# Sixteen Years on the Bench: Tobias Carlisle on Buffett's Three Best Deals and When Value Comes Back Source: Realpha Blog (blog.getrealpha.com) Original article and charts: https://blog.getrealpha.com/en/blog/behindbalance-2026-10-15-64-the-value-disciple-tobias-carlisle-on-buffett-s/ > Listening notes on Behind the Balance Sheet episode 64. Deep value investor Tobias Carlisle rereads Buffett through Sun Tzu: avoid ruin first, then talk about winning. Covers value traps versus quality traps, how many positions you need, and the real design behind GenRe, BNSF and the Japanese trading houses. Educational listening notes, no investment advice, no price targets or buy and sell calls. Published: 2026-10-16 Locale: en Tags: value investing, Buffett, Sun Tzu, diversification, quality factor TL;DR: Carlisle's working rule is to rule out the kind of mistake that takes you out of the game before reaching for return; he has also shifted from buying only cheapness to paying for quality. ![A deep aisle inside an emptied warehouse where a single shaft of high window light lands on one abandoned crate, a figure with a torch walking toward it](/covers/behindbalance-2026-10-15-64-the-value-disciple-tobias-carlisle-on-buffett-s-cover.png) > Numbers alone confer no advantage. Refrain from advancing rashly; gather your strength, read your opponent, win the people, and that is enough. The one who takes no thought and makes light of his opponent will be taken captive. > > — Sun Tzu, *The Art of War*, "On the March" (Spring and Autumn period; my own rendering) On 15 October 2026, *Behind the Balance Sheet* episode 64 had Steve Clapham interviewing Tobias Carlisle — Toby — of Acquirers Funds, a deep value investor with 26 years in the chair. He describes one good decade followed by 16 difficult years, and says today's market reminds him of the period around 2000. His new book, *Soldier of Fortune*, reads Buffett alongside Sun Tzu and pulls out one shared thread: make sure you cannot be destroyed, then worry about winning. In this episode he reworks three Buffett deals — GenRe, BNSF and the five Japanese trading houses — and admits he has changed his own mind about quality. Everything he says rests on one premise: nobody can tell in advance which company is dying. ## Sixteen years in the cold. What makes him say it turns? His evidence is a count of historical cycles rather than a trigger. Toby says this large-growth-beats-small-value cycle has run six times before, that the last one ended with the millennium, and that it was followed by roughly 15 good years for value and quality through to about 2015. The inference: a pendulum pushed far to one side tends to swing back. ![A long wave rises and falls: the left stretch repeats six small swings for past cycles, the middle swells into a tall peak for the fifteen years value led, then slides down into a deep trough for sixteen years on the bench, and a dashed line at the end tilts back up.](/figures/value-cycle-pendulum-en.svg) What held my attention was how little he claimed. He says he cannot place us in the cycle, only that there is a reasonable chance we are near the end of one, and he quotes the old line about never asking the barber whether you need a haircut. He adds one side observation: the episode was recorded shortly after the SpaceX listing, and waves of large IPOs have historically clustered near market highs — while noting that nobody in the moment ever thinks they are at the top. ## If you buy hated companies, how do you separate temporary from terminal? His answer stopped me: you cannot tell in advance. They look identical, and the difference shows up only in hindsight. Once you concede that, the whole design has to be built around the hole. His answer is diversification on two axes — across industries, and across strategies. He uses his Booking.com position as the illustration: take the historical financials, assume the business keeps working the way it has, and today's price hands you a large discount. If the market's narrative is wrong — people stop flying, everyone searches flights through an AI — price returns to value. If the narrative is right, that kind of structural decline moves slowly, and he already holds the discount as a cushion. He calls it a handicapping race rather than picking the winner. ![On the left, value sits above price with the gap between them marked as the discount; on the right, two paths branch out from the price point, one climbing back up to value and one drifting slowly down, both staying above the discount line.](/figures/discount-cushion-fork-en.svg) He also owned Meta, bought when the stock was beaten up the first time Zuckerberg walked into the metaverse, sold after it recovered. Deep value in his hands does not mean only bad businesses. ## So what is a quality trap? Can good companies hurt you too? Everyone has discussed value traps for decades. Toby points at an equally common mistake nobody named: paying for growth that never shows up, paying for a certainty that was never there. He calls it the quality trap. This is the part of the episode I rate highest. The point is that the number of ways to be wrong does not shrink depending on which side you stand on. Stand with cheapness and you buy companies that are genuinely dying. Stand with quality and you pay today for an imagined twenty