George Soros: the trade that broke the Bank of England
On 16 September 1992 the British government raised interest rates twice in one day, spent reserves it could not spare, and still could not hold the pound above a line it had promised to defend. By evening the line was gone. The fund that had bet hardest against it reportedly made about a billion dollars. The reason the story is worth your time is not the billion. It is that the loss on that trade was capped at roughly six per cent before it was ever put on — and somebody had published exactly where the cap was.
KEY TAKEAWAYS
- The risk was capped in public. Sterling’s ERM band ran from DM 2.7780 to DM 3.1320 around a central rate of DM 2.9500. Everyone could read where the trade would be wrong.
- The reward-to-risk was about 1 : 2.4 — roughly 6.19% of adverse room against a fall of about 15% in sterling’s effective rate. That ratio is ordinary. The size was not.
- It was not Soros’s idea. Stanley Druckenmiller built the thesis and the position. Soros’s contribution was the sizing decision — “go for the jugular”.
- The most-quoted “Soros” line is Druckenmiller’s. “It’s not whether you’re right or wrong…” is Druckenmiller in The New Market Wizards, describing what he learned from Soros.
- The same instinct cost him too. Roughly $2bn in Russia in 1998, and a late exit from the technology bubble in 2000 that ended the Druckenmiller partnership.
Who is George Soros?
György Schwartz was born in Budapest in 1930 and survived the German occupation of Hungary as a Jewish teenager under a false identity. He reached London in 1947, worked as a railway porter and a waiter while studying at the London School of Economics, and graduated in 1952. He moved into merchant banking, then to New York, and in 1969 started the fund that eventually became the Quantum Fund.
His stated method is unusual for a trader in that he wrote it down and published it. The Alchemy of Finance (1987) sets out what he calls reflexivity: market prices do not merely reflect the underlying reality, they change the behaviour that produces it, so the two feed each other. You do not need the philosophy to use the trading consequence, which he states plainly.
My approach works not by making valid predictions but by allowing me to correct false ones.George Soros, The Alchemy of Finance (1987)
The same instinct shows up in the odder habit he reported in Soros on Soros (1995): he treated the onset of acute back pain as a signal that something was wrong in the portfolio — not a diagnosis, but a prompt to go looking.
That is a different job description from the one most new traders write for themselves. It does not require knowing what happens next. It requires knowing, in advance, what would prove you wrong — which is the entire content of a support or resistance level, and the reason the 1992 trade is worth studying rather than admiring.
What actually happened on Black Wednesday?
Sterling joined the Exchange Rate Mechanism on 8 October 1990 at a central rate of £1 = DM 2.9500, with a wide 6% band. That put a floor under the pound at DM 2.7780 and a ceiling at DM 3.1320. The British government was publicly committed to keeping sterling inside those two numbers — by spending reserves, by raising interest rates, or both.
By September 1992 that commitment had become expensive. German rates were high, British unemployment was rising, and holding the pound up required the exact monetary policy a weak British economy could least afford. The contradiction was visible to anyone who cared to look: the government had promised a price it no longer had a domestic reason to pay.
Stanley Druckenmiller, who ran money at Quantum, had built a short position of around $1.5bn against the pound. Sebastian Mallaby’s account in More Money Than God (2010) describes Druckenmiller proposing to add to it gradually, and Soros’s reply: go for the jugular. The position was reported to have grown to roughly $10bn — larger than the fund’s own capital.
On 16 September the Bank of England raised its minimum dealing rate from 10% to 12%, and announced a further rise to 15% for the following day. Sterling did not recover. At 7.40pm the Chancellor, Norman Lamont, announced that Britain was leaving the ERM. The 15% rate was never implemented.
Two numbers close the story honestly. Sterling’s effective exchange rate fell about 15% by early 1993. And the cost to Britain, which circulated for years as £13bn to £27bn, was put at a net opportunity cost of £3.3bn when Treasury papers were released under freedom of information in February 2005, plus about £800m of reserves spent defending the pound between April and September. We print the later figure because it is the documented one.
Why was the risk on that trade smaller than it looks?
Shorting a national currency against a central bank sounds like the most reckless thing on this website. Measured properly, it was one of the more conservative bets available that month — because the government had published the worst case.
Sterling was sitting on its floor of DM 2.7780. If the defence succeeded, the pound could climb back toward the central rate of DM 2.9500. That is 6.19% against a short. The absolute worst the band allowed was the ceiling at DM 3.1320, or 12.74% — and a currency being defended at its floor does not sprint to its ceiling in a few weeks. If the defence failed, there was no band at all, and no arithmetic limit on how far sterling could fall.
This is what an asymmetric bet actually means, and it has nothing to do with confidence. Soros and Druckenmiller did not need to be sure. They needed a situation where being wrong was small and bounded, and being right was not bounded at all. The band did half the work of a stop-loss for them, because someone else was obliged to enforce it.
What did the trade really pay, per unit of risk?
Here is the part that gets left out of every retelling. Take the reported figures and divide.
Risk to the ERM central rate: 6.19% — about $600m on a position reported at roughly $10bn. The move that actually arrived: about 15%. That is a reward-to-risk ratio of roughly 1 : 2.4. It is a perfectly ordinary ratio. Plenty of losing traders take 1 : 2.4 trades every week.
The reported profit reconciles too. About $1bn on a position reported at about $10bn is a move of roughly 10% — consistent with what sterling did, and a useful check that the legend and the arithmetic have not drifted apart.
