We spend our time trying to find disruptors that can become secular winners which we can hold for many years. Oftentimes, we are drawn to companies whose outcomes are still somewhat debated because they are earlier in their journeys of establishing market dominance. For example, Carvana before it was clear they had the most profitable business model in auto retail. Cogent before the wavelength business generated significant new business. Not everything has to be so debate-worthy - we are also happy to own companies whose paths are pretty clear and our estimates require continued solid execution into healthy and growing end markets but the total opportunity is underappreciated, like Royal Caribbean.

I’ve spoken to many other investors and analysts on Cogent after they discovered some of our other Insights on the company. Common feedback I hear on it is: “I can’t figure out if I should buy it now, or if I should wait until it’s clear wavelengths are working.” I completely empathize with this internal debate. I have it with myself often, which was the inspiration for this article.

This post will discuss the challenges and opportunities of being early vs. waiting for so-called certainty.

Cogent Case Study

Let’s start with our goal: to own high-quality companies for many years, allowing them to generate significant value. Let’s frame a common decision we face when we find earlier-stage companies that are interesting.

Let’s look at Cogent, a live example in our portfolio that we have written about at length. If the wavelength business materializes as we expect, Cogent will generate over $10/share of FCF and likely will trade above $150/share within 3 years. If it goes really well, shares could trade well above that level.

These are the scenarios we face when thinking about investing now or waiting until we know more:

1. Buy the stock now at around $50. If it goes well, we will make 200% on the stock, plus $20 of dividends = about 240% total return.

OR

2. Wait for proof on wavelengths. Once we have sufficient proof, the price likely will have moved substantially and we may be able to buy the stock at $75. If so, we would make 100% on the stock plus a little less than $20 of dividends (because we waited), or about a 125% total return.

Would we rather have 240% potential return while uncertainty is higher, or 125% potential return once there has been stronger fundamental validation of our thesis?

Scenario 1 would be about a 50% IRR including cumulative dividends. Scenario 2 would be about a 30% IRR including cumulative dividends. Both IRRs are great and should be vastly superior to index-level returns. It’s hard to blame anyone for wanting to wait, but that doesn’t mean it’s an optimal way to invest in this type of scenario.

Of course, we should be rigorous about risk-adjusting these scenarios.

If we invest in scenario 1 but are wrong, we might lose another 20-30% on Cogent. For other companies earlier stage situations without valuable fallow assets to protect the downside case (which Cogent has), we might lose 50%.
If we invest in scenario 2 but our thesis validation turns out to be a head fake, we might lose 30-40% as “thesis change” capital flees the story. In other companies with less dispersion going forward but we got a growth catalyst wrong, we may lose 20-30% by the time it was clear we were wrong.

The better path really depends on how we weigh our probability of being right. But rather than evaluate the quality of our research on a single company or single situation, it makes sense to think about it statistically across larger sample sizes.

If each such situation were a coin flip, with50/50 odds, what should we do on average?

Scenario 1 Expected Return:

240% upside x 50% - 25% downside x 50% = 120% - 12.5% = 107.5% expected return.

Scenario 2 Expected Return:

125% upside x 50% - 35% downside x 50% = 45% expected return.

Expected returns overwhelmingly favor being early. However, this assumes the downside in Scenario 1 is lower than the downside in Scenario 2 because we are buying the stock significantly cheaper. For Cogent, this is rather unique because of its idiosyncratic fact pattern with substantial fallow assets.

What if we look at other types, where there’s more downside in Scenario 1 and less downside in Scenario 2?

Scenario 1: 240% upside x 50% - 50% downside x 50% = 95% expected return

Scenario 2: 125% upside x 50% - 25% downside x 50% = 50% expected return

When we analyze coin-flip probability distributions, the expected value of Scenario 1 still is substantially higher. Now, let’s see how how wrong would we have to be to be indifferent between Scenario 1 and Scenario 2.

50% = 240% upside x N – (1-N) x 50%, solve for N, which is the probability of being right.

Our indifference point is about when N = 35%. So we have to be wrong more than 65% of the time to choose Scenario 2 and come out ahead.

Despite pretty compelling mathematical reasoning to choose Scenario 1 when confronted with these situations, I would guess that most fundamental investors choose Scenario 2 most of the time. We believe they often choose the sub-optimal path because most investors prefer the psychological comfort of feeling like they “know” the outcome and they do not properly assess their probabilities of being right and wrong, and because they may not want to sell other ideas to “wait and see.” Additionally, most stock-picking analysts and firms are afraid of looking wrong or dumb to their bosses or to their clients, so they have a bias to look “right” more often by committing capital later, even if it’s at the expense of higher returns.

Readers of our materials hopefully would agree that our process is strongly oriented around trying to see around corners and conduct research that raises our confidence in future outcomes well above 50%. That is true, but investing involves uncertainty across multiple dimensions all the time. We can’t “know” anything with certainty. As such, at Recurve we have a preference for finding known uncertainties, digging on them as much as possible, and figuring out if we can be right often enough to make the math above work for our returns and for our partners. We make our assessments, size our bets, and sometimes we will be wrong - and that’s okay. If we aren’t wrong some of the time, we aren’t finding interesting enough situations where nailing a fundamental thesis can be rewarded with significant returns.

This topic reminds me of Jeff Bezos’ quote from an interview last year which really resonated with me:

“I think it’s generally human nature for humans to overestimate risk and underestimate opportunity.”

— Jeff Bezos

Bezos mentioned this in an entrepreneurial context, but I believe it also holds true for investors. Equities are inherently uncertain. Even some of the biggest companies in the world like Google and Apple have significant investment uncertainties that can materially swing their multiples and their terminal values. It’s better to recognize that we are all making a bet on the future that may or may not materialize, and it often pays to invest before the crowd feels comfortable with what it “knows” about needle-moving debates. We use fundamental research to tilt the odds even more in our favor, but it’s a tilting of odds - not an elimination of risk.

We haven’t mentioned position sizing yet, and we knowingly presented this as a binary choice: either invest now, or wait and invest later when we have more certainty. Oftentimes, we find that it makes sense to own a smaller position now and to be willing to buy more when we get the fundamental validation we are hoping for. This helps us maintain balance in the portfolio between early-stage bets we are making and more certain fundamental paths that have already materialized.

Closing

This article isn’t meant to be a rallying cry to take more risk on uncertain outcomes. Rather, I hope that it compels readers to think about the price they are paying to “know” some key information, and to serve as a reminder that, at times, it is perfectly rational to make a good, early bet - even if it turns out wrong in the end.