Dec 21 And this column is one I've wanted to write for a long time - one that challenges the easy use of numbers like "trillion", quadrillion", "bajillion" and the like.
"Quadrillion" got a few moments of fame recently because part of Donald Trump's efforts to overturn the US election result used that number'. I'm not certain those who quoted it fully understood it, or the particular way they were using it, so this column is for them as well. How many zeroes in one quadrliion: https://www.livemint.com/opinion/columns/how-many-zeroes-in-one-quadrillion-11608225027611.html Let me know how many zeroes you come up with. cheers, dilip --- How many zeroes in one quadrillion? One day last week, someone urged me to watch a short Youtube clip which features that BJP veteran of a thousand crummy but loud TV appearances, Sambit Patra. In this particular clip, Patra pronounces that familiar mantra: India will soon become a 5-trillion dollar economy. Another guest on the show immediately asks Patra to reveal to the audience how many zeros there are in a trillion. Patra's stumbling evasions tell the tale: he has no clue. He uses the number to make his grand pronouncements, but he has no clue. Oddly enough, that same day I heard the word "quadrillion" tossed about. This time it was from the other side of the world, courtesy Donald Trump's press secretary, Kayleigh McEnany. She was quoting the latest Trumpian effort to overturn his election loss to Joe Biden, a lawsuit in the US Supreme Court by the Attorney General of Texas. Now Texas is a state I called home, and happily, for eight fine years. But if this is the calibre of its highest elected legal authority today, I'm not sure I want to mention that residency aloud any more. The suit speaks of four other states that ended up in the Biden column in the recent election, even though the early counting had given Trump leads in all four. The chance of this turnaround happening in any one state, the suit claimed - and McEnany trumpeted - was about one in one quadrillion. Thus the chance of it happening in all four states was - McEnany trumpeted again - one in one quadrillion to the fourth power. Mind-boggling numbers, and what they amount to is the assertion that it's impossible that Biden won those states. Therefore, dear Supreme Court, won't you please overturn the election? Let's examine this quadrillion number a bit, though. The lawsuit is loosely based on a study - if that’s the right word - of voting patterns in those states in the 2016 and 2020 elections. It makes two broad assumptions. First, remember that in 2016 Clinton won 46% of the vote in Georgia (picking one of the four) and lost the state because Trump got 51%. The suit assumed that the party-wise breakup of eligible voters in 2020 follows that 2016 distribution: 46% are Democrat, 51% are Republican. Now about five million of these Georgia voters actually voted this year, and Biden’s vote share among them was higher than Clinton’s in 2016. (Higher, too, than Trump’s in 2020.) Which, for Trump faithful, raises this question: if we make the assumption above, what's the chance that Biden gets half of the votes, or 2.5 million? (Slightly more than half, really, because he won the state. But to make this easier, let’s leave aside that detail). To better understand this dilemma, try a couple of thought experiments. One, imagine you have a drawer crammed to the brim with a jumble of hundreds of socks. All you know is that 90% of them are black, 10% white. With your eyes closed, you reach in and pull out two socks. What's the chance that you'll have in your hand a white pair? Intuitively, you know that it's pretty unlikely. Two, think of this. An all-girls’ school has recently started admitting boys as well. But in a student body of 1000, they still have only ten boys. The principal wants to choose six students at random to make a team for an interschool contest whose only stipulation is that exactly half the team should be boys. How likely is it that her randomly-chosen team will meet this requirement? Not very likely again, you’ll agree. With somewhat different numbers, this is analogous to what the lawsuit sought to persuade the Supreme Court and the rest of us about Georgia. Indeed, if the state’s voters are split between the parties as they voted in 2016 - 51% Republican to 46% Democrat - it is extremely unlikely that Biden would get half of the 5 million votes cast in 2020. How unlikely? A chance of one in a quadrillion sounds about right. Hold on to that while we take a look at something else in the suit. Specifically, the second broad assumption. Trump did have a lead in Georgia at the end of 3 November, Election Day. Most of the in-person votes - the ones cast that day - had been counted by then, and a majority of those favoured Trump. But what had not yet been counted at the end of the day were all the ballots that had been mailed in. The suit assumed that those would follow the same distribution as the already-counted in-person votes; that is, it assumed that the same majority of those who would choose to vote by mail leaned Republican and thus to Trump. Yet when the mailed-in ballots were counted, a majority was for Biden; enough of a majority, in fact, to erase Trump’s lead and hand Georgia to Biden. Again, Trump faithful asked the question: if we make that assumption about the preferences of those who might mail in their ballots, what’s the chance that when they are counted, Biden gets substantially more than half? The same thought experiments apply, and the same answer applies: it’s extremely unlikely that Biden would get such a majority of the mailed-in ballots. Again, one in a quadrillion sounds right. So with these assumptions, the suit made the reasonable-sounding claim that it is impossible that Biden could have won Georgia. And yet, that’s just what he did. Therefore, it suggested, there was fraud on a massive scale. Or was there? Could there be another explanation instead? For example: might the assumptions themselves be mistaken? Take the one about the mailed-in ballots first. For months leading up to the election, Democrats had urged voters to vote early, by mail — partly so as to avoid personal contact of any kind on 3 November, partly to ensure their vote was cast and would count. In contrast, Trump himself had spent months alleging fraud in the ballot-by-mail system and urging his voters to vote in-person on 3 November. Therefore, it made no sense to assume the Republican-Democrat split was the same with both kinds of ballots. This is why many election watchers had for months predicted that Trump would likely have a lead with in-person votes, but that it would shrink when the mailed-in ones were counted. That’s exactly what happened. As for the other assumption: well, as more than one election observer has pointed out, voters’ political preferences do indeed change between elections. This is, after all, why countries hold elections in the first place. If their preferences didn’t change, we could as well hold an election once in a nation’s history and keep replicating that result. In our country, for example, that would mean the BJP and other parties could never hope to come to power. Because the Congress won the first general election in 1951-52, the Congress remains in office forever. That seems absurd to you? That’s exactly how absurd the Texas Attorney-General’s lawsuit was. And luckily nobody in India has (yet) propounded such an absurdity. Luckily too, the US Supreme Court refused to hear this absurdity. Yet the Texas AG also knew: wrap absurdity in a number like “one quadrillion”, and even “one quadrillion to the fourth power”, and you’ve produced material that the McEnanys and others like her - a Patra, perhaps? - will spout with abandon, without question. Which might raise still another question: how many zeroes in a quadrillion? -- My book with Joy Ma: "The Deoliwallahs" Twitter: @DeathEndsFun Death Ends Fun: http://dcubed.blogspot.com -- You received this message because you are subscribed to the Google Groups "Dilip's essays" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web, visit https://groups.google.com/d/msgid/dilips-essays/CAEiMe8q43X%3Dhm%3Dom7dBjcpApekiSLHwUnnSi3SXsyK%2BczZeaHA%40mail.gmail.com.
