Heisenberg's principle and Monte Carlo algorithms

When you have a more random process, you might (possibly counterintuitively) be more gritty - that's the only way to get value out of your process

Now I have two data points when it comes to babies. So I can generalise and concoct theories, without being accused that I’m extrapolating with a sample of one.

The son is now six months and a bit old. Sometime a week ago he sat. He is now in the process of learning to crawl. While he can get himself into what my daughter calls a “table position” (all fours) he is yet to figure out that he can move opposite arms and legs together. And so he struggles, trying to take one step at a time, and then just sitting down.

He also happens to be a quantum particle. Observation changes his behaviour. If you leave him alone, he behaves one way, exploring stuff and maybe making progress. The moment he knows that you are watching him, he becomes self-conscious and changes his behaviour, either trying to impress or deciding to play safe. So when he is up to his tricks (and he is incredibly naughty - much naughtier than my daughter was at the same age), we need to try and watch him without letting him know that he is being watched.

Now, speaking of the two data points I have, one thing we have found is that our son and daughter, based on observations so far, learn and do things rather differently.

The most obvious difference was in the way they learnt to get off their respective bouncers. My daughter was over five months old when she achieved this process, and she did so in an extremely careful manner.

She would keep bouncing herself, flip around to hold the side of the bouncer, and then ease herself down facing the bouncer, and slipping off the bouncer in an extremely risk-free manner. After a couple of times, we even stopped watching her when she did this, and would peacefully leave her on the bouncer, confident she won’t hurt herself even if she were to get off.

The boy is not like that. He started attempting to get off the bouncer much earlier, and he did it in a more direct manner. He would try to simply sit up from his position on the bouncer, and get off the bouncer straight! No care about risk or danger. All he wanted to do was to just get off.

And so he would keep trying. And so we would keep tying him up in the bouncer, for we knew that if we had to untie him we needed to keep a close watch. I’ve lost count of the number of times when he’s headed straight for a fall, and I’ve had to catch him to prevent him from hurting himself.

Rule 11 of Jordan Peterson’s 12 rules for life talks about how boys inherently want to take risk (we’ve evolved that way), and if you take away one means for them to take risk, they will take risk in another (and possibly more dangerous) way. What I had not realised until I observed this guy is that this risk-taking behaviour starts rather early!

In any case, there is also a difference in the way my daughter and son learnt to sit. Again, the daughter was more careful, making sure she had figured out a foolproof way of sitting and balancing herself before even attempting to sit. The son is far more cavalier, trying to sit up at a time when his neck was barely stable, and getting himself up into a sitting position at a time when he was barely able to balance himself!

And so we have had to do the risk management for him. Every time he has sat, during the last few days, one of us has rushed to put a pillow behind him, lest he fall on his head. This hasn’t prevented him from hitting his head multiple times, though - on Friday morning, for example, he suddenly decided he will crawl, one of his elbows gave way and down he went. He bawled. I picked him up and rubbed his head. And there he was, trying to crawl once again.

So what we have here is one person with a very careful, deliberate and risk-free process, and another who is far more risk-loving.

The related pertinent observation is - my son tries much much harder to do what he has to do than my daughter did at a comparable age. It is almost as if he is aware that the method that he has is stochastic, and the only way he can get some results out of it is to apply it a large number of times. My daughter, comfortable in that her method is foolproof, tries once, and if she fails, doesn’t try much more!

In computer science, my son’s approach is like randomised (specifically, Monte Carlo) algorithms, like the ones used for primality testing. You know your algorithm is random, and so the only way to get any value out of it is to apply it multiple times. And so you don’t care about the result of one instance of the algorithm and keep applying it, and the ensemble of all these applications gives you the answer you need.

My daughter’s approach is like a deterministic algorithm. You know the outcome is fairly certain, and so if it doesn’t work the one time you apply it, you just move on.

In finance, their approaches are like a volatile stock and a bond. When you know you are holding a volatile instrument, you have no choice but to work hard to hedge and thus manage your risk. When your asset is less volatile, you can be more chill in terms of hedging - you know it can’t hurt too much.

The only worry there, possibly based on how I have turned out, is that sometimes when you have an asset that seems to have very low volatility, there can be some hidden tail risks - which are less likely to be hidden in highly volatile instruments.

Anyway, as I continue to observe them (and in a manner that doesn’t affect their behaviour), I’ll write more on this here.