The Theory of Poker by David Sklansky

Poker and Investing

The Theory of Poker by David Sklansky has brought a level of clarity to my investment process that it previously lacked.  Poker is a useful proxy for investing because it is a cocktail of luck and skill that involves decision making with incomplete information.  The obstacles to analyzing one’s investing ability systematically are that each investment’s starting conditions differ, the time limitations on analyzing so many investment opportunities let alone conducting a meta analysis, and the lack of immediate feedback.  The nuances of poker can be more easily analyzed systematically, but the underlying lessons can be transferred to investing.  Using two examples, I want to convey the depth of the comparability in a way that encourages you to read the book and make your own comparisons.

Sklansky’s Theory of Poker is,

Every time you play a hand differently from the way you would have played it if you could see all your opponents’ card, they gain; and every time you play your hand the same way you would have played it if you could see all their cards, they lose.

The goal of the investment research and portfolio management process is essentially to make sure you are playing as close to the way you would if you could see everyone’s hand.  In poker, Sklansky thinks the ability to read hands it the most important thing on the path to perfect play.

“When a good player makes a play, there is a sensible reason for it, and your job is to find the reason and put that player on a hand.”

Actionable information can be created by observing how others act.  Everyone does something for a reason, but smart people are more likely to do things for smart reasons.  Across many fields, it seems easier to find the solution to a problem knowing that it is solvable.

Sklanksy is specific about how to approach reading hands.  “Before you can technically analyze what your opponents might have, you must have played with them for a considerable length of time, seen how they play their hands against you, and most importantly watched them play hands in which you are not involved.  Even when you are not in a hand, you should not relax your concentration.”  You can apply this in public markets easily because of access to a long track record for many actors – businesses, investors, board members, private equity, activists.

Some players have an imperative to share their thinking on a specific investment or discuss their philosophy.  Others never articulate a process, but you can study their actions to discern a behavior pattern.  You can also learn from how they have played their hands in a set of circumstances, such as David Einhorn with option like levered equities in 2009 and then go find similar circumstances for you to play that hand similarly.  Knowing how someone plays a hand is much more informative than the returns they’ve generated when looking at their activity for ideas – as in poker there are bad players who win hands and good players who lose them.

The concept of reading hands can be expanded upon by seeking out books that offer up a credible account of a corner of Wall Street.  Only reading accounts by investors is an unnecessarily limited vantage from which to make investment decisions.  Fleecing the Lamb, about the Vancouver Stock Exchange, lays out how numerous promoters played their hands.  Dead Bank Walking is by a senior executive at a struggling bank. It shows how he played his hand and the bluffs he makes.  Metal Men shows the culture of commodity trading, written by a great journalist whose research included working at a commodity trader for a year. The lessons could be easily expanded to any kind of trading.  All of these represent perspectives you will encounter in investment situations.

A book I read after digesting Sklansky was The Caesar’s Palace Coup, which solidified for me the overall value of his framework.  It shows the hands of distressed debt investors, executive officers, private equity firms, investment banks, and corporate lawyers at a level that should aid in reading their hands better.  This ranges from the dynamics of an “independent” board to the mercenary realities of a “fairness opinion.”  It is credible because it is based on documents and testimony that came out during a bankruptcy trial, not a one sided victory lap account, surface layer PR spin, or an axe grinder.  The authors can compare sworn testimony and exhibits to what people say in research interviews for the book.  Events such as the Caesar’s bankruptcy had conflicts and that typically causes everyone to talk because they want their side of the story told.

“One of my favorite types of player is the one who never bluffs.  You have a tremendous advantage over these players because you just about always know where you’re at.”

Playing with managers who don’t bluff – who speak plainly, share a thought process grounded in reality not fantasy, or are incentivized on outcomes they can control (ROIC vs relative TSR) – has a deeper benefit beyond being a pleasant way to do business.  Poker, as with investing, involves calculating the reward, risk, and making a decision based on the ratio of the two.  Quantitative measurements can contain self deception because “numbers don’t lie *winky face*.”  With managers who don’t bluff, you will find that the numbers lie a lot less – compare a SPAC presentation or a share based compensation “adjustment” to financial results with a Texas Instruments capital allocation presentation.

Bluffing is not desirable in a business partnership.  This encapsulates a key disconnect when investing in capital intensive commodity businesses.  The current commodity boom aside, shale companies spent the better part of the last decade publishing investor presentations claiming alluring IRRs on their wells in order to attract both debt and equity capital, without which they would not have a business.  Yet they never produced free cash flow and many became financially distressed.  This dynamic makes it difficult to know what cards the company is actually holding, which short circuits attempts to accurately quantify the reward and obscures risks.  Capital intensive businesses are structurally in the position of bluffing because offering a compelling return is how capital is attracted, which is self defeating at best in a commodity business.  This is the truth behind the joke that a gold mine is a hole in the ground with a liar next to it.

As an exercise, of the following examples who do you think as a management team is closer to reality and revealing what is going on in their business and who do you think is stringing you out on increasingly unproven assumptions they want you to conflate with present reality?  Which one is easier to value with accuracy?  Who is more likely to be responding to reality in their daily decisions running the business?

Here is a slide from a company that has grown EPS at 16% annually for the last 10 years:

Here is an excerpt from a conference call of an unprofitable consumer unsecured finance company that claims to be a payments company:

I'll pull the curtain back a little and while I will keep it very high level in part to avoid giving out trade secrets, feel free to tune out for a few minutes while I nerd out.

 [Our company]'s underwriting advantage begins before any of our models are interrogated for a decision with our product design. Because [our company] is predominantly offered at the point of sale, we have a natural opportunity to explain our value and transparent approach to the consumer. As a result, we avoid much of the adverse selection that often comes with traditional lending.

 Coupled with SKU-level data we receive from our partners, our models tend to split the risk far better than those used in traditional consumer loans. Another fundamental structural advantage [our company] has is its total separability of transactions. Unlike providers of lines of credit, we underwrite transactions individually, modeling a consumer's ability to pay us back as well as their propensity to do so.

 This notion of separability is also recursive, a consequence of our product because repayment schedules are highly predictable, our models operate at an individual installment level. This separability is a powerful tool for modeling as well as managing risk. We're able to deliver a reliable forward-looking picture of both consumers and our own cash flow.

 Our proprietary network of directly integrated merchants as well as other sources of nontraditional underwriting data offers us a significant raw data advantage into feature engineering. We maintain a library of over 500 features that we select from as we create new models or update existing ones, while continuously looking for and eliminating any potential for disparate impact in our decisioning, both at individual variable and model levels.

 We train our models using academically well-understood gradient boosting technique with significant proprietary modifications we've invented that help us improve results. Because from the very beginning, we focused equally on consumer and merchant information, we ended up with a large number of models that are specific to our products and merchants who use them.

 Moreover, as we launch new products with new and existing partners, we acquire new types of data that we incorporate into the models and over time, give incremental weight, too. Underwriting models decay over time as macroeconomic conditions and consumer behaviors change. Even the very best performing ones can lose a few percentage points of their area under the curve every few months.

When to fold ‘em

I’ve expanded on how reading hands and bluffing in poker maps deeply onto the investment process.  It unfortunately took several years before I picked up Theory of Poker after it was recommended.  When other people have mentioned the parallels to poker it was only a sentence or two, which made it hard to internalize the value because I didn’t actually understand the deeper mechanics of poker they were referencing.  There is a lot more in the book that is a useful and actionable framework in investing - position sizing, position management, mistakes, adaptability – that makes it worthwhile to read the book on your own and draw the connections.