Quant Analytics: from scratch to pro · Microstructure, execution and costs · Pro
Market fragility: the 2010 flash crash and the 2015 franc shock
Liquidity is thinnest when it is needed most. Two episodes, the US equity flash crash of 6 May 2010 and the end of the Swiss franc floor on 15 January 2015, show how a rule that ignores price, or a price that stops being defended, turns a thin market into a gap.
In this lesson
- Explain how a volume-linked selling rule that ignores price can feed on falling prices
- Explain why a price floor is a promise that can end at once, and what that does to stop orders
- Compute the gap loss of a stop order and of a leveraged position
- List controls that limit damage from thin liquidity
Read first: Slippage and market impact: the square-root rule, Gaps and weekend risk
Markets feel liquid until many participants want the same thing at once. Then the spread widens, depth disappears and prices jump from one level to a distant one. Two well-documented episodes show two different routes to the same place: an automated selling rule that kept selling into a thin market, and a policy price that was defended until it was not.
6 May 2010: a rule that ignored price
The joint report of the US futures and securities regulators (30 September 2010) describes how a large sell program in the E-Mini S&P 500 futures market met a day of falling and volatile prices. At 2:32 p.m. a large fundamental trader, a mutual fund complex, started a program to sell 75,000 contracts, valued at about $4.1 billion, to hedge an existing position.
The program ran through an automated algorithm. Its rule was to sell at 9% of the trading volume calculated over the previous minute, without regard to price or time. Between 2:32 p.m. and 2:45 p.m. it sold about 35,000 contracts, valued at about $1.9 billion. That is 46.7% of the program in 13 minutes, at an average of 54,667 per contract for the whole order.
The mechanism follows from the rule. When prices fall fast, volume rises. A rule tied to recent volume then sells faster, into a market whose depth is falling. It contains no condition that would slow it down as the price drops. The report describes a market already fragile and volatile that day. Later academic work using transaction-level data, such as Kirilenko and co-authors (2017), studied the behaviour of high-frequency traders in the same minutes. Read the primary report for the full sequence; this lesson uses only the part that teaches the mechanism.
15 January 2015: a price that stopped being defended
A policy floor holds a price in place by a promise. The Swiss National Bank had set a minimum exchange rate of CHF 1.20 per euro. On 15 January 2015 it discontinued the minimum with immediate effect and ceased the foreign currency purchases that enforced it. In the same announcement it lowered the sight deposit rate to -0.75%.
When a pegged price is released, the market has to find a new level. Orders resting in the book, including stops, are handled at whatever prices the thin market offers next.
Worked example: what a gap does to a stop
The numbers here are hypothetical. A position is entered at 100 with a stop order at 98, a planned loss of 2%. A stop becomes a market order when the trigger is touched. If the market gaps and the next available price is 90, the order fills there, and the loss is 10%. That is 5 times the plan.
Leverage turns this into a margin problem. At 20 times leverage a gap of 5% against the position costs 100% of the margin posted. A gap of 8% costs 160%, which is more than the margin, so the account owes money. A stop does not prevent this, since it cannot fill where there are no trades. See gaps and weekend risk.
What the two episodes share
- Liquidity is not a constant. The depth quoted in normal times says little about depth under stress.
- Rules that ignore price are fragile. Volume-linked or time-linked selling has no brake.
- Promised prices can end at once. A floor, a peg or a cap hides risk that appears all together.
- Stops are triggers, not guarantees. The fill depends on the book at that moment.
- Leverage turns a gap into ruin. Size must come from the gap scenario.
Common mistakes
- Sizing positions from the normal daily range and ignoring gap risk.
- Assuming a stop order caps the loss at the stop level.
- Reading a quiet history of a pegged price as low risk.
- Using leverage at which a plausible gap exceeds the margin.
- Treating a fast recovery in one episode as proof that thin markets are harmless.
