Automatic MT5 Journaling: What Gets Recorded and How

“Automatic journaling” sounds like it should be simple: read the trade history, write it to a table. In practice MT5's history is not a list of trades — it is a list of deals, and a trade only exists once two of them are paired up correctly. Most of what makes automatic sync worth trusting, or not, happens in that pairing step.

A trade is two deals wearing one position ID

MetaTrader doesn't store “trades.” It stores deals— one for opening a position, one for closing it — and both carry the same position ID. A script that wants a single record per trade has to walk the deal history, find the entry deal and the exit deal that share an ID, and merge them into one row: entry price and time from the opening deal, exit price and P&L from the closing one.

Get that pairing wrong and the failure is quiet. A half-matched trade doesn't crash anything — it just shows up with no entry price, or gets silently dropped, and the journal looks fine until someone notices a trade from last Tuesday never made it in.

Where the stop-loss actually comes from

This is the field most automatic journals get wrong, because the obvious place to read it — the closed position itself — doesn't have it anymore. The stop and take-profit that were set live on the orderthat opened the position, not on the deal. Reading the wrong ticket space for that order is an easy mistake: an order ticket and a position ticket look like the same kind of number, but on most brokers passing the position ticket where an order ticket belongs silently returns nothing, and every synced trade comes back reporting “no stop” — even for trades that had one.

There's a real limit here too, not just a bug to fix: a stop moved later by dragging the line on the chart, rather than through a new order, doesn't always leave a history record on every broker. What an automatic sync can honestly report is the stop the trade was opened with — which, for judging the decision rather than the management, is usually the more honest number anyway.

What a script gets right that manual entry doesn't

• It never rounds.Entry, exit, volume and P&L come from the broker's own record, not from someone reconstructing them from memory at the end of the day.
• Open positions update on their own.If a stop gets moved while a trade is still running, the next sync just re-sends the current position state — there's nothing to remember to go back and edit.
• The order comment survives.Anything typed into an order's comment field at entry — a setup name, a shorthand for the reason — rides along as the closest thing MT5 has to a stated reason for the trade. Empty for anything opened by hand, which is worth keeping as a fact rather than filling in after the fact.

Why re-sending the same trade twice has to be safe

A sync that runs on a timer, rather than on a trade-closed event, will always overlap with itself a little — a trade that closes in the same second as the last sync's cutoff needs to be picked up again on the next pass, or it falls through the gap entirely. The fix is cheap on one side only: the client re-sends a small overlap window every cycle, and the server has to treat a resend as an update to the same row, keyed on the position ticket, not as a new trade. Get the second half wrong and a quiet market produces a journal with duplicate entries instead of missing ones.

The part worth checking yourself

Anything that reads a trading terminal and talks to a server is worth being skeptical of by default, and the honest answer to “can I trust this” isn't “trust us” — it's being able to read the thing yourself. getALPHA's MT5 sync script ships as plain, commented source, not a compiled file: it sends open positions and closed trades outward on a timer, and that is the entire list of things it does. It never receives instructions back, never places an order, never modifies one — there is no code path in it that could.