The five highest-conviction setups on our desk right now, the numbers behind each, and exactly what would tell us we're wrong.
Money goes where it gets paid the most. Almost everything below is a version of that one sentence.
Think of each currency as a savings account in a different country. The American account currently pays about 4.35% a year. The Swiss account pays 0.2%. If you had serious money to park, where would it go? Everyone else is asking the same question, and to buy the American account you first have to buy dollars. That demand is the move.
The number that matters isn't the central bank's headline rate, it's the two-year government bond yield: what the market expects the rate to average over the next couple of years. And direction beats level: a gap that is widening pulls money in, a gap that is closing pushes it back out.
Every gap quoted on this page is a two-year yield, measured at the 24 July close. One caveat if you hold a position overnight: the interest that actually lands in your account follows the central banks' rates today, not the two-year, so the cash figure runs a little below the gap shown. On GBP/JPY the two-year gap is 2.9% while the policy rates are 2.75% apart.
There are two versions of that yield and they answer different questions. The nominal yield is the number you're actually paid, and it's what pushes money between currencies week to week. The real yield is what survives inflation, and it tells you whether being paid is worth anything. When the two disagree, that gap is usually the interesting part. Australia pays more than Britain on paper, 4.68% against 4.42%. But Australian inflation runs 4.0% against Britain's 2.6%, so in real terms sterling returns 1.82% and the Australian dollar just 0.68%. Same two currencies, opposite answers.
Two of the pairs below don't involve the dollar at all. Those work the same way, just between two other countries. When both currencies are weak, you're trading which one is weaker, not which one is strong.
Every economic release runs through the same five steps. Once you can see the chain, headlines stop being noise.
Step three is where most people go wrong. Data only matters if it changes what people expect the central bank to do. But the chain breaks most often at step four, and this week gave a perfect example. Australia posted a jobs number nearly five times better than forecast, and its expected rates did rise. America's rose by the same amount over the same days, so Australia never ended up paying relatively more, and its currency fell to an eleven-week low anyway. Step four is always about the gap between two countries, never about one on its own.
When the rate channel gives nothing either way, slower forces fill the space: how nervous markets are, and where a currency sits on the risk spectrum. That matters most for currencies tied to global growth, like the Australian dollar. But it stays a secondary driver. The clearest evidence is that gold fell rather than rose through this month's escalation, the opposite of what genuine fear would do.
Two things happened this month, and both push prices up rather than down.
The United States put new tariffs of 10 to 12.5 percent on goods from roughly sixty countries. At almost the same time, conflict in the Middle East pushed oil above $100 a barrel, though it has since fallen back below $90 as the strikes paused. Tariffs make imported goods more expensive. Expensive oil makes transport, manufacturing and heating more expensive. Neither is caused by people spending more. They're caused by things costing more to make and move.
That distinction matters more than almost anything else here, so it's worth being slow about it. A central bank has one real lever: it makes borrowing more expensive, which makes people and businesses spend less. When prices are climbing because everyone is spending too freely, that lever goes straight at the cause and it works.
This month's inflation didn't come from spending. Oil costs more because of a war. Imports cost more because of tariffs. No interest rate decision puts oil in the ground or cancels a tariff, so the only lever they have is pointed at the wrong problem. Rates could still drag prices down, but only by choking demand hard enough to damage the economy, and no central bank reaches for that willingly. So they wait: prices rising, and nothing comfortable to do about it.
Meanwhile the inflation figures that have already been published have been falling. So the market is holding two contradictory ideas at once. Yesterday's data says inflation is cooling. Tomorrow's costs say it's about to climb again. This is the week that gets tested.
The market has already made up its mind about something: it expects US interest rates to sit roughly half a percent higher a year from now. That's already baked into the dollar's price today. Knowing it isn't an edge.
That changed over the weekend. Oil above $100, the tariffs and a surge in AI-driven demand have pushed markets to price roughly a 40% chance the Fed actually raises rates on Wednesday, up from near zero when this was first written. US bond yields hit multi-year highs and traders are now discussing dissenting votes if the Fed holds.
So there are two questions on Wednesday, not one. Do they move? And if they hold, do they sound worried enough to keep that 40% alive? A week ago only the second question existed.
The event is not the trade. The gap between what was expected and what actually gets said is the trade. A hold that everyone predicted moves nothing. The same hold with one unexpected sentence attached moves everything.
Our desk rates thirty-five pairs. These are the five highest-conviction ones, ranked, and every one is marked with its conviction level so you can see where it sits. Reasons, not instructions. No entry prices, on purpose. Tap any pair to open it.
No dollar in this one, and here the interest payment is the whole trade. UK government bonds pay about 2.9% more than Japanese ones, and you collect roughly that much a year just for holding the position while Japan sits still.
Switzerland pays savers almost nothing. America pays over four percent. If you're holding a large amount of money, one of those is obviously better, and to get it you buy dollars and sell francs. That's the entire trade, and it's the widest gap the dollar holds over any major currency.
This is the one that teaches the most, because on paper it looks wrong. Australia's economy is doing well. Its jobs report this week beat forecasts by nearly five times. Australia even pays more interest than America does. And the currency fell to an eleven-week low anyway.
Canada was hit with new US tariffs alongside about sixty other countries, on top of a separate threat aimed only at them. Tariffs make Canadian goods more expensive for Americans, so Americans buy fewer of them. Fewer sales means a slower Canadian economy and less reason to hold Canadian dollars.
The second pair with no dollar in it. Both the euro and the yen are soft, so this comes down to one question: which one is weaker. The answer is the yen. German government bonds pay about 1.4% more than Japanese ones, and that gap widened again this week.
