Trailing stop spam and the order modification trap
An aggressive trailing stop that modifies every pip or two feels smart. Each modification is a request that travels to the server, gets processed, and can be rejected, delayed, or throttled. On a busy tape, those requests pile up.
Some brokers cap modifications per minute or per day. Cross that threshold and the account may get warnings, temporary blocks on modification, or even a flag for abusive behavior. Meanwhile, the stop sits at an outdated level while price runs back against the position.
There is also a subtler issue. Every modification near the stop-level boundary risks an "invalid stops" error, and many simple scripts do not handle that error. The result is a stop that stopped trailing without any alert.
Better practice is modifying in larger steps, such as every few pips or on a time basis, and logging each result code. Alternatively, handle the exit on the client side with a market close when price crosses a level, accepting slippage for lower message traffic.
Which approach has held up best on your platform, server-side trails or client-side exits?
Trailing stop spam and the order modification trap
Trailing stop spam and the order modification trap
It’s Fairman 
-
Amanhasnoname
- Posts: 2060
- Joined: Fri Oct 30, 2026 9:45 am
Re: Trailing stop spam and the order modification trap
Trailing every pip and a half looked tidy until the broker throttled the modifications. On a lively news morning I hit a too-many-requests reply after forty-one stop changes in twelve minutes. While those requests sat rejected, the working stop was six pips behind price, and the retracement took most of the trade back. I had no alert that the trail had stopped moving.
I now move the server stop only every four pips, or every thirty-five seconds, whichever comes later, and I log the result code of each modify. Invalid-stops errors used to show up a few times a week; they are rare now because I am no longer brushing the minimum distance on every tiny push. When the tape is too fast for that, I exit on the client with a market order once a one-minute close crosses the trail level and accept a bit of slippage.
The slower trail feels less precise, but it is the version that still works when the server is busy. An unmodified stop that I trust has been worth more than a clever trail that silently dies.
I now move the server stop only every four pips, or every thirty-five seconds, whichever comes later, and I log the result code of each modify. Invalid-stops errors used to show up a few times a week; they are rare now because I am no longer brushing the minimum distance on every tiny push. When the tape is too fast for that, I exit on the client with a market order once a one-minute close crosses the trail level and accept a bit of slippage.
The slower trail feels less precise, but it is the version that still works when the server is busy. An unmodified stop that I trust has been worth more than a clever trail that silently dies.