Does gold rise when the dollar falls?
No, not measurably. Gold cleared its cost 45.4 % of the time against 45.2 % on an ordinary day. The gap is inside what chance produces at this sample size.
Measured on daily data from 2020-11-10 to 2026-10-12, after the cost of trading. Recomputed every time this page loads.
When the dollar index falls on the day, does gold outperform its baseline over the next 1 day?
A question only means something if it is stated precisely enough to be wrong. This is the form the engine measured.
Gold is priced in dollars, so a weaker dollar should buy more of it. The relationship is quoted so often that it is usually stated rather than checked.
482 signal days since 2020-11-10. On 219 of them gold rose more than break-even over the next 1 day (45.4%). On an ordinary day the same LONG works 45.2% of the time — so the condition adds +0.2 points.
Typical move on a signal day: +0.08% against +0.08% on an ordinary day. Worst signal day: -10.78%.
Chance: a coin that lands LONG-right 45.2% of the time, flipped 482 times, gives 219 or more hits in about 93% of runs (p = 0.9271). That is not distinguishable from luck.
Consistency: first half of the period 43.6%, second half 47.3% — not both above baseline; one half carries it.
What it would take: the excess over baseline is so small that no realistic sample would settle it — roughly 240,100 signal days for p < 0.05, against the 482 there have been. A question that needs eighty years of data is not a question with an answer.
A relationship that only appears in one year is usually that year, not a relationship.
the hit rate over the next 30 signal days falls to the baseline (45.2 %) or below — the relationship disappears out of sample
Stated before the next observation, not after. The page recomputes, so if this stops being true the answer here changes with it.
Change the condition, the asset, the direction or the holding period and measure it against the same six years.
Open this in the tester