In the years before central banks began publishing intervention receipts on quarterly schedules, the pattern desk had to read the tape sideways — through cross-currency correlations, order-book depth, and the specific rhythm of a bounce that felt too clean to be organic. That rhythm has not changed. When the won printed near 1,550 against the dollar and pulled back within hours, traders old enough to remember the defenses of 1997 and 2008 recognized the choreography immediately. A managed hold at a round figure, absorbed offer-side liquidity, a coiled reversal. The exit — not the entry — is where the pattern breaks the most accounts.
The Bounce-Is-The-Top Fallacy
Here is a pattern I keep seeing when a defended figure holds. A round number gets tested, the tape prints a rejection wick, and within an hour the retail Telegram groups have declared the top. Traders who caught the bounce from 1,550 down to 1,528 close out — pleased with the win, convinced the intervention is over.
They are usually wrong about the top. And here is what nobody in those groups will tell you: the first pullback after a defended figure is almost never the terminal exit. It is the pause between the two waves. What you saw was the first order-book flush. What you did not see was the second-tier corporate hedgers waiting for confirmation before they bring their offers in size. That second flush comes on the next session open, not in the first six hours.
If you have traded through a Bank of Korea defense window before, or read the 1997 and 2008 debriefs closely, you know the shape. A round figure holds, price retreats twelve to twenty pips, then coils sideways as Asian corporate flows re-set their hedge ratios. The exit signal is not the first bounce. It is when the coiling breaks and volume returns without the central bank being visible in the order book. That usually takes twenty-four to seventy-two hours to confirm.
The trader who closes on the first pullback books a real profit. Fine. The problem is what happens next. Convinced the bounce was the top, they re-enter short at 1,532 expecting continuation. When the second wave reverses back toward 1,545 on real flow rather than intervention flow, the re-entry stops out, and the round-trip commission has now eaten most of the original gain. This is where the pattern desk sees the majority of intervention-day accounts close red on days when directionally they were correct.
The Commission Drag That Kills Late Exits
The second pattern is quieter and more expensive. It is not about direction. It is about which broker you are exiting through, and how the pricing model interacts with the wider spreads that always print during an intervention window.
Let me show you the math, because this is the part that hides itself. The historical commission-model brokers — Pepperstone standard, IC Markets standard, Exness Pro-tier — separate spread from cost. You pay a small raw spread plus a fixed round-turn commission. The marketing-led brokers of the same period, the ones running zero-commission structures like the standard-tier accounts at Exness or FBS or the AvaTrade retail books, embed cost inside the spread markup.
On a normal-liquidity day the difference is small enough to argue about. On an intervention day it stops being small.
Take five standard lots held through a suspected KRW intervention bounce. Under normal conditions, a marked-up-spread broker prints a one-pip spread, blows out to roughly four pips during the peak intervention window. Four pips times five lots times ten dollars per pip is two hundred dollars on entry, another two hundred on exit — call it four hundred dollars round-trip in spread cost alone, zero commission. Now the commission-model broker: normally 0.1 pip raw, blows out to maybe 0.8 pips during the same window. That is 0.8 times five lots times ten, or forty dollars entry, forty exit, plus roughly seven dollars per lot round-turn commission across five lots, which is thirty-five dollars — total one hundred fifteen dollars.
Four hundred versus one hundred fifteen. On a suspected intervention trade that captures thirty pips gross — thirty times five times ten equals fifteen hundred dollars — the marked-up book nets eleven hundred and the commission book nets one thousand three hundred and eighty-five. The commission structure delivers roughly twenty-six percent more retained P&L on the same gross move, purely from execution mechanics during the wide-spread window.
Now compound that. The trader who mistakes the first bounce for the top, exits, re-enters against the second wave, and stops out has paid the round-trip twice. On a marked-up book that is eight hundred dollars in spread drag against a fifteen-hundred-dollar gross. On the commission book it is two hundred thirty. The trader who does not know the pattern is not just wrong about the exit. They are wrong through the pricing structure that punishes them hardest for being wrong.
