning, you can identify two obviously wrong choices much faster than the exactly correct one," says Puri. But as a principle, it's best to go into rapidfire guessing mode only towards the slog end. Say, in the last seven minutes available.

But don't overdo it. Progressive negative marking is always a fear. Guesswork is just part of an overall strategy to maximize your score. To give you an idea of what works well, consider what the four toppers of CAT 2004 did. They did between 80-110 questions of the 123 total, using smart guesswork on some 10-15 per cent of those. It did the trick.

But then, the CAT can be tricky too. It could spring yet another surprise this year. So it's terrific if you've read Peter Bernstein's classic on risk. But the two minutes you spend in strategic thought before nibbing forth your pencil may make all the difference.

Read carefully "" for shifting odds. You never know, you might have to get wiser by the minute. That's the thrill of it, isn't it... discovering differential penalties? Whatever. And, hey, best of luck.