Programmable Trading Discipline / 可编程交易纪律

Definition

Programmable trading discipline is a precommitted decision system that defines what may be traded, which observable states trigger entry or exit, and how much exposure is allowed before live market emotion can rewrite the rules.

Current Synthesis

写一份《纪律.md》，代替我入市投资 uses “纪律.md” to turn Trend Following from a chart-reading idea into an operating contract. Data collection, signal generation, task delivery, position targets, exception handling, and execution can be delegated to software or a fixed operator, but the investor still owns rule design, market selection, risk tolerance, and review. The purpose is not to automate certainty; it is to preserve a chosen repeated-game structure when profit, loss, commentary, and fear would otherwise change behavior.

Key Claims

• The discipline begins with an eligible universe constrained by access, liquidity, asset class, and position capacity.
• Entry and exit rules must cover ordinary transitions as well as skipped states, bad data, failed orders, and market-specific exceptions.
• Instrument size, portfolio exposure, and opportunity density belong in the same rule set as the signal.
• Automation can separate observation and execution from public opinion, but it cannot prove that the underlying strategy has an edge.
• A mature discipline should become simpler after experience identifies which exceptions matter and which parameters merely fit history.
• Long-run adherence is part of the strategy because frequent post-loss parameter changes destroy comparability across repeated trades.

Evidence

• Rule architecture - 写一份《纪律.md》，代替我入市投资 specifies market scope, entry and exit states, and position management as the three layers of “纪律.md.”
• Execution separation - 写一份《纪律.md》，代替我入市投资 describes end-of-day signals, email delivery, opening orders, and a fixed executor who reduces discretionary interference.
• Exception completeness - 写一份《纪律.md》，代替我入市投资 says rules must address abnormal data, order failure, skipped market states, and conflicting conditions.
• Statistical boundary - 写一份《纪律.md》，代替我入市投资 keeps low win rates, payoff asymmetry, repeated trials, and overfitting risk inside the operating model.

Counterevidence & Qualifications

The episode reports one practitioner’s method and does not provide audited performance, complete signal definitions, transaction-cost analysis, or an independently reproducible backtest. Clear code can execute a bad strategy consistently. Delegation also creates operational, data, brokerage, and supervision risks, while a discipline designed for one person’s capital and tolerance may be unsuitable for another.

What Changed

• Created a concept separating explicit trading operations from the broader ideas of trend following and investment risk management.

Related Concepts

• Trend Following - supplies the market-state and right-side entry logic used by the discipline.
• No-Prediction Trading - explains why the rules can operate without certainty about the next trade.
• Position Sizing - converts signals into survivable exposure.
• Stop-Loss Discipline - defines the loss-bounding exit mechanism.
• Quantitative Overfitting - constrains parameter refinement and historical optimization.
• Human Agency Under AI - keeps objective and acceptance responsibility with the human even when software executes the workflow.