The textbook rule says a stronger dollar means weaker gold. In practice, there are regular periods when both rise at the same time, and recognising them avoids wrong-footed trades.
When gold and USD tend to rise together:
Acute risk-off: both serve as safe havens, so a sharp equity sell-off can lift both.
Central bank buying: steady official-sector demand for gold can outweigh dollar strength for months.
Geopolitical shocks: flight-to-safety flows into both at once.
When the classic inverse link is strongest:
Days driven by US rate expectations, where real yields move clearly.
Calm markets with no safe-haven demand.
For intraday traders, the practical check is simple: before using DXY as a signal for gold, look at the rolling correlation over the last 20 days. If it is near zero or positive, the usual rule is not working at the moment.
Have you seen periods where gold and the dollar moved together, and how did you adapt?
FXS Prop Desk wrote: look at the rolling correlation over the last 20 days.
Worth computing that on returns rather than price levels, because two trending series show a high correlation on levels even when the day to day moves aren't linked at all. Daily percentage changes over 20 days is the usual version. Even then, 20 points is a small sample, so a reading of 0.2 or minus 0.2 is basically indistinguishable from zero. A rough rule of thumb is that with 20 observations you need something beyond about 0.45 in either direction before it's meaningfully different from no relationship. For intraday use, a correlation on M15 returns over the last few sessions might be more relevant than daily, since the link can flip around big risk events within a week.
Fairman, the returns version is what I use too. Price level correlation on gold and DXY looked sticky near 0.6 for weeks last spring while daily returns flipped between positive and negative around risk events.
I keep a 20 day return correlation on both daily and M15, and I treat anything inside plus or minus 0.4 as noise the way you described. Below that I stop fading gold just because the dollar is bid.
img_13662_r1.png (88.18 KiB) Viewed 105 times
The illustration marks three short episodes where both gold and USD rose together during equity sell offs. In those windows my usual inverse hedge was the wrong trade twice before I added the filter.
Central bank buying weeks are slower but the same idea applies. If the link is positive I trade gold off its own levels, not off DXY.
Have you switched to M15 returns as the default for intraday, or do you still start from the daily read? I still glance at daily first, then trust M15 for the session.
Something worth checking before trusting any gold versus DXY number is what DXY actually holds. The euro is about 57.6 percent of the index and the yen about 13.6, so gold against DXY is mostly gold against EURUSD with a little yen mixed in.
When Tokyo is the driver, that matters. Gold can fall while DXY looks flat because the yen leg moved, so I keep a separate return correlation against USDJPY for those sessions.
For the M15 version there is a data trap too. Most gold feeds stop for an hour around the New York rollover, so the first bar after the break carries a stale return that can swing a short window by itself.
I drop that bar and the one before it from the calculation. With only a few sessions of M15 data, one bad return out of a couple of hundred was enough to move my reading by 0.1 to 0.15.
Has anyone tried gold against the US 10 year real yield instead of the dollar? It has been steadier for me on rate days but useless in a pure risk off flush.
Gold, the Swiss franc and the Japanese yen are all labelled safe havens, yet each reacts to different triggers. Treating them as one trade overlooks important differences.
Gold responds strongly to real interest rates and to central bank buying, so it can rise during periods of dollar weakness even without fear in the market. The franc is influenced by Swiss National Bank policy, which has included currency intervention and negative rates in the past, and by European risk events given the close economic ties. The yen's haven status comes from Japan's large net foreign asset position, which can lead to repatriation during stress, but a wide interest rate gap can offset that behaviour.
A scalper watching USD/CHF and USD/JPY alongside gold can notice when the three disagree. If gold climbs while both pairs stay flat, the move may be rate-driven rather than fear-driven. If the franc and yen strengthen together with equities falling, defensive demand is more probable.
Limits remain. Haven flows can reverse quickly once the triggering headline fades, and the yen has at times weakened during turbulence when rate differentials dominated. Evidence from one episode rarely transfers neatly to the next.