Why does gold mean reversion lose so much on trend days?

Because the stretch trigger is satisfied repeatedly all the way up or down. A trader filtering on stretch alone takes the trade three or four times in the same direction while the anchor follows price, and each loss is larger than the last.

How can I tell a rotational gold session from a trend day before entering?

Count session VWAP crosses and read structure. Two or more crosses, overlapping candles and no consecutive higher highs with higher lows means rotational. One cross or fewer, or over ninety minutes on one side of VWAP, means trend.

Should I fade gold after a CPI spike?

No. A release moves gold's fair value instantly while session VWAP moves slowly, so a 3.0 ATR reading can be correctly priced. Wait forty-five to ninety minutes for new rotational structure, and anchor VWAP from the release.

Is it wrong to add to a losing gold reversion trade?

Yes. Price extending against the entry is evidence that the excursion was not thin flow, which is the one condition under which the position should be smaller. One entry per excursion, with no exceptions.

Why do my small reversion trades lose money even when they work?

Because a $0.30 round-trip spread is 7% of a $4.20 target. Require both a minimum $5.00 distance to the anchor and a minimum 2.0 ATR stretch before entering.

What is the earliest sign that gold mean reversion is degrading?

A rising share of time-stop exits. When trades stop reaching the anchor within four 15-minute candles, gold is shifting from rotation to trend, usually one to two weeks before the hit rate falls. Halve size above 35%.

When Gold Mean Reversion Fails — XAUUSD Trend Day Failure Modes

Gold mean reversion fails on trend days where the stretch trigger is satisfied all the way up or down, when excursion volume is expanding, when a release reprices VWAP itself, when stretch is too small to clear costs, when a trader averages down into the excursion, and around rollover where the anchor is unreliable.

Gold mean reversion fails in six ways: trend days where the stretch trigger is satisfied repeatedly while session VWAP follows price, excursions carrying above-average volume that migrate the anchor towards price, tier-one releases that reprice fair value while VWAP lags, stretch too small for a $0.30 round-trip spread to be worth paying, averaging down into an excursion, and rollover or early-session windows where VWAP has too little volume to be a meaningful anchor.

The dominant mode is the trend day, and it is fully detectable before entry: fewer than two session VWAP crosses, consecutive higher highs with higher lows on the 15-minute chart, large candle bodies with one-sided wicks, and over ninety minutes on one side of the anchor.

Minimum entry conditions are 2.0 ATR and $5.00 distance to the anchor, flat or declining excursion volume, no tier-one release within ninety minutes, and an established session VWAP — no entries before 02:00 UTC or after 21:00 UTC.

Degradation is monitored on a rolling twenty-trade window through hit rate, median stretch at entry, the share of losses that were rule violations rather than valid entries, and the share of trades exiting on the four-candle time stop. Time-stop exits above 35% signal a shift from rotation to trend.