years. Both hurt; the difference is which one puts you out of the game. So he routes the problem back to two operable things — valuation discipline, and position sizing. His phrasing: when you are wrong, make sure it is not catastrophic for the portfolio. ![One plank has a pit dropping away at each end, the left pit labeled value trap and the right pit labeled quality trap, while two pillars labeled valuation discipline and position size hold up the middle.](/figures/two-traps-one-ruler-en.svg) He then describes what you actually need to understand about a business: how it behaves through a cycle. Some businesses are Taleb's turkey — years of good numbers and then one catastrophic loss, and that loss is exactly what the share price is discounting. Semiconductors are his example. They have been deep value stocks several times over, their cyclicality makes them hard to value, and he says the market has forgotten they are cyclical. ## How many positions do you need? Doesn't Kelly tell you to bet big? Toby wrote *Concentrated Investing* on this question. The academic answer sits around 20 to 30 names, but he notes the academics are answering a different question — replicating the market with as few stocks as possible, assuming a random walk. Among value investors, Graham said 30. On the Kelly-style sizing that Buffett and Munger favour, he raises an objection I had not considered: Kelly is designed for a serial game, one hand after another, like sitting at a blackjack table. Investing is a parallel game — an entire universe of candidates is in front of you at once, all with positive expected return, so each deserves some size in proportion to it. Account for that and a 40% single position becomes hard to justify, which is where Kelly-following value investors often start. ![On the left, cards lined up in a row arriving one hand at a time, so a single bet can be large; on the right, a whole field of dots laid out at once, each needing a share, which makes any single bet small.](/figures/kelly-sequence-vs-universe-en.svg) His conclusion is that the old value intuition was about right: roughly 30 names. But it depends on what you hold. His small and micro cap ETF carries 100 names, because those businesses are smaller and more volatile; even at 100 names the portfolio swings two or three times as much as his 30-name mid and large cap strategy, with individual names moving 15% to 20% on a given day. Asked whether that is a liquidity effect, he says likely — crossing the bid-ask spread. That "I don't know why, the numbers just look like this" honesty does more for my confidence than a tidy explanation would. ![Two wave lines stacked: the upper one, marked thirty holdings, has a small swing, while the lower one, marked a hundred holdings, swings two to three times wider.](/figures/thirty-vs-hundred-volatility-en.svg) ## Why those three Buffett deals? Because all three carry the wrong label. **GenRe (1998).** It looked expensive, Buffett used Berkshire stock, and he talked about synergies — three things he had criticised in his own letters shortly before. Toby's reread: Berkshire had put a third of its capital into Coca-Cola in the late 1980s and early 1990s, Coke rose roughly 14 times by 2000 and traded near 60 times earnings, and it was a large slab of Berkshire's net worth. People ask why Buffett never sold Coke. Toby says he did — by selling part of Berkshire. GenRe's portfolio was all bonds; issuing Berkshire stock to GenRe holders diluted the Coke position substantially and brought a mountain of bonds in. Then 2000 arrived, and that crash was broader than dot-coms: Walmart was no tech stock and still got expensive and still fell; so did Coke; Berkshire itself had traded at three times book. On the other side the bonds rallied, Buffett sold them down into equities, and the next 25 years were set up from there. Toby calls it a defensive acquisition. The twist is that Buffett spent years afterwards criticising the deal for importing financial weapons of mass destruction — derivative contracts where both sides could book a profit. Toby says he wanted to correct the record. ![Two stacked bars side by side: the first has a large Coca-Cola block, while the second shows that block squeezed smaller with a big bond block piled on top, and below the bars stocks fall and bonds rise in the crash.](/figures/genre-dilution-en.svg) **BNSF (2009).** Read at the time as a violation of Buffett's own high-return-on-equity doctrine, because railroads are capital hungry. Toby admits he had not fully priced in how much the railroads had improved. Someone later showed that Buffett pulled roughly half his capital back out in short order, cutting the risk, and what he bought was a vast regulated empire earning about 10% on capital expenditure in an era of near-zero rates. Clapham offers the reading I liked most: Buffett was looking at five, ten, fifteen years out, knew the cash would be hard to house, and bought a business that could absorb enormous capex. Toby agrees, and notes Buffett has said as much. **The Japanese trading houses.