Skip to content

04 · Event-Driven Trading

Event-driven trading (Event-Driven Trading): positioning around high-impact events at known times (earnings, NFP, CPI, central bank decisions, elections, geopolitical conflicts). Its biggest difference from other styles: you don't need to predict direction — you prepare before the event happens and execute your plan after — "buy the expectation, sell the fact" is the core of this logic.

But it is also the window where retail traders lose money fastest: the gap at the moment of a data release can erase a month of profit in one second. This article covers how to prepare, how to participate, how to control position size, and the most common ways to die.


1. Event Taxonomy: Which Events Are Worth Trading

Event typeTypical examplesInstruments affectedPredictability
Earnings seasonA-share earnings previews/reports, US quarterly reportsIndividual stocks, related sectorsTiming certain, outcome uncertain
Macro dataNFP, CPI, GDP, PMI, unemploymentEquity indices, FX, gold, TreasuriesTiming certain, market expectations exist
Central bank decisionsFed/ECB/BoJ rate meetingsGlobal risk assets, FX ratesTiming certain, path expected
Supply-side eventsOPEC+ meetings, cut/boost decisionsCrude oilTiming roughly certain
Political eventsElections, referendums, policy releasesThe country's assetsTiming certain
Geopolitical conflictsWar, sanctions, supply cutoffsSafe havens (gold), energy, shippingTiming uncertain, most sudden
Crypto project eventsToken upgrades, mainnet launches, large airdropsThe specific tokensTiming roughly certain

Classification (determines your response):

TypeDefinitionHow to handle
Certain time + uncertain outcomeEarnings, CPI, rate meetingsPosition in advance via an event calendar, trade the expectation gap
Uncertain time + huge impactGeopolitical conflicts, black swansCannot pre-position; rely on position control and "never ride naked"
Certain good/bad newsForced delisting, unambiguous policyUsually already priced in — beware sell-the-news

2. Preparing in Advance: Event Calendar and the Expectation Gap

2.1 The event calendar

Build your own event calendar 1-2 weeks ahead (sources: economic calendar sites/apps, exchange announcements, company investor-relations pages):

markdown
## This Week's Event Calendar (example)
| Date | Time (Beijing) | Event | Affected instruments | Market expectation | My plan |
|---|---|---|---|---|---|
| Monday | 21:30 | US Nonfarm Payrolls | Gold/USD/equities | +200k jobs | Trim before data, watch reaction after release |
| Wednesday | 02:00 | Fed rate decision | Risk assets | 80% chance of +25bp hike | No adds, wait for dot plot |
| Friday | 09:30 | Company X earnings | That stock | EPS expected 1.00 | No gambling before earnings, follow trend after |

2.2 Expectation and expectation gap: why "buy the expectation, sell the fact" works

Price reflects "expectation", not "fact". When good news is fully expected by the market, it has already been bought; at the release moment, unless the result beats expectations, price has no reason to keep rising — profit-takers cash out instead. That is "buy the expectation, sell the fact".

Numeric example (gold vs NFP):

text
Event: US Nonfarm Payrolls, market expects +200k jobs
Before: consensus sees strong employment → USD index and Treasury yields rise early,
        gold falls from 1980 to 1930 (gold-negative expectation, already priced)
Release: actual +180k, below the 200k expectation
Reaction: expectation gap = actual − expected = −20k (USD negative)
        gold spikes from 1930 to 1960 (buying the fact: below expectations = gold positive)

Three outcomes of the expectation gap (historical statistics, not predictions; defer to actuals):

CaseResult vs expectationTypical move
In line with expectationGap ≈ 0Reversal or drift after release — the news was priced
Beats/misses materiallyClearly better/worse than expectedA run along the surprise direction (but the first leg often fakes)
Double whammyData bad itself + worse than expectedLarge one-sided move (e.g. CPI beat + prior revised up)

💡 The core concept of event trading

The market trades the gap versus expectations, not whether the number itself is "good". A "good" print that merely matches expectations carries no new information.

