Thread by @Sanjay__Bakshi

So many lessons here... Here's my interpretation of a couple of them:

https://twitter.com/iancassel/status/1141883784847081473

"My largest positions aren’t the ones I think I’m going to make the most money from. My largest positions are the ones where I don’t think I’m going to lose money. — Joel Greenblatt"

This is logical. We have a tendency to focus on the UPSIDE. We estimate per-share intrinsic business value. We estimate expected returns. Both measures focus on the expected outcomes for which both upside and downside have been considered.

However, what Ian is saying here (based on Greenblatt's quote) is that when it comes to position sizing, you should put the MOST money in that business which has the LOWEST downside.

It's logical because the objective is not to maximise returns. Rather, the objective should be to maximise returns per unit of risk.

In other words, to find stocks that qualify for INCLUSION in a portfolio, use intrinsic value or expected returns. But when it comes to POSITION SIZING focus on the downside.

Theoretically speaking, all businesses have a worst case scenario of a 100% loss. So what do Greenblatt and Ian mean when they say that for position sizing downside risk matters? Isn't the downside risk for every stock the same -100%?

To see the way out of this problem, read the Greenblatt quote again:

"My largest positions are the ones where I don’t think I’m going to lose money."

He is saying that while in THEORY all stocks can go to zero, on a PRACTICAL level some have a far more chance of doing that.

Which stocks have far more downside risk? Those whose business models are fragile. A very highly leveraged company, asset-liability mismatch, DEPENDENCE on one product, technology, plant, vendor, government support etc etc.

If the source of fragility materialises, the business can lose almost all its earnings. So the key question to ask here is that is there something that can DEVASTATE earning power? If so, then if you still like the stock, keep position size low.

In contrast, if a business qualifies for inclusion in a portfolio but has MULTIPLE sources of UNCORRELATED earnings streams so that any given event (earthquake, fickle fashion, adverse credit markets, withdrawal of govt support) won't DESTROY earnings, put more money in it.

Here's another point that Ian makes. He asks:

How many of your successful investments occurred in the timeframe you originally predicted? How many times do you ever remember saying, “Well that happened a lot sooner than I thought?”

And he answers for himself:

Exactly. Not often enough. A majority of my successful investments took longer, in many cases a lot longer than I originally expected.

This is so true. At least it has been for me.

Bill Gates once said: “Most people overestimate what they can do in one year and underestimate what they can do in ten years."

I think Gates' words are particularly relevant for investors. We investors are too optimistic about the businesses we love. Even when we are right about the businesses, we turn out to be wrong about the timing.

Almost NEVER do they deliver earlier than expected. I am not referring to STOCK PRICE here. Rather, I am focusing on EARNINGS. Stock price has a component - the P/E multiple, which is not in our control. But earnings expectations are.

We expect the businesses we love to do wonderful things FASTER than they will actually do. We are right about the QUANTUM of earnings but wrong about the TIMING. And of course that affects both intrinsic value and expected return estimates.

I do think that we should make adjustments for this type of over-confidence factor in our models - explicit ones in excel or the ones in our head...