** Toby's decomposition is carry plus diversification. Berkshire could issue yen bonds at close to zero while those trading houses paid dividends of seven, eight, nine percent; each is a diversified conglomerate in its own right, and he bought five. The spread between dividends received and financing cost alone brought in about $600 million a year. Toby's verdict: a Buffett deal, where the capital walks back to you. ![A long bar of invested capital, half of it flowing straight back along a return arrow while the other half feeds into a downward-widening funnel labelled regulated and about ten percent.](/figures/bnsf-capital-return-en.svg) Clapham also asks about the recent Alphabet participation — roughly $10 billion with about a 6% discount. Toby stays guarded: Google is well positioned in AI and can integrate it across its product suite, so the deal ticks the other criteria; but this is a public-market private placement in a hot name near the top of the market, and on price he says he does not know. You could have had it cheaper in recent years, and you do not get to go back in time — you only get to decide now. ## What can I do with this myself? What I took away was not his bullishness on small value but the passage where he changes his mind. Toby says that recording a podcast every week and talking to a lot of investors has moved him on quality: he used to sit at the junk end of the quality scale and argue for it, and now he says he wants to be agnostic about where on that scale he stands — he will buy better companies if he can get his price, and some companies are worth paying up for. The evidence he cites is "Quality Minus Junk" (AQR, 2016), where quality means high return on assets, high return on equity, low variability of those returns and reasonable growth, and where the finding is that the market underpays for it. The same paper discusses the price of quality: when quality gets expensive it underperforms, when quality gets cheap it outperforms. He checked the performance since publication and says we now sit at a point where quality is cheap — and the path to cheap quality was a decade of quality doing badly. So his method today fits in a line: value the quality on offer, then demand a discount to buy it. He says that is traditional value investing, and he just wants the researched version with explicit rules bolted into his process. For someone like me, who also oscillates between cheap and good, the use of this passage is the repair it demonstrates. He did not swap out a belief; he wrote the conditions of the belief in finer detail. The old rule was "buy it if it is cheap enough." The new rule is "price the quality into the value, then still demand a discount." The first rule traps you at the junk end through a decade when quality is expensive. The second does not. My own losses over the past few years have mostly come from treating "this worked for a while" as "this works" — his repair gives me a sentence pattern to copy. ![The top row has a single gate leading to buy, while the bottom row has two gates in series, first pricing the quality and then demanding a discount, before leading to buy.](/figures/old-rule-new-rule-gates-en.svg) One more thing I want to keep. Asked how he finds time for a fifth or sixth book, Toby says writing is how he thinks, that he does not know what he thinks until he writes it down, and then adds that you would be surprised how many ideas you can kill simply by writing them down. A year into writing these listening notes, that matches what I find — plenty of thoughts that felt clever collapse by the third sentence of typing. ## Further Reading - *Behind the Balance Sheet* episode 64 (15 October 2026), host Steve Clapham, guest Tobias Carlisle - Carlisle's books: *Soldier of Fortune* (the new one, Buffett and Sun Tzu), *The Acquirer's Multiple*, *Concentrated Investing*, *Quantitative Value* - Clifford Asness and co-authors, "Quality Minus Junk" (AQR, 2016): the definition of the quality factor and the price of quality - The aphorisms of Baltasar Gracián, source of Buffett's sins of omission; Toby calls him Machiavelli with a soul - Roger Lowenstein, *Buffett: The Making of an American Capitalist*, and Alice Schroeder, *The Snowball*: Toby's picks for watching Buffett evolve - *The Art of War*: the book Toby says has shaped him most as an investor ## One Thing to Take Away One idea: ask whether the worst outcome takes you out of the game before you ask what the upside is. Everything Toby covers — diversification, sizing, demanding a discount, understanding a business across a cycle — is one idea in different clothes. He concedes he cannot see the future, so his design has a single goal: when he is wrong, he is not carried off the field. Stay on the field and time does a lot of the work for you. A practice I have tried, and it works outside investing too: take a sheet of paper, write down three decisions you face this week, and after each one ask whether you could carry on as normal tomorrow if it went completely wrong. The stake is not only money — time, credibility, health and relationships all count. For any item where the answer is no, halve the stake: half the money, half the hours, half the scope of what you promised anyone. The first time I did this I halved a project I loved, and the half that survived is still alive today. ![Three decisions each get a bar; the first two pass the question and keep full length, while the third answers no and is cut through by a trim line down to half, with money, time and commitment shown halved alongside.](/figures/cut-the-stake-in-half-en.svg)