A Man for All Markets, By Ed Thorp By Jesse Koltes, TheCharlieton.com
When I was eight years old, my dad caught me trying to cheat at blackjack.
“People get shot for doing that,” he said. If I wanted to win, he told me I needed to count cards. My dad became calm as he methodically explained a simplified counting method involving the five card.
Canyon Profits On Covid Crisis Refinancings
Canyon Partners' Canyon Balanced Funds returned -0.91% in October, net of fees and expenses, bringing the year-to-date return to -13.01%. However, according to a copy of the firm's investor correspondence, which ValueWalk has been able to review, the fund quickly bounced back in November, adding 7.3% for the month. Net of fees, the letter reported, Read More
Even as a boy, I sensed his fascination, and caught an early glimpse of my dad’s gambling streak. I wanted to connect with him about whatever this was.
Seeing that I wasn’t quite following, and still reeling from the casual death threat, my dad handed me a dog-eared copy of Ed Thorp’s book Beat the Dealer, and said that that was how he had learned.
Twenty years later, Ed Thorp’s autobiography, A Man for All Markets, was the only book Charlie Munger recommended at the 2017 meeting of the Daily Journal Corporation. That recommendation was my most important takeaway from the 4-hour affair. *
While Mr. Munger’s own comments from the meeting are worth reading, I’d suggest reading Thorp’s book first. In many ways, Thorp is a second Charlie in terms of both intellectual power and scope, which is perhaps the highest praise I could offer another person. Thorp’s book is chock- full of knotty lessons for investors, thinkers, and business people, but because Thorp is far less well covered than Munger, many of these ideas felt new and let me see them with fresh perspective.
My biggest takeaways are below.
Extreme Knowledge and Circles of Competence
“We often find that winning systems go almost ridiculously far in maximizing or minimizing one or a few variables” – Charlie Munger
“I’m no genius. I’m smart in spots—but I stay around those spots.”
— Tom Watson Sr., Founder of IBM
One of my favorite ideas from Charlie is that the secret to winning systems is the maximization or minimization of one or two key variables. Think of the minimalism at Costco, fee reductions at Vanguard, or portfolio concentration in the Kelley formula: In order to have better than average results, you need to have different than average behavior.
But one of my other favorite ideas from Munger is that investors should stay within their circle of competence, and only bet big when they have a high conviction understanding. As Buffett has said, he greatly prefers stepping over one foot hurdles to heaving himself over ten foot hurdles.
Both of these ideas are fantastic advice, but Thorp’s book made me realize that when you combine them there is a tension. In knowledge work fields (like investing), the variable you want to maximize is actionable understanding. You have to constantly improve your probabilistic understanding of how events will unfold in the future, while at the same time constraining your faith in your own abilities to understand the things you’ve learned about. You are forced to push and pull in opposite directions—to an extreme degree—at the same time.
Ed Thorp is an excellent and understudied example of how to thread this needle between growth and humility. When Thorp began to study the stock market, he was surprised and encouraged to discover “how little was known by so many.” He quickly recognized that the stock market was full of the same ignorance and lazy analysis that had pervaded the gambling world. Where others saw an unbeatable game of chance, Thorp saw a system of probabilities and payoffs masked by a veil of psychology and whim. If he could maximize his understanding of the system, perhaps he could beat the market, just as he had beaten Vegas.
His ambitious endeavor paid off. By applying the tools of physics, computer science and math to finance, Thorp created the world’s first “quant shop,” and a trading system that functioned profitably for decades with few drawdowns.
What I find incredible about Thorp’s example is not only that maximized his understanding and beat the market, but that he avoided the quackery and hubris that can so often bedevil people who have ventured so far from the average.
To make this point clear, consider how Thorp’s example compares with inventors of the Black-Scholes model. Thorp, and later Fischer Black, Myron Scholes and Robert Merton, stretched their understanding of the world to the extreme, and were able to deduce a theoretical model for pricing derivatives. Thorp took his discovering and traded it profitably for years. Merton and Scholes won the 1997 Noble price for the same discovery, but were unable to control themselves and use it safely within their circles of competence. Their hedge fund, Long Term Capital Management, blew up in 1998 and had to be bailed out by a government led consortium.
