Mastery
Market studies
These are short histories of markets that broke someone's assumptions. The lesson in each is about liquidity, leverage, or crowding. None of them is a setup to repeat.
On this page
Read these for the mechanism, not for the drama. Dates and percentages below are the public record, rounded only where the sentence says about. A field manual is not a research paper. Where prints were chaotic, the page says so instead of inventing a tidy low.
Section 01
October 19, 1987
On Black Monday the Dow Jones Industrial Average fell 508 points, 22.6 percent, in one session. It is still the largest one-day percentage drop in that index. The S&P 500 fell about 20.5 percent the same day. Portfolio insurance was a strategy that tried to mimic a protective option by selling stock-index futures as prices fell. Selling into a fall, in size, in a market that was already falling, helped turn a decline into an air pocket. Program trading carried the selling from futures into stocks.
The useful lesson is not "computers are bad." Computers were doing what the humans had written: sell more as price drops. The lesson is that a model which assumes you can always transact near the last price will fail in the only moment you desperately need it. Circuit breakers were built because that day happened. A halt can slow a stampede. It does not refund a gap.
Section 02
The dot-com unwind, 2000 to 2002
The Nasdaq Composite peaked at 5,048.62 on 10 March 2000. By October 2002 it was near 1,114, a loss of roughly 78 percent. Many of the favorite names had no earnings. The story was the product. The S&P 500, which is broader than technology, also fell nearly 50 percent over that bear market. The decline was not a one-week crash. It was a long education in how slowly a narrative can finish dying.
Two lessons sit side by side, and both are true. Paying any price because a technology is real is not a valuation. And being early on the other side, with leverage and no defined loss, can ruin you while you are still eventually right. A correct idea is not a risk plan.
Section 03
2008 and the year that followed
From the October 2007 peak to the March 2009 low, the S&P 500 fell about 57 percent. Lehman Brothers filed for bankruptcy on 15 September 2008. Funding markets seized. In November 2008 the VIX, the options market's price of near-term S&P 500 volatility, closed above 80. Risky assets that had looked diversified started to fall together, because the thing investors were selling was risk itself, and because leverage forced selling that did not ask which story you preferred.
Diversification across five technology stocks was never diversification. In a crisis, diversification across anything funded the same way can fail too. The study belongs next to your position-size cap. If every position in the journal would have been crushed by one liquidity shock, you do not have five risks. You have one.
Section 04
The flash crash, May 6, 2010
U.S. equity indexes fell violently and recovered a large part of the drop within minutes. The Dow's intraday swing that day was about 1,000 points. The joint report from the SEC and the CFTC described a large sell program in E-mini S&P 500 futures that executed against volume, not against price or time, meeting a market where liquidity was being withdrawn. High-frequency firms that had been buyers became aggressive sellers as they passed inventory along. Later, a separate case ended with Navinder Sarao pleading guilty to spoofing in that futures market. Prosecutors said the spoofing contributed to the conditions. The 2010 report and the later case do not have to be flattened into one villain. The mechanism you can use is simpler.
Displayed size is not a promise. Quotes can cancel. A market order says "fill me at whatever is left." On an ordinary day that sentence is cheap. On this kind of day it is a blank check. Know which of your orders are blank checks.
Section 05
The Swiss franc, January 15, 2015
Since September 2011 the Swiss National Bank had enforced a minimum exchange rate of 1.20 francs per euro. On 15 January 2015 it discontinued that policy. The franc jumped. In the first minutes the euro's drop against the franc was on the order of 15 to 30 percent, and some quotes printed far beyond that because liquidity was gone. Stops did not fill at the stop. They filled wherever a bid finally existed, or they did not fill until the damage was larger. Some retail foreign-exchange firms failed. Alpari UK entered insolvency. Clients discovered that a stop on a leveraged currency bet is not insurance.
A peg is a policy. Policies change on a Thursday. If your plan's survival requires a central bank to keep a promise, you do not have a plan. You have a dependency. Size any trade that can gap as if the gap is allowed, because it is.
Section 06
March 2020
The S&P 500 peaked on 19 February 2020 and, by the low on 23 March, had fallen about 34 percent. That is one of the fastest bear markets on record. On 16 March 2020 the VIX closed at 82.69, the highest close in its history. U.S. equity circuit breakers triggered more than once. A very large monetary and fiscal response followed, and the index recovered far faster than it had after 2008. Sitting in the middle of March, that recovery was not a fact. It was one possible future.
"Buy the dip" is not a system until it has a size, a maximum number of buys, and a point where you are wrong or done. The people who were hurt worst were not only those who sold the low. They included people who bought every dip with size that assumed the first dip was the last. Speed was the teacher. Your process has to be writable on a bad morning, or it will be rewritten by the morning.
Section 07
GameStop, January 2021
GameStop was a heavily shorted stock. Reported short interest rose above 100 percent of the estimated public float. That can happen because the same shares can be lent more than once. In pre-split prices, intraday trades printed as high as $483 on 28 January 2021. The company later split, so a modern adjusted chart does not show 483. Read the scale before you repeat a number you saw in a screenshot.
Several brokers restricted new buys that day. The operational reason they gave was collateral demanded by the clearinghouse, which rises when volatility and volume rise. The episode is still argued about, including in hearings and lawsuits. This manual does not settle the argument about motives. The mechanics that are solid enough to study are these: a crowded short, a limited float, listed call options whose dealers buy shares as those calls move in the money, and trading rules that can change during the session. Together they can feed a move that has little to do with next year's earnings.
Section 08
Long-Term Capital Management, 1998
Long-Term Capital Management ran highly leveraged convergence trades: relationships that "ought" to tighten. In August 1998 Russia effectively defaulted on domestic debt and the fund's spreads widened instead. The balance sheet was on the order of a hundred billion dollars against a few billion of capital, with much larger derivative notionals off to the side. In September 1998 the Federal Reserve Bank of New York coordinated a private recapitalization by a group of banks. That was not a government check written to the fund. It was a forced lesson in who gets to hold a position when the margin clerk arrives.
You will probably never run a relative-value book. The transferable sentence is still yours. Leverage decides whether a temporary mark against you is a bad month or an ending. A relationship is not a law just because it held in the sample you trust.
Check yourself
What did portfolio insurance contribute on 19 October 1987?
Where the facts come from
- Federal Reserve History, Stock Market Crash of 1987, and the Brady Report (1988), for the Dow's 22.6 percent drop and the role of portfolio insurance.
- Published index history for the Nasdaq Composite peak of 5,048.62 on 10 March 2000 and the roughly 78 percent decline into October 2002.
- S&P 500 peak-to-trough decline of about 57 percent from October 2007 to March 2009. Lehman Brothers filed for bankruptcy on 15 September 2008.
- SEC and CFTC, Findings Regarding the Market Events of May 6, 2010. Later U.S. cases addressed spoofing in the E-mini that day.
- CBOE Volatility Index: a close of 82.69 on 16 March 2020, the highest close on record. The S&P 500 fell about 34 percent from 19 February 2020 to 23 March 2020.
- Swiss National Bank press release, 15 January 2015, discontinuing the minimum exchange rate of 1.20 francs per euro.
- FINRA short-interest reports and contemporary broker disclosures around 28 January 2021. GameStop later split; unadjusted prices are not today's scale.