Practical trading:
Turn ideas into executable, stop‑lossable and reviewable processes.

1. First pick the market and instrument: trading opportunities carry execution costs

1.1 Liquidity is not just trading volume

When choosing an instrument, don't focus only on daily volume or a single attractive price spike. Liquidity should at least include bid‑ask spread, depth, execution stability in normal and event periods, your planned position relative to market depth, and the ability to exit when a stop is triggered. Small caps, illiquid futures, FX crosses and weekend crypto may seem tradable when calm, but can produce large slippage during news or liquidity contractions.

1.2 Spot, futures and leveraged products cannot share the same risk assumptions

Spot P&L is usually price change × quantity; futures also require contract multiplier, tick size, margin, settlement and roll; options have premium, expiry, volatility and time value. If multiplier, currency, tick size or margin specs aren't entered, the calculator can only run conceptual estimates and should not output seemingly precise, orderable contract quantities.

1.3 One asset card

Record instrument, exchange or platform, quote currency, trading hours, contract multiplier, minimum move, fees, shorting ability and primary liquidity risks on the instrument card. The card’s purpose is not prediction but to avoid misapplying rules across different markets.

2. Trading plan: turn subjective narratives into pre‑defined rules

2.1 A complete plan must answer at least six questions

• What observable conditions trigger an entry?
• Which price or event would invalidate the original assumption?
• Is the expected loss after stop execution still within budget?
• Is the target price derived from structure, volatility, or rules?
• Under what conditions do you not trade — e.g. event risk, liquidity or time‑of‑day constraints?
• What fields must be recorded after a trade to enable a proper review?

2.2 Single-trade risk/reward and historical expectation are not the same thing

A long plan can be written as risk_distance = Entry − Stop, reward_distance = Target − Entry, plan R:R = reward_distance ÷ risk_distance; for shorts, swap price directions. This only describes the geometric distance of a single trade. To estimate whether a strategy has an edge, you also need win rate, realized average profit and realized average loss:

2.3 Define "no-trade" conditions

A trading plan should not only state when to enter, but also when to abstain: estimated spread exceeds limit, stop distance so large the position becomes too small, before/after major events, signals appear only in low‑liquidity conditions, or the day has reached the preset loss limit. Plans without no‑trade rules tend to be rewritten under maximum pressure.

3. Orders & execution: price determination and fill confirmation are two different risks

3.1 Market orders, limit orders and stop orders

Investor.gov classifies common order types as market, limit and stop‑loss. Market orders typically prioritize immediate execution but do not guarantee execution price; limit orders set price boundaries but may not execute; stop‑loss orders convert to market orders when the stop price is reached, so in fast markets the execution price may differ from the stop price.[1] FINRA also cautions that quote delays, execution time and volatility can cause actual execution prices to differ from displayed prices.[2]

3.2 Write slippage into the model

If you assume entry and exit slippage as percentages, apply them directionally: long entry gets more expensive, exit gets cheaper; short entry (sell to open) cheaper, cover more expensive. Fees should be computed on actual execution amounts; fixed fees, taxes and securities lending fees listed separately. Costs are not a later “decimal”, they change break‑even win‑rates and are strategy parameters.

Can use withCost‑adjusted expected‑value toolRun single‑trade sensitivity first, then inTrading journal analyzerView the ex‑post distribution of actual costs for completed trades; pre‑trade you can usePre‑trade rules checkerVerify item‑by‑item that the plan left the necessary records.

3.3 Execution records

At minimum record signal formation time, order submission time, order type, expected price, fill price, filled quantity, reason for cancel/modify, slippage and unfilled portion. Without execution records, post‑trade review easily mislabels “no fill” as a “strategy error”, or treats favorable slippage as a reproducible edge.

4. Position sizing & risk control: decide first how much you can lose at most

4.1 Back out quantities from account risk budgeting

Position size should not be decided by "how right I feel this time"; it should be back‑calculated from the single‑trade loss you can bear. A simplified spot model is shown below; for futures, include per‑point multiplier and contract costs into the per‑unit risk:

If slippage increases loss after a stop, using only Entry−Stop will underestimate risk. The addedPractical trading plan and position checkerIncludes entry/exit costs and slippage in risk per unit, and shows overall budget utilization.

4.2 Correlation and simultaneous positions

That each trade conforms to 1% risk does not mean the account carries only 1% risk overall. Multiple positions in the same industry, the same macro factor, or the same currency direction can trigger stops simultaneously under stress. In review, calculate portfolio‑level maximum loss, same‑direction exposures, margin usage and liquidity — not just single trades.

4.3 Consecutive losses and maximum drawdown

Fixed fractional risk, fixed dollar sizing and position sizes that vary with equity produce different drawdown paths. Before trading you can run stress scenarios — e.g., 5, 10 or 15 consecutive losses — then check remaining capital, psychological resilience and whether you'd be forced to change rules. After trading useRisk‑return and maximum drawdown analysis toolDisentangle the order of peaks and troughs in an equity sequence and the recovery periods.

