Why orders get rejected right at the start of a new minute
Some scalpers notice a strange cluster of rejections or delayed fills in the first second or two after a minute rolls over. It looks random until the timestamps get lined up.
Several things can converge there. Many algorithmic strategies fire on bar close, so a wave of orders hits the same liquidity at the same instant. Price feeds can also stall slightly while the platform builds the new candle. Bridge software sometimes batches requests on a timer, which creates a small traffic jam.
For a scalper entering on bar-close signals, the result is a systematic disadvantage. Everyone is reacting at the same moment, and the book thins out for a beat.
One fix is to randomize entry timing slightly. Waiting two or three seconds after the close, or entering on a tick-based trigger inside the bar rather than at the boundary, often improves fills noticeably.
It will not turn a bad setup into a good one, but it removes a hidden tax that most backtests never include.
Has anyone compared fill quality at second zero against second five on the same strategy?
Why orders get rejected right at the start of a new minute
-
Amanhasnoname
- Posts: 2060
- Joined: Fri Oct 30, 2026 9:45 am
Re: Why orders get rejected right at the start of a new minute
I ran the same bar-close signals two ways for five weeks so I could answer that comparison properly. Version one sent the order at the first tick of the new minute. Version two waited until roughly the fourth second. The immediate version collected nineteen rejections or requotes and a median adverse excursion of six tenths of a pip on the fills that did go through. Eleven of those fills were more than one point two pips worse than the close that triggered them.
Waiting four seconds cut rejections to three and the median adverse move to a quarter of a pip. I did miss seven fast bursts that were gone by the time I joined, which is the cost of standing out of the pile-up. Even so, expectancy rose from eight hundredths of a pip to nineteen.
I now delay any bar-close entry until somewhere between the third and fifth second, or I trigger inside the bar on a tick rule so I am not lined up with every other close-based expert. The setup did not change. The hidden tax at second zero did.
Waiting four seconds cut rejections to three and the median adverse move to a quarter of a pip. I did miss seven fast bursts that were gone by the time I joined, which is the cost of standing out of the pile-up. Even so, expectancy rose from eight hundredths of a pip to nineteen.
I now delay any bar-close entry until somewhere between the third and fifth second, or I trigger inside the bar on a tick rule so I am not lined up with every other close-based expert. The setup did not change. The hidden tax at second zero did.