Three AI-generated opportunities do not create room for three trades. If your position-sizing rule permits only one, the disciplined decision is to approve one at most and reject or defer the others.
In April 1970, Apollo 13 was losing electrical power roughly 200,000 miles from Earth. After an oxygen tank failed, NASA flight director Gene Kranz and teams in Houston faced a hard constraint: the spacecraft had more systems than its remaining power supply could support.
Every system had a purpose. Some could help with navigation, communication, temperature control, or the eventual return through Earth’s atmosphere. Yet keeping everything active would consume the power needed later. NASA had to preserve limited capacity for the one outcome that mattered most: bringing James Lovell, Jack Swigert, and Fred Haise home.
NASA’s history of the mission, Apollo Expeditions to the Moon, documents how controllers developed a reduced-power plan and protected the command module’s batteries for re-entry. They could not approve every useful demand on the system. Capacity decided.
A beginner looking at three plausible trades faces a smaller version of that decision. The account has limited risk capacity. Three opportunities can be valid on their own while the portfolio can safely hold only one.
Position size comes before opportunity count
Suppose an account holds $800 and the trader’s rule limits planned loss to 1% of equity per trade. That creates an $8 risk budget.
Position size then depends on the distance between entry and invalidation. If a setup enters at $40 with a stop at $39.60, the planned loss is $0.40 per share. Dividing the $8 risk budget by $0.40 permits 20 shares, before accounting for fees, slippage, or platform constraints.
The same calculation applies to each AI-generated opportunity:
Position size = permitted account risk ÷ risk per unit
The AI may identify three setups that meet its signal criteria. The position-sizing rule answers a different question: how much exposure can the account carry if the trade fails?
A signal describes an opportunity. A risk rule sets the boundary around acting on it.
Three separate trades may share one source of risk
The morning becomes harder when all three setups point in a similar direction. Perhaps they involve three technology stocks reacting to the same market move, or several crypto assets rising with Bitcoin.
Treating them as independent can hide concentration. One reversal may hit every stop within minutes.
A beginner with room for $8 of planned loss cannot automatically assign $8 to each trade. Doing so raises total planned loss to $24, or 3% of the $800 account. Correlation could make that exposure behave like one larger bet.
That is why the approval gate matters. The trader can compare the queued ideas before execution:
- Which setup has the clearest invalidation level?
- Which requires the smallest exposure to stay within the risk limit?
- Are the opportunities driven by the same market move?
- Would approving one consume the day’s remaining risk capacity?
- Does existing portfolio exposure already point in the same direction?
The highest-confidence signal does not deserve an exception to the sizing rule. Confidence is an estimate. Planned loss is a boundary.
This is the same portfolio-level issue behind Marcus rejecting a fourth correlated position and Lena reviewing 11 clustered signals. Counting signals tells you how many ideas appeared. Counting shared risk tells you what the account can absorb.
Rejection is part of the trading process
Rejecting two AI-generated opportunities can feel wasteful, especially when both later move in the expected direction. That feeling creates a dangerous lesson: the trader starts judging discipline by the outcome of a single morning.
A rejected trade can rise without making the rejection wrong. The decision should be judged using information available before entry: account size, stop distance, existing exposure, correlation, and the trader’s written risk limit.
An approval-gated assistant keeps that decision visible. It can generate and queue opportunities, but the human chooses which order, if any, proceeds. Nothing needs to execute merely because a model produced a signal.
That separation also creates a useful trading journal. Record all three opportunities, the calculated position size for each, the reason two were rejected, and what happened afterward. Over time, the journal can show whether the selection rule is consistent and whether the AI’s reasoning holds up across wins, losses, and drawdowns.
Protect capacity for the decision that matters
Apollo 13’s remaining power had to be allocated according to mission survival, not according to how many spacecraft systems could make a reasonable case for staying on. The unused systems were not necessarily defective. The available capacity was smaller than the total demand.
Your account works under the same constraint. When three opportunities arrive, calculate the permitted loss first. Then examine combined exposure and approve only what fits.
Tomorrow morning, write the account-risk limit at the top of the journal before opening the signal queue. For each opportunity, record entry, invalidation level, risk per unit, permitted position size, and correlated exposure. If only one fits, approve one. If none fit, the correct position size is zero.
Educational content, not financial advice. The figures above are illustrations and exclude costs such as fees and slippage.
Comments
No comments yet.