Two concentrations to see before sizing anything. Three of these five are long dollars. USD/CHF, AUD/USD and USD/CAD all improve if the dollar rises, so Wednesday's Fed hits all three at once. And two of them are selling the yen: GBP/JPY and EUR/JPY both lose badly if Japan intervenes on Friday. That's five positions but really two bets. Size the bets, not the tickets.
For each event, the outcomes worth preparing for and what they'd do to our five pairs. Percentages are our desk's own odds where it publishes them.
Arrows show which way the pair moves, not whether it's good or bad for you.
A week ago this was near zero. Oil above $100, the new tariffs and AI-driven demand have put it firmly back on the table, and US bond yields have hit multi-year highs pricing it in.
The strongest outcome for the dollar side of the board. Note it is still the less likely of the two branches, so a hold is not a surprise.
Still the more likely branch. But a hold is no longer the quiet outcome it was last week, because the market has to decide what the hold means.
Hold and sound worried: the 40% stays alive, the dollar keeps its bid, and the board behaves much as a hike would, just less violently.
Hold and sound relaxed: that 40% comes out of the price fast, and it has further to fall than it did a week ago. Three of five go against you at once. This is now the sharpest risk on the page.
Either way the two yen pairs barely react. That is exactly why they are on the board, and it matters more this week than last.
The market prices this as a near certainty, so the decision itself won't move much. Mildly yen-negative, which suits both pairs.
The government buys its own currency. A large, fast, one-directional move that hits every yen pair at the same moment. Our desk's instruction is blunt: close both, don't buy the dip. Interventions aren't a level to fade.
This is the biggest risk on the page, and it can land any day, not just Friday. Two of your five positions die together.
Nobody positions for this one. The interest gap narrows sharply and the reason to own either pair thins out.
Already in the price, so muted. Worth noticing it quietly contradicts a hawkish Fed from the night before.
Validates the hawkish path. The best outcome for the dollar side of the board.
If the Fed sounded tough Wednesday and this lands soft Thursday, the market has to unwind a view it took on eighteen hours earlier. Those are the sharpest moves you get.
Every economist surveyed expects this. The pound keeps its big interest advantage over the yen and the trade carries on earning.
UK business activity beat sharply. If the Bank leans into that rather than dismissing it as a weather-and-football blip, sterling's advantage widens further.
Government borrowing is already above plan. If it breaks the stated rules you'd get a fiscal shock and a carry unwind on the same pair at once. That's the worst case here.
Canada earns more, and nervous markets sell growth-linked currencies.
Hurts the Canada trade, helps the Australia one. A partial natural hedge.
This has now happened. Strikes paused for a second night and Brent fell back below $90 from above $100. The dollar softened and the Australian dollar rallied on the risk-on move, which runs against the AUD short in the short term.
Australia pays 4.35%. America pays 3.75%. Australia pays more, had a jobs report five times better than forecast this week, and its currency still fell to an eleven-week low.
The market doesn't price where rates are, it prices where they're going, and only relative to everywhere else. Australia's expected rates rose this week, but America's rose by the same amount, so the gap didn't move. No change in the gap means no new reason to buy.
Three of these pairs are long dollars and two are short yen. If the Fed surprises Wednesday, three lose together. If Japan intervenes, the other two lose together. Five tickets, two outcomes.
The mistake is sizing each one as though it's independent, then discovering on Wednesday evening that you had three times the dollar exposure you thought.
Anything on this page with a dotted underline links straight to the article it came from. Everything below opens in a new tab.
Britain's Home-Grown Price Pressures Finally Easing for the BOE
. Private-sector pay growth below 3% for the first time since 2020 and services inflation falling for a year. The Monday morning dent in the GBP/JPY case.
Will the Fed Pivot on Rates or Just Get Stuck?
. Markets pricing close to a 40% chance of a rise this week, and Brent back below $90 after a calmer weekend. The source for the reframed Wednesday.
Bond Traders on Edge as Risks of Fed Rate Hike This Week Mount
. US Treasury yields at multi-year highs as traders price the hike risk.
Fed Faces Growing Pressure to Hike Rates as Price Risks Rebound
. Oil, tariffs and AI-driven demand rebuilding inflation risk, with dissent possible if the Fed holds.
Oil Tumbles as US and Iran Pause Military Strikes
. Brent under $90 on the pause, the dollar softer and the Australian dollar firmer. The de-escalation scenario actually happening.
Australian Prime Minister Albanese to Press Trump on New Tariffs
. Australia among roughly 60 economies hit by the 12.5% tariff, with the prime minister appealing directly.
US Business Activity Picks Up on World Cup, July 4 Spending
. US business activity at an eight-month high of 53.6, above the 50 line separating growth from contraction.
Charting the Global Economy: Growth Picks Up, AI Boom Continues
. US unemployment claims at a near 60-year low, giving the Fed room to stay focused on inflation.
Wall Street Bulls Are Staring Down $100 Oil, Tariffs, AI Angst
. Oil above $100 and new US tariffs of 10–12.5% across roughly 60 economies. Explicitly names the Australian dollar as weakening on slower global growth.
Trump's Tariffs Are Likely to Stick Around Despite Unpopularity
. Tariffs look permanent because they raise revenue and provide negotiating leverage. Behind the USD/CAD case.
Euro-Zone Resilience and More Inflation Seen in First Data for ECB
. Q2 growth of 0.2% forecast with July inflation at 2.9%, supporting the case for further European rate rises. Behind EUR/JPY.
Yen's Slide Bolsters Appeal of Hawkish Bank of Japan Hedges
. A very weak yen has major banks expecting faster Bank of Japan rate rises, and raises the odds of government intervention. The main risk to both yen pairs.
US Pauses Iran Strikes for Second Night as Red Sea Tensions Rise
. The most recent development, and the main reason oil sits where it does.