Read the last two rows of that table together. The ratio in both columns is the same. The results differ by a factor of about twenty million. Everything separating them is capital and access to it — not insight, not courage, and not a secret. The ratio is the transferable part. The size is the part that belongs to a balance sheet you do not have.
PRACTICE CORNER
Soros had the luxury of a level someone else was defending, published in advance. You do not — so the number that caps your loss has to be one you write down before the position exists. Take any chart you already follow, find the level below which your reason for being long stops being true, and work out what that distance costs you at the size you were about to trade. Every major exchange shows the stop and the position quantity on the same order screen, so the arithmetic can be checked before the trade rather than after.
Referral links — they never change our assessment. Education only; most retail traders lose money.
Our position size calculator turns that distance into a quantity, and position sizing covers the mechanics.
Did Soros say “it’s not whether you’re right or wrong”?
No. The most widely circulated “Soros quote” on the internet is not a Soros quote.
The line is: “It’s not whether you’re right or wrong that’s important, but how much money you make when you’re right and how much you lose when you’re wrong.” It appears in Jack D. Schwager’s The New Market Wizards, in the Stanley Druckenmiller interview, where Druckenmiller is describing the most important thing he learned from Soros. It is a report of Soros’s thinking by the man who worked beside him — which is close, and is not the same thing.
It’s not whether you’re right or wrong that’s important, but how much money you make when you’re right and how much you lose when you’re wrong.Stanley Druckenmiller — interviewed in Jack D. Schwager, The New Market Wizards (1992), p. 207, on what he learned from Soros
Attribute it correctly and it gets more useful, not less. It stops being a maxim from a man with $10bn to deploy and becomes a working note from an employee about his boss’s actual decision rule — a rule you can see operating in the ERM band, where the entire trade was designed around how much would be lost if it failed.
The habit matters more than this one instance. If a quote can be misattributed for thirty years without anyone checking, so can a return figure, a drawdown, or a claim that someone never had a losing year. Checking the source is the transferable skill.
What did it cost him?
A profile that prints only the winning trade teaches nothing, so here is the rest of the record.
October 1987. Soros was positioned for a crash in Japan rather than in the United States, and the crash came in the United States. The Quantum Fund took heavy losses in the days around 19 October, and finished 1987 up about 14% — a respectable year that was very far from where it had stood before the crash.
Russia, 1998. Quantum’s Russian holdings were reported to have lost around $2bn when the rouble was devalued and Russia defaulted. The positions were substantial and illiquid, and the exit was not available at the price the model assumed.
The technology bubble, 2000. Soros later acknowledged he had been late getting out. The losses ended the Soros–Druckenmiller partnership and forced a restructuring of the fund. That episode is told from the other side in our profile of Stanley Druckenmiller, who bought the top he had himself called.
The uncomfortable point is that these are not failures of a different faculty. The instinct that says size up now is the same instinct in 1992 and in 2000. It produced a billion dollars once and a partnership-ending loss later, and no amount of study of the 1992 trade tells you which one you are living through.
Where this stops working
Nobody is defending a level in your market. This is the big one. Soros’s downside was capped because a sovereign government had publicly committed to buying at a fixed price. No such commitment exists anywhere in crypto. A level that “always holds” on your chart is held by traders who can change their minds in a second and owe you nothing — which is exactly what support and resistance is really describing.
Your cap is a stop, and a stop is an instruction, not a promise. The ERM band was enforced by a central bank with reserves. Your 6% is enforced by an order book that may not be there when you need it. Read slippage and liquidity and spread before treating a stop distance as a guaranteed loss size.
He could survive being early; leverage usually cannot. A $10bn position needs margin to sit through weeks of nothing happening. On a retail account with leverage, being right eventually is indistinguishable from being wrong, because liquidation arrives before the thesis does.
One trade is not a method. Quantum ran for decades and the pound trade is one entry in the ledger. Any lesson drawn from a single outcome — including this one — is drawn from a sample of one.
Common mistakes when learning from Soros
Copying the conviction and skipping the cap. “Go for the jugular” is the famous half. It was said about a position whose downside was already bounded and known. Treating asymmetry as a feeling. It is two numbers you can write down before entering, or it is not asymmetry. Reading reflexivity as permission to predict. Soros’s own summary is the opposite: the method corrects false views rather than producing true ones. Giving Soros credit for Druckenmiller’s work. The thesis, the research and the original position were Druckenmiller’s; the sizing was Soros’s.
FAQ
How much did George Soros make on Black Wednesday?
Reports put the Quantum Fund’s profit at roughly $1bn, with some accounts saying $1.1bn. There is no official published figure from the fund, so the number should be read as widely reported rather than confirmed.
What did Black Wednesday cost the UK?
Treasury papers released under freedom of information in February 2005 gave a net opportunity cost of £3.3bn, plus about £800m of reserves spent defending sterling between April and September 1992. Earlier estimates of £13bn to £27bn predate those documents.
Did Soros say “it’s not whether you’re right or wrong”?
The line is Stanley Druckenmiller’s, in The New Market Wizards (1992), p. 207, describing the most important lesson he learned from Soros. It is routinely quoted as if Soros said it himself.
Can a retail trader copy the 1992 trade?
The reward-to-risk shape, roughly 1 : 2.4, is ordinary and copyable. The cap on the downside is not: it came from a government’s public commitment to a price. In markets with no such commitment you have to create the cap yourself with a stop, and stops can slip.