An unexplained position should trigger containment: pause new exposure, identify its size and exit conditions, and reconstruct the decision before you decide what to do next. If nobody can state the trade thesis, invalidation price, and risk limit in plain English, the position has already escaped the process meant to control it.
At 6:42 a.m., Tomas stood in his kitchen in Chicago holding a mug that had gone cold. His brokerage screen showed a crypto position he did not recognize, opened overnight at a size larger than he usually used.
He could see the entry price. He could see the unrealized loss ticking with each update. What he could not find was the reason for the trade.
His first impulse was familiar: read the chart, find a story, and decide that the position probably made sense. A support level, an overnight move, a headline he had missed. But the trade had no written thesis, no recorded stop, and no defined point where its idea would be wrong. If the market moved further against him, Tomas risked turning confusion into a larger loss simply because closing the position felt like admitting a mistake.
This is an illustrative composite, but the situation is real enough to recognize. The danger begins when an order arrives without an explanation a trader can repeat.
A visible position can still be an unknown risk
A trade is not understood because the ticker symbol is familiar or because the chart can be interpreted after entry. Every position needs a plain-language answer to a few basic questions:
- What condition created the entry?
- Where does the idea fail?
- How much capital is at risk if it fails?
- What would justify holding, reducing, or closing the position?
Without those answers, a trader may mistake activity for a plan. The chart provides endless material for a post-hoc explanation. That explanation can feel persuasive because it arrives after the order exists.
This is why autonomous trading systems create a particular accountability problem. An order may be technically traceable while remaining practically opaque to the person carrying the risk. The Bank of England has warned that complex AI models can create challenges around predictability, explainability, and transparency, including the possibility that autonomous trading amplifies shocks without a human manager’s awareness.
For an individual trader, the immediate issue is simpler: you cannot assess a risk you cannot describe.
Containment comes before interpretation
When Tomas saw the position, he did not need to decide whether the market would reverse in the next hour. He needed to stop the situation from expanding while he gathered facts.
Containment can mean pausing any system that may add to the trade, cancelling related queued orders, and checking whether the current position breaches a daily, weekly, or per-trade risk limit. It also means resisting the urge to average down before the original risk is known.
Then document what is observable: entry time, entry price, current size, account exposure, existing stop order if one exists, and any linked orders. These details do not explain the trade, but they prevent the record from disappearing beneath a new decision.
The next step is a plain-language reconstruction. “I entered because price broke above X, the trade fails below Y, and I planned to risk Z” is a reviewable thesis. “It looked strong” is a feeling. Feelings can be useful signals to examine, but they cannot carry the whole risk decision.
The same discipline applies when a system supplies the idea. An AI-generated signal can be useful only when the human reviewing it can inspect the reasoning, compare it with their limits, and reject it without friction. An approval gate creates that pause before execution, when the trade can still remain a proposal rather than becoming a problem to explain later. For a closer look at that distinction, read Approval-gated AI vs autonomous trading bots.
The missing explanation is information
By 7:05, Tomas had found the source of the order: an old automation rule he had left enabled after a test. It had no current position-sizing rule tied to his account and no stop attached.
That discovery did not tell him where the market would go. It told him something more actionable: the position did not meet his conditions for staying open.
He reduced the exposure, disabled the rule, and wrote a brief note in his trading journal describing what he could verify and what he could not. The morning ended without a dramatic win or loss. It ended with the account back inside a process he could audit.
This is the point often missed after an unexplained trade. The absence of a reason is itself evidence. It may reveal a broken automation rule, an approval process that was bypassed, a missing journal habit, or a position-size calculation that never ran. Treat that gap as a control failure to investigate, rather than a puzzle the market must solve for you.
A later review can then focus on the process: Why was the rule still active? Which approval or alert failed? Could the account have opened another related position? What guardrail would have stopped it?
Build the explanation before the order exists
The safest time to demand clarity is before entry. A short approval checklist gives each trade a place to state its thesis, invalidation price, position size, and maximum downside in terms you understand.
The checklist does not predict price. It makes the decision visible while changing it remains easy.
Tomas returned to his desk that evening and wrote one rule on the page beside his monitor: “No plain-English reason, no open position.” The next queued signal had a chart, an entry idea, and a suggested size. It also had a missing invalidation level. He rejected it.
That rejection was a disciplined trade decision. Nothing needed to execute for the process to work.
Educational content, not financial advice.
Comments
No comments yet.