The bounce is not the top. The bounce is the moment your broker's pricing model starts telling you the truth about what you actually pay to be in this position.
The Regulator-Signal Substitute
Here is where I need to concede something before I take it apart. The instinct to weight tier-one regulation heavily is not stupid. It is defensible. Segregated client funds, negative-balance protection under FCA and ASIC rules, published capital adequacy — these things matter. When a broker fails, the reader with a tier-one regulator on their account statement recovers a higher percentage of their balance than the reader who chose an offshore-only book. The 2015 Swiss franc episode confirmed this, and no honest analyst will tell you otherwise.
Fine. Now here is what the tier-one badge does not tell you, and what the retail flow gets systematically wrong on intervention days.
Regulatory tier and market-structure execution are two different things. FCA authorization tells you what happens if your broker goes insolvent. It tells you almost nothing about how your broker's dealing desk handles a widening-spread event at the top of an intervention window. It does not tell you whether your stop will be filled at your requested level or ten pips beyond. It does not tell you whether your broker's liquidity provider is warehousing your flow or passing it through. It does not tell you whether the commission structure you signed up for actually holds during the specific fifteen-minute window when the Bank of Korea last defended the 1,550 line.
The pattern here is worth naming. Retail traders substitute the regulatory question — is this broker safe from insolvency — for the execution question, which is entirely separate. AvaTrade holds ASIC, FSCA, ADGM, CBI and FSA authorizations, and on a normal Tuesday the spread and slippage figures are entirely defensible. That does not tell you what the execution profile looks like at 1,550. Exness Pro-tier under FCA authorization holds raw spreads near a tenth of a pip on quiet days. That does not tell you what the raw spread looks like during a defended figure test.
The commission-versus-spread choice is a market-structure choice, not a regulatory one. The historical broker disclosure documents on this — the ones from 2011 to 2016 that separated dealing-desk practice from custodial arrangements — made the distinction cleanly. The retail marketing after 2018 collapsed the two, because it sold better. Do not let the collapse become your framework.
The Second-Wave Reversal Pattern
Look at the tape from suspected KRW interventions across the last two decades and a specific shape recurs. Defense holds. First pullback. Coiling for six to eighteen hours. Second-wave test back toward the defended figure. Third-day confirmation, either a break or a hold, on volume that no longer needs the central bank to appear.
This is not unique to Korea. The same rhythm shows up in the historical Bank of Japan defenses of 1998, in the JPY intervention rhythm of 2022, and, with different numbers, in the emerging-market defenses of 2013 and 2018. The pattern is not a coincidence. It is what happens when a central bank absorbs a specific tranche of speculative selling, corporate hedgers wait to see whether the defense will hold overnight before adjusting their hedges, and then the real-money flow re-tests within one to three sessions.
The exit windows for this pattern sit at three distinct points. Window one is the first-wave pullback, twelve to eighteen hours after the defended figure. Window two is the second-wave failure, forty-eight to seventy-two hours later, if the defense holds. Window three is the break-of-defense scenario, which requires a different playbook and is not the modal outcome.
Most retail flow tries to catch window one and re-enter for window two. The pattern desk sees this attempt liquidated a majority of the time, because window two is not always short-directional. If the defense held and the second-wave selling gets absorbed, window two is where dollar weakness begins — and the re-entry short becomes the wrong side of the trade. The historically consistent exit is window one for the trade you took, then flat until window three confirms which regime you are in.
So What Do You Actually Do
Exit the first-wave bounce and stop trading it. That is the whole discipline. Book the gain, close the platform, and do not re-enter until either forty-eight hours have passed and the second-wave direction is confirmed, or the defended figure has broken decisively on volume that does not need central bank absence to survive.
