An engineer sizing a beam doesn't guess. They calculate the load, choose a section, and apply a safety factor — because the load estimate might be wrong, and the structure still has to stand.

Trading is the same problem with worse data. The signal is a load estimate. Position size is the section. Drawdown tolerance is the safety factor. Most retail traders design without one: they size to the best case and fail on the first surprise.

MacroParity treats it as an engineering problem. It does not predict where price will go. It shows where currencies stand on fundamentals, whether the market has already priced that in, and what position size survives the range of outcomes you are willing to sit through.

The same signal makes one trader 30% and wipes out another. The difference is never the signal.

How the ranking is built

Eight currencies are scored on three inputs, standardised against each other and combined by weight:

Forward real rate (50%) — policy rate minus inflation, adjusted for the expected change in policy over the next six to twelve months. Capital seeks real return; this is the strongest single explanatory factor over a three to twelve month horizon.

Rate path (30%) — the direction of policy, entered by hand from central bank guidance, vote splits and market pricing. A currency with a low real rate but a tightening path can outrank one with a high rate that has peaked.

Current account (20%) — external balance as a share of GDP. A structural, slow-moving pressure on the exchange rate.

The output is a relative ranking. A score of +0.90 means a currency sits well above the group average on these inputs. It does not mean the currency is 90% undervalued.

Why price positioning matters

A macro divergence is only tradeable if the market has not already moved to reflect it. Each pair is measured against its own one-year, five-year and ten-year range.

Where macro direction and price position agree, the pair is marked ENTRY ZONE. Where the direction is right but price has already travelled, it is LATE. Where the short and long horizons disagree, it is TRAP. Where the gap is below 0.30 it is WEAK and no direction is shown — small gaps are noise.

Catalysts

For any pair the model recomputes the whole ranking assuming one central bank shifts its rate path by 50bp, 25bp, nothing, or the reverse. The output shows what happens to the gap: whether the signal holds, falls below threshold, or flips direction. This is arithmetic on the inputs — not a forecast of how price will react to the decision.

Position sizing

The sizing page works backwards from the drawdown you accept carrying. It sets an invalidation level at the edge of the one-year range, measures the distance in pips, and derives the largest total position that survives that move. The ladder spreads entries into the move rather than at a single price.

Limitations

Valuation gaps persist. A currency being cheap on fundamentals says nothing about when — or whether — the market will agree. Gaps have stayed open for years.

The rate path column is a judgement. It is entered by hand from central bank communication. It is not a market-implied probability and it can be wrong.

Weights are not fitted. They are set by hand, informed by the literature but not optimised against historical data. Changing them changes the output.

Developed markets only. In emerging markets the mechanism inverts: a high real rate signals risk premium rather than strength, and the model reads the sign backwards. TRY, ZAR, MXN and similar are deliberately excluded.

Flows are invisible here. Positioning, risk appetite, intervention and capital flows drive short-term price action and none of them appear in these three inputs. JPY and CHF in particular can trade against their fundamentals for long stretches.

Prices are end of day from Yahoo Finance and may differ from your broker's quotes by several pips.