Advertisement IC Markets

Why mean reversion tends to work better in the dead hours

Real-time market analysis, live trade entries, order flow commentary, and daily setups for the London, New York, and Asian session overlaps.
Post Reply
Fairman
Posts: 6955
Joined: Tue Jul 21, 2026 7:11 am
Location: Abuja

Why mean reversion tends to work better in the dead hours

Post by Fairman »

Why mean reversion tends to work better in the dead hours

During quiet hours, such as the late New York evening and the early Asian morning, price often oscillates around a mean without strong directional pressure. Few participants, little news, and a lack of large orders create conditions in which small deviations tend to revert.

That makes mean-reversion strategies more natural for this period. Entries at the edges of a short-term range with small targets can show good hit rates. The challenge lies in costs: wider spreads and weaker fills reduce the net edge.

There is also a hidden risk. Occasional breaks from the range, triggered by a headline or a large order, can produce losses that cancel many small wins. In thin markets the break can be abrupt and the stop slips.

To assess the real edge, test the strategy specifically in these hours with realistic costs and slippage. Compare the distribution of losses, not just the win rate, since the tail often decides the outcome.

If the edge survives, size down relative to daytime trading because the tail risk is larger.

Does the strategy hold up in the dead hours once realistic costs are included?
It’s Fairman :geek:
Recommended broker for automated trading & scalping IC Markets
Amanhasnoname
Posts: 2060
Joined: Fri Oct 30, 2026 9:45 am

Re: Why mean reversion tends to work better in the dead hours

Post by Amanhasnoname »

Dead-hour mean reversion on AUDNZD looked excellent until I put real costs on it. I tested the late New York to early Tokyo window across fifteen weeks. The raw win rate was sixty-eight percent with small targets at the edges of a short-term range, the kind of figure that makes automation look obvious.

Then I applied the spreads I actually paid. Average spread in that window was one point seven pips against about zero point seven five in London, and stop-outs picked up another third of a pip in slippage. Expectancy fell from plus fifty-three hundredths of a pip to plus eleven hundredths. Two range breaks, one on an official-rate headline and one on a large order, cost fourteen point six and twenty-one pips and erased dozens of small wins.

The edge barely survives, so I run it at one third of daytime size and I stand aside if the spread is wider than two pips. The tail decides the result, not the hit rate. Without those cost assumptions I would have overstated the idea.
Post Reply