Thorp’s example is like an anti-Icarus, skilled enough to balance hubris and humility, flying neither too low nor too high.
Ed Thorp - Nontransitive Odds
Although Munger’s observation that winning systems often maximize one or two variables, Thorp’s book points out an an important exception: non-transitive games.
In Thorp’s telling, Warren Buffett introduced him to the subject at a fateful meeting of the two greats. Buffett would offer a guest their pick from a set of three dice. The guest was asked to pick the “best” die, and Warren would pick the “second” best. Both players would roll the dice at the same time, and the highest number would win. The trick was that no matter which die the guest selected, Warren could always beat them by picking second.
“This puzzles people because they expect things to follow what mathematicians call the transitive rule: if A is better than B, and B is better than C, then A is better than C. For example, if you replace the phrase “better than” by any of the phrases longer than, heavier than, older than, more than, or larger than, then the rule is true. However, some relationships don’t follow this rule. For instance, is an acquaintance of, and is visible to, do not. The childhood game of Rock, Paper, Scissors, is a simple example of a nontransitive rule.”
A Man for All Markets, page 158.
What’s remarkable about the non-transitive rule is that it bends the mind away from thinking that issues are always about straight-forward maximization or minimization. Sometimes, the edge doesn’t always go to the actor who boldly picks the apparently best option, but to the person who adapts to the first mover, and counters with something better.
Institutions and Resilience
For all the similarities between Munger, Buffett, and Ed Thorp, it is notable that Berkshire stands as a colossal monument to the first pair’s intellectual achievements, while no such entity exists in Thorp’s wake. All three men possessed an enthusiasm for the power of compound interest, and enjoyed great returns and long careers. What explains the difference in institutional outcomes?
Part of the answer has to do with how Thorp and Buffett constructed their backup systems. In Berkshire’s conglomerate model, excellent businesses are acquired, but remain totally unintegrated. This requires that Berkshire forego the “synergies” and cost reductions that typical come with merger integrations.
But decentralization gives Berkshire a valuable firewall against negative contagions. When bad practices pop up at Wells Fargo, they can’t quickly spread to Kraft-Heinz or even to Omaha. Because no single pillar is central to Berkshire’s success, perhaps no longer even Buffett himself, the company has grown into one of the largest and most stable businesses in the world.
Compare that structure to Thorp’s now defunct Princeton-Newport Partners. Thorp arranged PNP such that the business was split into two geographically separate but equal offices. The East Coast office handled all the business administration and trading execution work that Thorp found tedious. Thorp was delighted to outsource the administrative parts of the business that he despised, and set up his own office on the West Coast where he designed the firm’s mathematical trading strategies.
Critically, Thorp claims that both offices acted independently, with each exerting full control over their own hiring decisions and cultures. Thorp may have thought that this added a protective layer of redundancy to his business as he often thought deeply about worst case scenario planning in nearly all aspects of his business. Thorp even asked Goldman Sachs what would happen to PNP’s account if a nuclear bomb went off in New York harbor (Goldman allayed his concerns by having a backup system in Colorado).
Unfortunately for Ed Thorp, his backup system overlooked how co-dependent the reputation of two separate offices, with two separate bosses and oversight systems could be. That was a fateful mistake.
When the top five officers at PNP’s East Coast office were charged with stock manipulation and other white collar crimes in the late 1980s, the system Thorp created did have the resilience to guarantee the firm’s survival. The decentralization of the two offices had let one veer dangerously off track, but the interrelatedness of the two parts meant that neither could survive independently.
Though Thorp’s West Coast office was innocent in practice, it was fatally wounded by its relationship with the East Coast counterpart. Seeing no way forward, Thorp was forced to wind down his partnership before it could achieve the scale we might have expected of it.
By Thorp’s telling, the opportunity cost of this closure was massive. His ideas went on to inspire the success of the next generation of great quantitative firms such as D.E. Shaw and Citadel, each now valued at many billions of dollars.
*My second best takeaway was taking the opportunity to ask Mr. Munger a brief question in person. Unfortunately, things did not go exactly as planned. I racked my brain for months searching for a line of attack that might shake loose a new, previously undisclosed nugget of wisdom. Unfortunately, Mr. Munger deflected my question before launching into a stump speech that while interesting, was well known.