In practice
A gap scenario asks what is lost if the next price is far from the last one, and whether the account can bear it. Limit orders do not turn into market orders in a thin market, as stops do, and low leverage keeps a plausible gap below the margin posted. Research on FinUties is where studies built on methods like these are published, with their method and limits stated. Related ideas: tail risk and rare events.
Worked numbers
Each result is recomputed from the formula shown.
| What | Formula | Result |
|---|---|---|
| share of the program sold in the first phase | 35000 / 75000 | 46.7% |
| average value per contract | 4.1e9 / 75000 | 54,667 |
| planned loss of a stop at 98 from 100 | (100 - 98) / 100 | 2% |
| loss when the stop fills at 90 | (100 - 90) / 100 | 10% |
| actual loss as a multiple of the planned loss | gap_loss / planned_loss | 5 times |
| illustrative gap against a position | 0.05 | 5% |
| loss on margin at 20 times leverage, small gap | gap_small * 20 | 100% |
| loss on margin at 20 times leverage, 8% gap | 0.08 * 20 | 160% |
| larger illustrative gap | 0.08 | 8% |
Check yourself
A stop order at 98 on a stock trading at 100 executes at 90 after a gap. What loss did the trader plan, and what was the loss?
Planned: 2%, from 100 to 98. Actual: 10%, from 100 to 90. The stop fixes the trigger, not the fill price.
A position uses 20 times leverage. A gap of 5% against it costs what share of the margin posted? And a gap of 8%?
5% x 20 = 100% of the margin. 8% x 20 = 160%, which is more than the margin, so the loss exceeds the money posted.
Why does a sell rule tied to 9% of recent volume behave differently from one tied to price?
Volume rises when prices fall quickly, so the rule sells faster just as the market gets thinner. A price condition could have paused it; this rule has none.
A central bank defends a floor with unlimited purchases. What risk does a trader short the floor's currency face?
The floor is a policy choice that can be dropped at once. The risk is not a gradual drift but a single large gap, and the position may be closed far from any stop level.
Name two controls that limit damage from a thin market.
Limit orders instead of stops that become market orders; size set from a gap scenario, not from the normal daily range; leverage low enough that a gap does not exceed margin.
Key terms
Checked facts
Every figure below carries how it was checked: Primary checked (read in the cited primary, official, academic or data source), Two sources agree, or Approximate (an order of magnitude only).
- Primary checked
At 2:32 p.m. on 6 May 2010 a large fundamental trader, a mutual fund complex, started a program to sell 75,000 E-Mini S&P 500 futures contracts, valued at about $4.1 billion, as a hedge.
- Primary checked
Between 2:32 p.m. and 2:45 p.m. the sell algorithm sold about 35,000 contracts, valued at about $1.9 billion, of the 75,000 intended.
- Two sources agree
The algorithm targeted an execution rate of 9% of the trading volume calculated over the previous minute, without regard to price or time.
- Primary checked
On 15 January 2015 the Swiss National Bank discontinued the minimum exchange rate of CHF 1.20 per euro with immediate effect.
- Primary checked
The same announcement lowered the interest rate on sight deposit balances above an exemption threshold to -0.75%.
Sources
- Findings Regarding the Market Events of May 6, 2010. Staffs of the CFTC and SEC, report to the joint advisory committee on emerging regulatory issues, 30 September 2010, 2010. Accessed 9 Nov 2026
- Swiss National Bank discontinues minimum exchange rate and lowers interest rate to -0.75%. Swiss National Bank, press release of 15 January 2015, 2015. Accessed 9 Nov 2026
- The Flash Crash: High-Frequency Trading in an Electronic Market. The Journal of Finance 72(3), 967-998 (Andrei Kirilenko, Albert S. Kyle, Mehrdad Samadi, Tugkan Tuzun), 2017. Accessed 9 Nov 2026
- Trading and Exchanges: Market Microstructure for Practitioners. Oxford University Press (Larry Harris), 2003. Accessed 9 Nov 2026
- Regulators blame computer algorithm for stock market 'flash crash'. Computerworld (Lucas Mearian), 2010. Accessed 9 Nov 2026