5. Strategy framework: different strategies fail in different ways

5.1 Trend‑following and breakouts

Trend strategies accept some small losses while waiting for a few large trends to pay the costs; a common failure is repeated false breakouts during consolidation. When evaluating, don't look only at win rate — also check whether average profit is sufficient to cover consecutive losses, slippage and holding costs.

5.2 Mean reversion and range trading

Mean‑reversion often has higher win rates, but when true regime, fundamental or liquidity changes occur, deviations can continue to widen. Treating "price returning to the mean" as inevitable leads to moving stops farther out. In backtests, examine tail losses and extreme samples, not only average days.

5.3 Pairs, grid and leveraged strategies

Pairs trading requires checking each leg's Beta, currency, trading session, multipliers and basis; grid trading requires checking inventory and margin after a one‑sided breakout; leveraged strategies require checking liquidation price, margin calls and gaps. Same name does not imply same risk — strategy descriptions must include clear stop/kill conditions.

5.4 Indicators are conditions, not guarantees of causation

MAs, RSI, MACD or volatility indicators are functions of historical data, suitable for defining observable conditions, filtering environments or building test features; they do not automatically provide future direction. Can be combined withArticle on blind spots of technical indicators與Article on backtesting and overfittingRead.

6. Backtest & validation: prevent data leakage first, then assess performance.

6.1 Minimum contract for backtest

• Data frequency, timezone, adjustment method and cut-off date are fixed.
• Signal formation and fills should be separated by verifiable execution intervals.
• Research, validation and out‑of‑sample periods separated by time.
• Include fees, spread, slippage, taxes and no‑fill rules.
• Does not use future prices, future constituents, or data revisions that became known later.
• Report returns, MDD, trade count, losing streaks and distribution simultaneously.

6.2 Don't replace stress tests with a pretty curve

The more parameters and the more times you peek at the data, the higher the risk of overfitting. Double the costs, delay the signal by one bar, reduce fills, remove the best trades and see if the strategy still holds a reasonable risk/reward. If it only works for a single precise parameter set, the edge may be sample noise.

Category 13 will continue withQuant research workbenchIntegration: require a data contract first, then discuss features and backtests — do not masquerade research pages as real‑time execution signals.

7. Position management: don't invent rules on the fly under pressure

7.1 Trailing stops and breakeven are not free options

Moving a stop to breakeven may reduce single‑trade losses but can increase the share of trades stopped out by short‑term noise; moving stops can protect part of unrealized gains but may lower average returns. These behaviors must be validated in pre‑defined rules and historical records — don’t explain them by cherry‑picking winners.

7.2 Pyramiding, scaling in/out and partial exits

Every time you add to a position you should recompute the total worst‑case risk; partial exits will change realized average profits and R:R. If a tool only accepts a single Entry、Stop、Target, the output should be treated as a simplified baseline and does not represent the full trade distribution of multiple fills.

7.3 Events and overnight risk

Earnings, central bank decisions, regulatory news and weekend markets can cause gaps or spread widening. Your trading plan should state how you handle positions before/after events, acceptable liquidity, whether to reduce size, and fallback plans if orders can’t be filled.

8. Trade journal & review: break results into causes you can improve.

8.1 Journal fields

A complete trade record should include at minimum: date/time, instrument, direction, strategy tag, entry rationale, failure conditions, planned Entry／Stop／Target, actual fills, position size, fees, slippage, exit rationale, result R, rule compliance and post‑trade notes. Screenshots can preserve context but do not replace structured fields.

8.2 Look at groups first, then the overall average

Group trades by strategy, time window, direction, volatility regime, holding period, liquidity and rule compliance. Overall positive expectation may be driven by a single strategy or a few dates; only after grouping can you tell whether the edge is concentrated, whether costs erode performance in certain periods, or whether rule deviations caused the main losses.

Paste completed R resultsTrading journal analyzer，check win rate, average P/L, Profit Factor, expectancy, max consecutive losses and R drawdown；then useSingle‑trade cost toolCompare the plan with realized results.

8.3 Post‑mortem is not post‑hoc rationalization

When reviewing, first label trades as 'within rules'/'outside rules' before checking outcomes; you must not retrospectively dress rule violations up as good decisions simply because a trade made money, nor dismiss an originally valid process as wrong just because a trade lost money. The problems you can actually improve are usually execution deviation, cost estimation, oversized positions, data definitions, or strategy failure in specific environments.

9. Class 13: continuous upgrade path

Future upgrades will prioritize adding trade record import formats, grouped comparisons, cost sensitivity, rule versioning, out‑of‑sample reports and event risk checks; boundaries of client‑side processing, no fake market data and no automatic order placement on behalf of users will be maintained.