Limit orders around news only fill when they should not
Placing limit orders just before a release looks smart. A buy limit below the market catches a dip, a sell limit above catches a spike. Cheap entries, guaranteed price.
The catch is adverse selection. In the event of a fast move through the limit, the order fills. That move, if it keeps going, produces a loser. If the move is a brief spike that reverses, the limit often does not fill because the spike did not reach it, or fills and gets a great entry.
The pattern in practice: limit orders fill most reliably when price is moving through them with momentum, and that momentum usually continues. The orders that would have been the best, the ones at the extreme of a spike, are the ones that often miss by a tick.
So the average limit order near news ends up filled in the wrong situations and missed in the right ones.
This is a common property of limit orders in all markets, but news amplifies it dramatically.
A review of fill rates on limit orders around releases, split by outcome, tells the story clearly.
What fraction of the limit orders placed around news turned out to be winners compared with those that never filled?
Limit orders around news only fill when they should not
-
FXS HongKong Scalper
- Posts: 543
- Joined: Sat Oct 24, 2026 10:15 pm
Re: Limit orders around news only fill when they should not
The adverse selection point is the core of it. A limit that fills on a release fills because the market ran through it, which is the scenario where the trade is most likely wrong. Pulling resting limits a couple of minutes before high-impact data is a simple rule that avoids most of this.
FXS Official Team | Hong Kong Session Desk
-
Amanhasnoname
- Posts: 2060
- Joined: Fri Oct 30, 2026 9:45 am
Re: Limit orders around news only fill when they should not
I went through exactly this with resting limits on the US jobs report. For a couple of months I left buy and sell limits around ten pips either side of the price before the release, thinking I would catch the knee-jerk overshoot. When I finally sorted the results, the filled orders were losers far more often than winners, and the handful of times the spike reversed sharply my order was usually a pip or two short of the extreme.
The thing that changed my view was splitting fills by how fast price reached the order. Fills that came within the first few seconds were mostly bad, because price was running straight through the level with no hesitation. The few fills that came later, after the first burst had already exhausted itself, were the ones that worked.
These days I cancel everything about three minutes before high impact data and only consider a limit once the first minute has printed and spreads have come back to normal. I miss some nice entries, but the account stopped bleeding on news days.
The thing that changed my view was splitting fills by how fast price reached the order. Fills that came within the first few seconds were mostly bad, because price was running straight through the level with no hesitation. The few fills that came later, after the first burst had already exhausted itself, were the ones that worked.
These days I cancel everything about three minutes before high impact data and only consider a limit once the first minute has printed and spreads have come back to normal. I miss some nice entries, but the account stopped bleeding on news days.
Re: Limit orders around news only fill when they should not
That rule has paid for itself on my account. I went back through six months of logs last year: 41 limits resting through tier-one releases, 23 filled, and only 7 of those filled trades were winners. The 18 that never filled would mostly have been the good ones, exactly the adverse selection problem. With 0.5 lots and a 6 pip stop, the losers alone came to roughly 480 dollars, which is a lot for a setup that felt clever.FXS HongKong Scalper wrote:Pulling resting limits a couple of minutes before high-impact data is a simple rule that avoids most of this.
What I do now is close to what Amanhasnoname describes. Everything resting gets cancelled at T-2 minutes, and I don't place a new limit until the first one-minute candle has closed and the spread on EURUSD is back under 0.3. If price is still moving more than 5 pips a minute at that point, I simply don't trade the release.
The only exception I make is a limit far enough away that it would need a 25-30 pip run to reach, and even then I halve the size.
It’s Fairman 