▫️ The SEC Took Three Months to Work Out What Broke the Market.
At the low, nobody in the entire market would put up $25 million to buy it at a discount.
▫️ THE CORE TOPIC
Everybody checks the price. Some people check the volume. Almost nobody checks the only number that decides what happens when they try to leave.
Volume is a record of what already traded. Depth is how much the market can absorb next without the price moving against you, and the two can point in opposite directions at the same moment. The SEC's own reconstruction of one afternoon put it plainly: buy-side depth in the S&P 500 futures contract started the day at about 100,000 contracts and, over roughly four minutes, fell to about 1,000 contracts, less than one percent of the morning figure. Volume that afternoon was enormous. Capacity was gone.
Which means the honest measure of a holding is not its dollar value. It is how many days of ordinary trading that holding represents, because that is how long it takes to leave without becoming the reason it fell.
The gap matters most where depth is thinnest. A company in its first months as a listed business has a quoted price and little behind it.
▫️ THE MECHANISM
Four things separate what a market has traded from what it can take.
- Volume is history, depth is capacity. One tells you how busy the room was. The other tells you whether anyone will take the other side of your next order.
- The screen shows the shallow part. Depth close to the current price can look steady while the rest of the order book empties out underneath, which is precisely what regulators found had happened.
- Your own selling removes your buyer. Every order you fill consumes the bids you were counting on, so a large seller competes with himself the whole way down.
- Capacity thins when it is wanted most. Participants step back in fast markets because they cannot tell whether the prices they see are real.
This is why very large holdings are reduced in tranches over months rather than in one trade, and why the exit of a big position shows up as a long sequence of small ones.
▫️ THE CASE FILE
On the afternoon of 6 May 2010 a single firm began selling. At 2:32 an order went in for 75,000 E-Mini contracts worth around $4.1 billion, executed entirely by an algorithm.
The instruction was the problem. In the absence of price or time limits on the sell algorithm, regulators found the order executed far faster than anyone anticipated, and buying interest could not keep pace. All 75,000 went through between 2:32 and 2:51.
The result reached everything. American equity indices fell more than 5 percent in a matter of minutes and rebounded just as fast. At the low, Procter & Gamble traded 36 percent down on the day. Nobody in the market would put up $25 million to buy it at a 23 percent discount.
▫️ THE PRESSURE TEST
- Depth is adequate almost all of the time. The market absorbs ordinary orders without anyone noticing, which is why the problem stays invisible for years.
- Size is relative, not absolute. A million dollars is nothing in a large index future and a serious problem in a small mining company, and the dollar figure tells you neither.
- The rules changed afterwards. Circuit breakers now pause a stock that moves too far too fast, limiting the damage one order can do.
- Thin markets cut both ways. The same shallow book that punishes a hurried seller rewards a patient buyer, and the best entries often come from a forced exit.
The discipline that follows is straightforward. Size a holding against the days it would take to leave, not against the amount you would like to own, and who is bidding becomes a plan rather than an emergency.
▫️ AUTHOR'S LENS
In the pits, the first thing anybody taught you about a position was how long it would take to get out of it. Not what it was worth. How long. A good idea with no exit was still trouble.That habit never left me. I ask how many days a holding represents before I ask what it might return. And the slice I keep in metal is there for one unglamorous reason: on the worst afternoon of a decade, somebody somewhere is still bidding for it.Build the structure. Ignore the noise. |  |
▫️ PAST ISSUES