If you are still in the trade past the first pullback, run the math I walked through above on your own broker. Pull up your spread and commission history for the last three high-volatility windows and reproduce the numbers with your own lot sizing. The commission-versus-marked-up-spread difference is not theoretical. It is in your own trade log, and you can measure it. If you cannot measure it because your broker does not disclose the raw-spread and commission breakdown, that itself is the answer to the broker question. The historical commission-model books published this data. That was the point of the model.
And here is the question I do not have an answer to, because nobody in the public tape has confirmed it. When the Bank of Korea intervened around 1,550, we do not know how much of the absorbed offer was warehoused by state-linked commercial banks versus passed through to the interbank ladder. The rhythm of the defense suggests warehousing — the coiling is too clean, the second-wave absorption too orderly. But the receipts have not been published, and the quarterly foreign-reserve disclosure will not resolve it cleanly. If you sit closer to that flow and can see the answer, write it up. It is the missing piece of the pattern that has held for three defenses now, and none of us on this desk can close it from the tape alone.
FAQ
How do I know a bounce is intervention-driven and not organic dollar weakness?
The tape reads differently. Intervention bounces show up as a hard stop at a round figure, a rejection wick that is disproportionately large versus recent hourly ranges, and a pullback that coils sideways for six to eighteen hours rather than trending. Organic dollar weakness carries through into cross-currency correlation — EUR, GBP and JPY move in sympathy. If USD/KRW reverses while the DXY basket does not, that is the shape of a managed defense, not a broad dollar unwind.
Is the KRW 1,550 level a hard line the Bank of Korea will always defend?
No, and treating any figure as a permanent line is how traders get run over. Round-figure defenses are tactical, not structural. The 1,550 print gets defended when reserve levels, current-account posture and political tolerance for depreciation align. When any of those three shifts, the figure gets released. The historical record on Korean won defenses shows the figure moves — 1997 was defended at a very different level than 2008, which was different again from 2022.
What broker structure actually matters for holding through intervention windows?
The specific things worth checking are execution disclosure during widened-spread windows, whether the commission structure is separated from spread pricing, and whether the broker publishes historical slippage data. Tier-one regulation matters for custody. It does not tell you what your fill will look like at the top of a defended figure. Historical commission-model books — Pepperstone standard, IC Markets standard, Exness Pro-tier — separate these questions cleanly. Marked-up-spread structures embed the answer inside the pricing, which is why they cost more during the exact windows that matter.
If I miss the first-wave exit, should I hold for the second-wave move?
Usually no. The second-wave outcome is bimodal — either the defense holds and the direction reverses against your original trade, or the defense breaks and the move accelerates past your original target. Retail flow assumes the second scenario. The historical record suggests the first is more common on tactical defenses. If you missed window one and you are still in the position, the honest move is to size down to a fraction of your original exposure and set the exit at a level that does not require you to be right about which of the two scenarios plays out.
How much does the commission-versus-spread difference actually cost on an intervention day?
On the five-lot example walked through above, roughly two hundred fifty dollars round-trip against fifteen hundred dollars of gross P&L — about a sixteen-percent haircut on the marked-up structure versus the commission structure, purely from execution mechanics during the widened-spread window. Compounds badly if you re-enter and take the round-trip twice. Run it on your own broker's actual spread and commission history rather than trusting either the marketing figures or the round numbers here.
Does negative-balance protection matter for intervention trades?
Yes, but only in the tail scenario where the defended figure breaks catastrophically and your stops do not fill. Negative-balance protection is a solvency question, not an execution question. FCA and ASIC-regulated brokers provide it, which is one of the reasons the tier-one badge is worth something. It is not, however, a substitute for sizing the position such that you do not need the protection to trigger. Position sizing is the first line of defense. The regulator is the second.
What is the specific signal that the defense has actually broken versus held?
Volume returning to the pair without the central bank being visible in the order book, and cross-currency correlation re-establishing. If the won continues weakening while regional Asian currencies stabilize or reverse, the defense is still holding tactically and the pressure is idiosyncratic. If the won weakens in sync with a broader regional move, the defense has been released and you are in a different regime. That confirmation usually takes two to three sessions to become clean on the tape.