US dollar to South Korean won, from the ECB reference rate of Friday 6 November, re-anchored to the live market six times each working day.
Measured on the last three years of daily fixes, refreshed with every run.
Chance each fix beats the latest rate of ₩1,358.73. A higher rate means more KRW for every USD you convert. Model estimates, not promises.
No day this week is better than even odds to beat the latest fix.
Next fix forecasts roughly tie with a "no change" guess. Five sessions out they roughly tie, and the range caught 79% of outcomes against an 80% target.
| Date | Forecast | Actual | Miss | Range |
|---|---|---|---|---|
| Fri 6 Nov | 1,344.69 | 1,358.73 | 14.040 | Outside |
| Thu 5 Nov | 1,338.58 | 1,344.69 | 6.110 | Inside |
| Wed 4 Nov | 1,344.59 | 1,338.58 | 6.010 | Inside |
| Tue 3 Nov | 1,342.90 | 1,344.59 | 1.690 | Inside |
| Mon 2 Nov | 1,335.27 | 1,342.90 | 7.630 | Inside |
| Fri 30 Oct | 1,337.29 | 1,335.27 | 2.020 | Inside |
Set a level and we'll tell you on the day the reference rate crosses it.
One US dollar bought as little as ₩1,324.45 on Fri 23 Oct and as much as ₩1,558.09 on Wed 1 Jul, a swing of 16.4%. On a $10,000 transfer, that gap is worth about ₩2,336,400.
Our forecast for Monday 9 November is ₩1,358.73 per US dollar, with an 80% range of ₩1,347.00 to ₩1,370.56. The latest ECB reference rate was ₩1,358.73 on Friday 6 November.
The five session path drifts sideways to ₩1,358.73 by Friday 13 November, a change well inside the normal range, so treat it as steady. On this pair, five session forecasts currently roughly tie with a simple no change guess.
No day in the next five sessions is more likely than not to beat the latest fix of ₩1,358.73, so the model does not favour waiting.
Across 990 next fix forecasts the average miss was 6.012, against 6.013 for assuming no change, and the real rate landed inside our range 79% of the time.
The next fix starts from the latest market rate: before the fix we anchor to the live rate, after it to the fix itself. A drift is added only when the last 250 sessions show a statistically strong trend, which matters for crawling currencies like the lira and not for USD to KRW right now. The range comes from a volatility model that reacts to recent swings, calibrated so 8 in 10 outcomes land inside it. Nothing about the direction is guessed: on daily data no model we tested beats the market's own price, and we publish the score to prove it.
No. Banks and transfer apps add their own margin. Use the forecast for the direction and size of the move, not as your final quote.