Over the past several days, Nate Silver has gone from baseball and political forecasting guru to full-blown media darling. Since correctly forecasting the winning presidential candidate in 49 and 50 states in the last two elections, Silver has been seen everywhere from NBC to Comedy Central. All of this culminated Wednesday when Silver won the ultimate modern-day achievement: trending Twitter topic. And while #DrunkNateSilver has already predicted the next two presidential elections and the ending to Star Wars 7, his sober counterpart has lessons that we can apply to our GMAT studying.
The PECOTA System- Finding historical similarities
Nate Silver first became famous outside of his consulting job when he developed PECOTA, which sought to predict the statistics and career arc of major league baseball players, as well as projected team win totals. PECOTA wasn’t the first projection system for baseball, but it was the first to use other players’ previous performances instead of that player’s trends. By comparing each player with 20,000+ player’s seasons since World War II, Silver was able to make a probabilistic distribution for individual players and their teams.
The Official Guide released its 13th edition this past summer. But even with 13 different editions of this GMAC study guide, the essential math and verbal skills that the GMAT has tested has not changed much over the years. Between the 12th & 13th editions, 83% of the questions did not change. And the GMAT Prep software that mba.com offers to experience an authentic GMAT has included the same bank of questions over the past six plus years. What does this mean? Even though the scenarios and numbers change, the basic concepts that you need to conquer the GMAT do not. Don’t focus on the trends on your test (the last three questions have been C!), but compare each question to questions that you have seen before. When you look at a question and think, this looks like the number properties question where two variables were multiplied together, you can most likely use similar math and problem solving skills to answer the new question in front of you.
The Signal and the Noise- What’s important and what’s not?
Any given Sunday, a gambler can make a convincing argument for both teams using statistics: Team X is winning 100% of the games this season in which they lose the coin-flip, but Team Y has won the last six games on the second Sunday of November. And while statistics like this probably won’t make too much of an impact on the outcome of the game, there are many important signals hidden within the superfluous noise that can be useful in predicting winners and losers.
The GMAT is filled with useless information- from the scenarios disguising a math problem to the paragraphs that don’t relate to the specific question beings asked in a Reading Comprehension passage to the large portion of non-essential information in many IR problems. The better you are able to eliminate the noise in a GMAT problem, the easier it will be to find what’s essential and use that to guide your answer selection. And just like Silver, who goes back and looks at his algorithms after any season or election, you need to be able to look back at questions you get wrong to determine what noise you mistook as important in order to make the next test even better than this one.
Making Models vs Finding Models- Choosing your battles before your test
In the weeks leading up to the 2012 election, Nate Silver encountered the predictable backlash to his findings from people who were not pleased with what he was predicting. People brought up everything from his self-admitted rooting for Obama to his small stature. But the difference between a statistician and a pundit is that the former decides what is important and uses that to find an answer while the latter decides an answer and then searches for facts that help the cause. Silver’s algorithms for the 2012 election were created before the election season and before knowing what each state poll would say. While other news outlets tried to decide which polls were most important in the middle of the election season, Silver knew his answer months before.
One of the few predictions that I will make about you, the reader, is that when you go and take your GMAT, you will get several questions wrong. The majority of you will also find yourself rushed at the end of either the quant or verbal section or possibly both. But while many students will spend the last two minutes of the test guessing on the remaining questions, their more prepared counterparts will have already utilized a plan to finish the test. If a student has been consistently running out of time on the last three questions, that student might decide to guess on three questions earlier in the test on question types that he or she has found to be consistently difficult for him or her. Rather than have the GMAT decide which questions you will be forced to guess on, let yourself make those decisions before you sit down.
Burritos or Politics- Knowing when to give up on a project
While he is probably best known for his political prognosticating, Nate Silver has used his statistical analysis for everything from online poker to finding the best night of the week to, uh, meet someone at a bar. But one of his most important pieces of research was never completed. During the quarter-finals of his Chicago Burrito Bracket, Silver stopped working on his burrito blueprint in order to… blog about politics. Probably a good career move.
If Silver ever needs another statistical project to work on, he needs only to search Twitter for his own drunken persona. But neither he nor you are going to be able to do everything, whether that is making algorithms or answering GMAT questions. There is a huge value in giving up on a project if another, better project presents itself. Likewise, the GMAT will offer 37 quant questions and 41 verbal questions for you to prove your genius. But you probably won’t be able to answer all of them in your limited time, so learn which problems are worth the invested time and which ones aren’t. Remember, if a question is hard after 2 minutes, it will still be hard after 4 minutes, so think about whether the time that you will spend to answer the question will be worth it.