2.3 Where to find expectations (executable checklist)

Event typeExpectation sourceNotes
Macro dataEconomic calendars' "consensus estimate" column, Bloomberg/Reuters surveysInstitutional forecast medians exist weeks ahead; track revisions too
Fed ratesCME FedWatch tool (implied probabilities from rate futures)Hike/cut probabilities are a direct reading of the expectation gap
Central bank pathDot plot (FOMC), officials' speechesWording (hawkish/dovish) often matters more than the decision
EarningsAnalyst consensus (EPS/revenue median)Compute the gap as "actual vs consensus"
Oil/OPEC+Reuters surveys, producer officials' trial balloonsCut-size expectations revise dynamically with the news

Usage discipline:

  1. Track "expectation revisions", not just the final figure. If consensus revises from 250k down to 180k a week before release, an eventual 200k may be "below the original estimate" but "above the latest expectation" — completely opposite directions;
  2. The more unanimous the expectation, the more dangerous. With the whole market aligned (e.g. 100% hike odds), any deviation produces a big move;
  3. Decide on the difference between "my view vs market expectation", not on "good or bad data" — your personal opinion is worthless against market pricing.

3. Earnings Trading

3.1 A-share earnings previews / US quarterly reports

ScenarioDescriptionTypical play
A-share earnings previewFixed schedule; results quality knowable weeks earlyPreview games usually complete before the announcement; gap-up-then-fade on preview day is common (sell-the-news)
US quarterly reportsReleased pre/post-market; first trading day often gaps 5%-15% (historical statistics; defer to actuals)Gambling before vs chasing after earnings — see below

3.2 Gambling direction before vs chasing trend after: win-rate common sense

ApproachWin-rate common sense (historical statistics)P&L profileMain risk
Gamble direction before earningsNear coin flip (~50%)Big gap profits when rightEqually big gap losses when wrong; implied volatility expensive (options carry high Vega)
Chase trend after earningsSlightly higher (55%-60%, historical statistics)Clear information, follow throughThe gap has eaten most of the profit, risk-reward worsens
Fade after earnings (bet sell-the-news)Depends on "whether expectations were stretched"More effective the fuller the expectationJudging "fullness" is hard; easy to catch falling knives

Practical advice:

  1. Don't gamble direction before earnings. Coin-flip odds + rich implied volatility makes long-run negative expectation near-certain (especially buying options: post-earnings IV crush can lose money even when direction is right — historical statistics; defer to actuals).
  2. After earnings, only trade "pullbacks after breakout" or volume-confirmed trends — never chase the first jump. Wait for direction to clarify 15-30 minutes after the open and for confirmation signals.
  3. Cap per-trade size (see Part 5); gaps can jump straight over your stop-loss level — use "stop-loss level + gap contingency" as double protection.

4. Macro Events: Volatility Patterns Around Releases

4.1 The 30 minutes before release: the calm before the storm

TraitBehaviorResponse
Volume shrinksBig players watching, volume well below normalDon't open new positions in this window
Narrow oscillationPrice bounces in a tiny rangeExisting positions face "two-sided stop hunts"
Stops sweptInstitutions hunt stops during thin liquidityMove stops away / trim before release

4.2 The release moment (first 1-5 minutes): the first wave often fakes

  • Price lunges one way instantly, then often gives back sharply or reverses within minutes — the first wave is liquidity vacuum plus programmatic front-running; its direction isn't necessarily sustainable (historical statistics; defer to actuals).
  • Response: don't chase the first wave. Observe 5-15 minutes after release, wait for the second wave's direction (in most cases only the second wave reflects the real expectation gap).

4.3 The 30 minutes after release: direction clarification window

PeriodTypical pattern (historical statistics)Trading window
0-5 min afterInstant gap + first lungeChasing forbidden
5-15 min afterGiveback/reverse probe, then re-selecting directionWatch and confirm
15-60 min afterTrend unfolds along the true directionCan enter with the trend; widen stops by ATR

⚠️ The first wave at release often fakes — don't chase it

The first wave right at data release is often a fake-out. Price lunges one way instantly, then often gives back sharply or reverses within minutes — the first wave is liquidity vacuum plus programmatic front-running, and its direction isn't necessarily sustainable. Watch 5-15 minutes after release and wait for the second wave's direction before acting.

Volatility traits of major events (historical statistics, not predictions; defer to actuals):

EventTypically affected instrumentsVolatility profile
NFPGold, USD, equity indicesExtreme instant volatility; "spike then fade" common
CPITreasuries, USD, equities, cryptoSurprise = broad risk-off, effects lasting days
Fed decisionEverythingTwo phases: decision instant + press conference (wording matters more than the rate)
OPEC+ meetingCrude oilCut/boost magnitude versus expectations sets direction

5. Position Discipline for Event Trading: The Half-Position Rule

Event moves are 2-3× normal volatility (historical statistics, not predictions; defer to actuals) — the same position doubles its risk in event markets. The answer is one rule: the half-position rule.

