Start the schools already

Irrespective of when you open the schools, there will be a second wave at that point in time. So we might as well reopen sooner rather than later and put children (and parents of young children) out of their misery.

OK, I admit I have a personal interest in this one. Being a double income, single kid, no nanny, nuclear family, we have been incredibly badly hit by the school shutdown for the last nine months. The wife and I have been effectively working at 50% capacity since March, been incredibly stressed out, and have no time for anything.

And now that I’ve begun a “proper job”, her utilisation has dropped well below 50%. This can’t last for long.

Then again, this post is not being driven solely by personal agendas or interests. The more perceptive of you might know that on my twitter account, I publish a bunch of graphs every morning, based on the statistics put out by . And every day, even when I don’t log into twitter, I go and take a look at the graphs to see what’s happening in the country.

And the message is clear – the pandemic is dying down in India. It is a pretty consistent trend. The Levitt Model might not really be true (my old friend’s comment that it is “random curve fitting” when I first came across it holds true, I would think), but it gives a great picture of how the pandemic has been performing in India. This is the graph I put out today.

In most states in India, the Levitt measure is incredibly close to 1, indicating that the pandmic is all but over. However, you might notice that the decline in this metric is not monotoniuc.

However, if you look at the Delhi numbers on the top right, notice how nicely the Levitt metric shows the three “waves” of the disease in the city. And you can see here that the third wave in Delhi is all but over. And while you see the clear effect of Delhi’s third wave in the Levitt metric, you can also see that it coincided with a second wave in Haryana, and a (barely noticeable) second wave in Uttar Pradesh and Rajasthan.

This wave was due to increased pollution, primarily on the account of crop burning in Punjab and Haryana in October-November. The reason the second waves in Uttar Pradesh and Rajasthan (as seen in terms of the Levitt measures) were small is that they are rather large states, and the areas affected by the bad pollution was fairly small.

And along with this, consider the serosurveys in Karnataka (both the government one and the IDFC-sponsored one), which estimated that the number of actual infections in the state are higher than the official counts of infections by a factor of 40 to 100 (we had initially assumed 10-20 for this factor). In other words, an overwhelmingly large number of cases in India are “asymptomatic” (which is to say that the people are, for all practical purposes, “unaffected”).

In other words, we know cases only when someone is affected badly enough to get themselves tested, or has a family member affected badly enough to get themselves tested. And what happened in Delhi and surrounding states in October-November was that with higher pollution, everyone who got affected got affected more severely than they would have otherwise.

Some people who might have otherwise been unaffected showed symptoms and got themselves tested. Some people who might have not been affected seriously enough ended up in hospital. Pollution meant that some people who might have recovered in hospital ended up dying. And as the crops finished burning and pollution levels dropped, you can see the Levitt metric dropping as well.

And lest you argue that I’m making an argument based on a mostly discredited metric, here is the actual number of known cases in the most affected states in the country. The graph is a Loess smoothing, and the points can be seen here.

See the precipitous decline in Delhi (green line) and Karnataka (orange) and Andhra Pradesh (pink) in the last couple of months. The pandemic has pretty much burnt through in most states. We can start relaxing, and opening schools.

You might be tempted to ask, “but won’t there be a second wave when schools reopen?”. That is a very fair concern, since people who have so far been extremely conservative might relatively relax when the schools open. The counterpoint to that is, “irrespective of when you open the schools, there will be a second wave at that point in time“.

It doesn’t matter if we reopen the schools now, or in April, or in August, or in next December. There will always be a few vestigial (possibly unaffected) cases going around, and there will be a spike in known cases at that point. And by quickly dialling up and down, we can control that.

So I hereby strongly urge the state governments (especially looking at you, Government of Karnataka) to permit schools to reopen. A few vocal and overly conservative parents should not be able to hold the rest of the country (or state) to ransom.

The Prom

The other day, the wife and I were discussing about growing up, and about school crushes, and how relationships worked in school. It was a fascinating discussion, and it has already led to an excellent newsletter episode by her. Here is the key point of our discussion, as she wrote in her newsletter:

There are rumours that some boys have a crush on a couple of girls. You think that it’s a pandemic like the COVID-19, and it’s going to get us all, except it doesn’t. This unfortunately follows a power law, only a couple of boys and girls will be affected by the “crush”, the rest of us just have to be affected by the lack of – crushes, bosoms and baritones. Now, the problem with middle/ high school is that it operates on mob mentality – everyone is only allowed to have a crush on the crushable.

And then later on in the piece, she talks about proms.

You are most likely to fall in love organically and benefit from it early on in life. So, wasting these precious years of socialising is a sin.

So, when I think about it, “prom” is a great concept. It gives everyone a shot at gaining some experience. You’re better off going to prom at 16 rather than at 26.

This got me thinking about proms. I had no clue of the concept of a “prom” while growing up, and only came to know of it through some chick flicks I watched when I was in my late teens. However, I ended up writing about proms in my book (while describing Hall’s Marriage Theorem – yes, you can find Graph Theory concepts in a book on market design), and the more I think about it, the more I think it is a great concept.

The thing with proms is that it forces a matching. One on one. One boy gets one girl and vice versa (I really don’t know how schools that don’t have a balanced sex ratio handle it). And that is very different from how the crush network operates in middle and high school.

As Pinky described in her post, crushes in middle and high school follow a power law, because there is strong mob mentality that operates in early puberty. Before “benefits” get discovered, one of the main reasons for having a boyfriend/girlfriend is the social validation that comes along with it, and such validation is positive if and only if your peer group “approves” of your partner.

So this leads to a “rich get richer” kind of situation. Everyone wants to hit on the hottest boys and girls, with the result that a small minority are overwhelmed with attention, while the large majority remains partnerless. And they continue to be partnerless this way, friendzoning large sets of their classmates at an age that is possibly most suited for finding a long-term gene-propagating partner. 

In most Indian schools, the crush graph in high school looks like this. The boys and girls towards the bottom are the “long tail” – they are not cool to hit on, so nobody hits on them. In other words, they are unloved in High School. Notice that it’s a fairly long tail.

Also notice that most of the arrows point upwards (I’ve drawn the graph so the most sought-after people are on top). Because nothing prevents “one way crushes”, everyone just tries “as high as they can” to find a partner. And most of these don’t work out. And most people remain unloved.

So what does a prom do? Firstly, everyone wants to go to the prom, and to go to a prom, you need a date. Which means that everyone here in this long tail needs a partner as well. In the original setup, when crushes were based on mob-mentality, there was no concept of seeking “undervalued assets” (people nobody else is hitting on). Now, when everyone needs a unique partner, there is value to be found in undervalued assets.

Basically a prom, by providing immediate rewards for finding a partner (soon enough, the kids will discover other “benefits” as well), moves the schoolkids from a “crush network” to a “partner network”, which better represents real-world romantic networks.

Many people may not be able to pair with their first choice (notice in the above network that even the most sought after people may not necessarily match with their favourites), but everyone will get a partner. The Gale Shapley (or should I say Shapely Gal?) algorithm will ensure a stable matching.

Moreover, it doesn’t help your cause in getting a preferred (if not most preferred) partner for the prom if you make your attempt just before the prom. You need to have put in efforts before. This means that in anticipation of the prom, “pair bonding” can happen much earlier. Which means that schoolkids are able to get trained in finding a partner for themselves much earlier than they do now.

That will make it less likely that they’ll bug their parents a decade (or two) later to find them a partner.