Decision Tree for a Hike That Is Ninety Percent Priced In

[Fairman]

Decision Tree for a Hike That Is Ninety Percent Priced In

A central bank meets tomorrow. Interest rate futures imply a 90 percent probability of a 25-basis-point hike. You hold no position and want to understand the possible reactions.

Lay out the branches. Branch A, with about 90 percent probability: the bank hikes as expected. Branch B, about 10 percent: the bank holds.

Within branch A, the statement and press conference matter. A1: the bank hikes and signals more tightening to come. A2: the bank hikes and signals it is probably done. In A1, the currency may rise modestly, because the guidance is more hawkish than the baseline. In A2, the currency may fall despite the hike, since the market wanted more.

In branch B, a surprise hold with a dovish tone would likely cause a sharp fall in the currency, because 90 percent of the move was already in the price. A surprise hold with hawkish language could produce confusion, with a brief drop and a recovery.

So the decision tree shows that the hike itself carries little information. The information lies in the guidance, and in the 10 percent tail.

How should you act? Staying out until the press conference is the typical answer. If you must hold a position, size for a violent move and wide spreads.

Notice that probability and payoff differ. A branch with 10 percent probability may carry a larger move than the branch with 90 percent.

Your turn: draw the tree for the next meeting in your calendar, with the implied probabilities.

[FXS HongKong Scalper]

Many traders forget that in a priced in hike, the guidance matters more than the decision. Preparing for each branch in advance, with possible entries and exits, helps avoid reacting emotionally in the moment.

[Shadow Trader]

Export the history and plot how long each trade lasted. A scalper who thinks of the strategy as "two minutes in, two minutes out" often finds a surprise.

A tidy cluster around the intended duration is a good sign. A second cluster of much longer trades is a warning. Those are usually losers that were allowed to run past their stop, or trades that were forgotten, held in the hope of recovery.

Compare average profit for each cluster. Frequently the short trades are profitable and the long ones destroy the total.

A cluster of extremely short trades, a few seconds, often corresponds to panic exits or accidental clicks, and some brokers' minimum-hold rules would flag them.

The distribution also helps set rules. If most winners are closed within ninety seconds and the winners after that point are rare, a time stop at ninety seconds converts hopeful holding into a clean exit.

Has the time-in-trade histogram revealed a hidden second strategy mixed in with the first?

[Fairman]

The decision tree makes the information content of the decision clear. With 90 percent of a hike already priced, the hike itself tells the market almost nothing new. The guidance, whether more hikes are coming or this is the last one, carries most of the information in branch A. And branch B, the surprise hold, carries a large move because it contradicts what was priced. So a trader focusing on whether the bank hikes is focusing on the least informative part of the event. The useful preparation is to think about what the guidance might say and how that compares with what the market expects.

[Amanhasnoname]

The tree is a useful habit, and what changed my behavior was putting a pip range on each branch. Before a meeting I write the scenarios on a card. A priced hike with a hawkish tone I expect to drift maybe 20 pips. A hike with a hint that it is the last one I expect to move 40 pips against the currency. And the 10 percent hold I write as 80 pips or more, because everyone is leaning one way.

On a European meeting last year the EUR/USD spread went from 0.8 pips to over 4 in the first seconds, and the first candle went the opposite way to the next hour. A normal stop would have been gone.

So I place nothing before the press conference, and when I do enter it is at a third of my usual size with a stop twice as wide. The card keeps me from chasing the first candle.

[Fairman]

The A2 branch, where the bank hikes but signals it's done, is often the most counterintuitive for newer traders. The currency falls on a rate hike, which seems backwards. But markets trade expectations, and if the market expected more hikes to follow, a signal that this is the last one lowers the expected path of rates. That's dovish relative to expectations, even though the decision itself is hawkish. This pattern, sometimes called a dovish hike, has occurred at several central banks over the years and has caught many traders off guard.

[Fairman]

The probability and payoff point is important. The 10 percent branch might produce a move several times larger than the 90 percent branch. If the hike causes a 20 pip move and a surprise hold causes a 150 pip move, the expected move from the tail branch is 15 pips, comparable to the expected move from the main branch at 18 pips. So the tail branch contributes as much to the overall expected volatility as the main one. Anyone holding a position through the event is exposed to both branches, which is why sizing for the tail, as the post suggests, is essential.

[Fairman]

Staying out until the press conference is a sensible default for most traders. Spreads widen sharply at the moment of the decision, and the initial reaction can reverse as the statement and press conference are digested. For the ECB, the decision is released at 14:15 WAT and the press conference begins at 14:45 WAT, now that Central European Time has returned to standard time. For the Fed, the decision comes at 20:00 WAT in winter, with the press conference at 20:30 WAT. Has anyone here traded a dovish hike or a surprise hold? Sharing what happened would help others prepare.