CaseNormal positionEvent-market position
Per-trade risk≤ 1% of capital≤ 0.5%
Daily risk≤ 2%-3% of capital≤ 1%-1.5%
LeverageNormalHalved or none
Positions held into dataHold normallyTrim or hedge before data (reduce directional exposure)

Additional discipline:

  1. Stops must be placed in advance — there is no time to act manually at the gap instant; resting orders are the only execution guarantee;
  2. No new positions in the 30 minutes before release (leave that knife fight to quant firms and market makers);
  3. One event, one direction — no "clever cross-instrument hedging" — event reactions routinely exceed your model;
  4. Two losing events in a row → stop and review. Event samples are small; consecutive losses mean your expectation model is wrong, not that luck is bad.

💀 The half-position rule

Event moves are 2-3× normal volatility — the same position doubles its risk. One answer: the half-position rule — per-trade risk from ≤ 1% of capital down to ≤ 0.5%, daily risk from ≤ 2%-3% down to ≤ 1%-1.5%, and leverage halved or dropped. Place stops in advance; open nothing new in the 30 minutes before release.


6. Ways Event Traders Die: Checklist

DeathScriptAntidote
Front-runningHeavy bet 30 minutes pre-data, swept out by the fake first waveNo adds/new positions before data; wait 15 minutes after release
Chasing the first waveJumping into the gap at release, catching the exact reversal pointWait for second-wave confirmation; enter 5-15 minutes later
Data double whammyBeat expectations + prior revised up, losses amplifiedHalf-position rule — even a double whammy finds "no position to kill"
Sell-the-news distributionNews merely matches expectations, gap-up fades and traps youRemember: in line = no new information; don't chase highs on release day
Gambling pre-earningsBuying options/heavy before earnings; IV crush + wrong direction = double lossDon't gamble pre-earnings (Part 3); follow confirmed signals after
No event calendarForgot CPI is tonight; gap blows through your stopBuild a weekly event calendar; check positions and stops before events
Riding naked into geopoliticsFully invested, unhedged; one sudden move resets everythingKeep a cash buffer always; black swans can't be predicted but margin can be reserved
Over-modeling eventsBelieving "hikes must fall, cuts must rally"Markets trade the expectation gap: a fully-priced hike rising instead is common

7. Full Event-Trading Workflow (Action Checklist)

text
[1 week before] Build/update the event calendar; note affected instruments and market expectations
[1-3 days before] Fix your plan:
   ① What is the expectation? (calendar consensus value + recent revision direction)
   ② For each scenario — beat/in-line/miss — what exactly will I do?
   ③ Position: compute amounts under the half-position rule
[1 hour before] Check:
   □ Stops placed for existing holdings? Away from pre-data sweep zones?
   □ Any new-entry plan scheduled "15+ minutes after release"?
[Release instant] Do nothing; observe the first wave
[15-60 min after] Execute per plan:
   Confirm second-wave direction → enter with trend at half position → widen stop by ATR
[After] Review: directional call, entry timing, position vs plan

8. Quick Reference: Common Event-Trading Fallacies

FallacyReality
Bad data = must fallDirection depends on "data vs expectation", layered with liquidity behavior at release; no guaranteed move
Positioning early beats chasing lateEarly positions must survive pre-data volatility and stop sweeps; late entries give up some spread but get clearer signals
Hikes fall, cuts rallyA priced-in hike can flip to "bad news exhausted = good"; what's traded is the gap, not the event itself
Everyone makes money in earnings seasonEarnings volatility is zero-sum against institutions and algorithms; retail's edge is discipline, not information
Event trading requires insider infoPublic sources (calendars, consensus, probability tools) suffice; tips are mostly noise
One big event defines the yearNever double down after a single-event loss (stop after two straight); event trading is not a core strategy
Rush in as soon as data dropsFirst-wave fake-outs are the norm; wait 5-15 minutes and confirm the second wave

⚠️ Risk Warning

Event markets are the most violent and least controllable by gaps of all regimes; history is full of traders who called direction correctly yet were badly hurt by oversized positions or missing stops. All win rates, volatility magnitudes, and timing windows here are historical statistics, not predictions; defer to actual market conditions; treat event data and consensus values as subject to official release channels and market consensus. Participate only with money you can afford to lose — leveraged trades in event markets can wipe principal in an instant or produce a negative balance.

Further Reading

For study and research only — not investment advice. Markets are risky.