Bayes Theorem and Respect

Regular readers of this blog will know very well that I keep talking about how everything in life is Bayesian. I may not have said it in those many words, but I keep alluding to it.

For example, when I’m hiring, I find the process to be Bayesian – the CV and the cover letter set a prior (it’s really a distribution, not a point estimate). Then each round of interview (or assignment) gives additional data that UPDATES the prior distribution. The distribution moves around with each round (when there is sufficient mass below a certain cutoff there are no more rounds), until there is enough confidence that the candidate will do well.

In hiring, Bayes theorem can also work against the candidate. Like I remember interviewing this guy with an insanely spectacular CV, so most of the prior mass was to the “right” of the distribution. And then when he got a very basic question so badly wrong, the updation in the distribution was swift and I immediately cut him.

On another note, I’ve argued here about how stereotypes are useful – purely as a Bayesian prior when you have no more information about a person. So you use the limited data you have about them (age, gender, sex, sexuality, colour of skin, colour of hair, education and all that), and the best judgment you can make at that point is by USING this information rather than ignoring it. In other words, you need to stereotype.

However, the moment you get more information, you ought to very quickly update your prior (in other words, the ‘stereotype prior’ needs to be a very wide distribution, irrespective of where it is centred). Else it will be a bad judgment on your part.

In any case, coming to the point of this post, I find that the respect I have for people is also heavily Bayesian (I might have alluded to this while talking about interviewing). Typically, in case of most people, I start with a very high degree of respect. It is actually a fairly narrowly distributed Bayesian prior.

And then as I get more and more information about them, I update this prior. The high starting position means that if they do something spectacular, it moves up only by a little. If they do something spectacularly bad, though, the distribution moves way left.

So I’ve noticed that when there is a fall, the fall is swift. This is again because of the way the maths works – you might have a very small probability of someone being “bad” (left tail). And then when they do something spectacularly bad (well into that tail), there is no option but to update the distribution such that a lot of the mass is now in this tail.

Once that has happened, unless they do several spectacular things, it can become irredeemable. Each time they do something slightly bad, it confirms your prior that they are “bad” (on whatever dimension), and the distribution narrows there. And they become more and more irredeemable.

It’s like “you cannot unsee” the event that took their probability distribution and moved it way left. Soon, the end is near.

The Base Rate in Hitting on People

Last week I met a single friend at a bar. He remarked that had I been late, or not turned up at all, he would have seriously considered chatting up a couple of women at the table next to ours.

This friend has spent considerable time in several cities. The conversation moved to how conducive these cities are for chatting up people, and what occasions are appropriate for chatting up. In Delhi, for example, he mentioned that you never try and chat up a strange woman – you are likely to be greeted with a swap.

In Bombay, he said, it depends on where you chat up. What caught my attention was when he mentioned that in hipster cafes, the ones that offer quinoa bowls and vegan smoothies, it is rather normal to chat up strangers, whether you are doing so with a romantic intent or not. One factor he mentioned was the price of real estate in Bombay which means most of these places have large “communal tables” that encourage conversation.

The other thing we spoke about how the sort of food and drink such places serve create a sort of “brotherhood” (ok not appropriate analogy when talking about chatting up women), and that automatically “qualifies” you as not being a creep, and your chatting up being taken up seriously.

This got me thinking about the concept of “base rates” or “priors”. I spent the prime years of my youth in IIT Madras, which is by most accounts a great college, but where for some inexplicable reason, not too many women apply to get in. That results in a rather lopsided ratio that you would more associate with a dating app in India rather than a co-educational college.

In marketing you have the concept of a “qualified lead”. When you randomly call a customer to pitch your product there is the high chance that she will hang up on you. So you need to “prime” the customer to expect your call and respond positively. Building your brand helps. Also, doing something that gauges the customer’s interest before the call, and calling only once the interest is established, helps.

What you are playing on here in marketing is is the “base rate” or the “prior” that the customer has in her head. By building your brand, you automatically place yourself in a better place in the customer’s mind, so she is more likely to respond positively. If, before the call, the customer expects to have a better experience with you, that increases the likelihood of a positive outcome from the call.

And this applies to chatting up women as well. The lopsided ratio at IIT Madras, where I spent the prime of my youth, meant that you started with a rather low base rate (the analogy with dating apps in India is appropriate). Consequently, chatting up women meant that you had to give an early signal that you were not a creep, or that you were a nice guy (the lopsided ratio also turns most guys there into misogynists, and not particularly nice. This is a rather vicious cycle). Of course, you could build your brand with grades or other things, but it wasn’t easy.

Coming from that prior, it took me a while to adjust to situations with better base rates, and made me hesitant for a long time, and for whatever reason I assumed I was a “low base rate” guy (I’m really glad, in hindsight, that my wife “approached” me (on Orkut) and said the first few words. Of course, once we’d chatted for a while, I moved swiftly to put her in my “basket”).

Essentially, when we lack information, we stereotype someone with the best information we have about them. When the best information we have about them is not much, we start with a rather low prior, and it is upon them to impress us soon enough to upgrade them. And upgrading yourself in someone’s eyes is not an easy business. And so you should rather start from a position where the base rate is high enough.

And this “upgrade” is not necessarily linear – you can also use this to brand yourself in the axes that you want to be upgraded. Hipster cafes provide a good base rate that you like the sort of food served there. Sitting in a hipster cafe with a MacBook might enhance your branding (increasingly, sitting in a cafe with a Windows laptop that is not a Surface might mark you out as an overly corporate type). Political events might help iff you are the overly political type (my wife has clients who specify the desired political leadings of potential spouses). Caste groups on Orkut or Facebook might help if that is the sort of thing you like. The axes are endless.

All that matters is that whatever improved base rate you seek to achieve by doing something, the signal you send out needs to be credible. Else you can get downgraded very quickly once you’ve got the target’s attention.