EURUSD and Gold Correlation Measured: Same Direction on 2 of 3 Bars, and Trading Gold’s Lead Lost in 15 of 16 Setups

“EURUSD and gold move together” is one of those things everyone repeats and nobody measures. Both are priced in dollars, so when the dollar weakens, both tend to go up. The question I wanted answered is how strong that link really is on the charts people trade, and whether you can make money from it. The short answer: the link is stronger than I expected, and it is useless for predicting the next bar.

This is a companion to my gold vs EURUSD test, where the same rules ran on both markets separately. Here I compared them bar by bar.

How was the correlation measured?

• Data: 16,926 EURUSD M5 bars and 13,736 gold (XAUUSD) M5 bars. I kept only the bars that opened at the same minute in both files: 13,631 bars from June 16 to August 26, 2026.
• What was correlated: bar-to-bar percentage changes, not price levels. Price levels trend, and two trending lines look correlated even when they are not moving together bar by bar. On these same bars, correlating price levels instead of changes gave 0.91 — more than double the real 0.42.
• Timeframes: M5, then M15, M30, H1 and H4 built from the same bars, plus daily closes. A return is only counted when both markets had both bars — no gaps, no weekend jumps.
• Trades: EURUSD only, entered at the next bar’s open, 1 pip of cost per trade as measured in my spread test. Every result is split at July 15 into two halves.

How correlated are EURUSD and gold?

Pearson correlation of bar-to-bar changes. 1.0 would mean they always move together by the same proportion, 0 means no link.

Two things stand out. First, the correlation rises with the timeframe, from 0.42 on M5 to 0.57 daily. Short bars carry more noise that belongs to one market only — a gold order hitting COMEX, a euro headline — and that noise averages out on longer bars. Second, the share of bars moving the same way barely changes: about two out of three on every timeframe. On the third bar they disagree, and that third is exactly where a trader relying on the link gets hurt.

0.5 is a moderate correlation. It means gold explains roughly a quarter of EURUSD’s hourly movement (0.52 squared is 0.27). The other three quarters is the euro’s own business.

When are EURUSD and gold most correlated?

The link almost doubles when New York opens. By hour, the two strongest were 12:00 UTC at 0.70 and 17:00 UTC at 0.70. 12:00–13:00 UTC is the hour of the US data releases at 12:30, when one dollar number moves everything priced in dollars at once — my news release test found that candle is 2.5 times a normal one. The weakest hour was 22:00 UTC at 0.19, when both markets are thin.

The same happened on big bars. When gold’s 5-minute change was more than twice its average size (1,566 bars), the correlation was 0.60. On all the other bars it was 0.27. The link is mostly a dollar story, and it shows up when the dollar moves.

Does gold move before EURUSD?

This is the question that matters for trading. If gold reacted first, you could watch gold and trade EURUSD a bar later. So I correlated each gold bar with the EURUSD bar that came after it:

The correlation drops from 0.42 to 0.04 the moment you look one bar ahead. Gold has a tiny head start over EURUSD — 0.044 against 0.018 — but it is gone within 15 minutes, and it is far too small to pay a spread.

Can you trade EURUSD off a big gold bar?

I tested it anyway. The rule: when gold’s 5-minute bar moved more than k times its 14-bar ATR, buy EURUSD at the next open if gold went up, sell if it went down, and hold for 1, 3, 6 or 12 bars.

Net pips per EURUSD trade after a 1-pip cost; trades in brackets.

15 of 16 setups lost. Before costs, 11 of 16 were slightly positive — the head start is real — but the best of those earned 0.94 pips gross, less than the spread. The one winner, +0.60 pips on 37 trades, lost 2.39 pips per trade in the first half and made 2.64 in the second. That is two coin flips, not an edge. A rule that doesn’t beat its own cost is the same result I got for four of five rules in my algorithmic rule vs random entry test.

Does EURUSD catch up when it lags gold?

The other popular idea: if gold rallied over the last hour and EURUSD didn’t, EURUSD is “behind” and should follow. I measured both markets’ hourly move in units of their own volatility, and traded EURUSD toward gold when the gap was large.

Net pips per trade after a 1-pip cost; halves in brackets where the total was positive. Setups with a gap of 2.0 had fewer than 10 trades and are left out.

With 30 or more trades, betting on the catch-up lost in 5 of 6 setups. If anything the gap tended to stay open or widen: when EURUSD ignored a gold rally, it usually had its own reason. The one catch-up setup that was positive in both halves had 52 trades and 0.35 pips per trade — too thin to trust, and it flipped to a 2.79-pip loss when held an hour longer.

So what is the gold correlation good for?

• Risk, not signals. A long EURUSD and a long gold position are partly the same trade — a bet against the dollar. With a correlation of 0.5, holding both is not two independent positions. Size them together, as in my position sizing table.
• Reading the news hour. When both move hard together at 12:30 UTC, the move is about the dollar. When EURUSD moves alone, it is about the euro.
• Not for timing. Gold’s last bar says almost nothing about EURUSD’s next one. Whatever edge there is is gone inside 15 minutes and smaller than a pip.

Gold’s own behaviour hour by hour is in gold scalping session by session, and why a busy hour is not the same as a profitable one is in my session vs all-day trading test.

Frequently asked questions

Are EURUSD and gold positively correlated?

Yes. On 13,631 matched 5-minute bars from June to August 2026, the correlation of their bar-to-bar changes was 0.42 on M5, 0.52 on H1 and 0.57 on daily closes. They moved in the same direction on about two of every three bars.

Why do EURUSD and gold move together?

Both are quoted in US dollars. When the dollar weakens, EURUSD rises and gold priced in dollars rises too. The link was strongest at 12:00 UTC, the hour of US data releases, at 0.70, and weakest in the Asian session at 0.30.

Does gold lead EURUSD?

Barely. The correlation between a gold 5-minute bar and the next EURUSD bar was 0.044, against 0.42 on the same bar. Trading EURUSD after a big gold bar lost in 15 of 16 setups after a 1-pip cost.

Is the EURUSD–gold correlation stable?

Over the whole period, yes: both halves of the data were within 0.05 of each other on every timeframe up to H1. Over short windows it swings a lot — a 20-bar M5 window went from -0.66 to +0.96.

Can I use gold to confirm EURUSD trades?

Not as a trade signal on its own. As a filter on an EMA crossover, the best of 24 gold filters cut the loss roughly in half at one holding time and none made money.