Position, swing, and day trading use different decision schedules
Holding periods overlap, so the decision rule matters more than the label. Fidelity describes swing trades as lasting days or weeks and position trades as potentially lasting months or years. This page uses a multiweek trend example; it does not require a losing trade to stay open for a minimum number of weeks.
| Style | Typical objective | Decision schedule |
|---|
| Day trading | A move within one session | Intraday decisions and a session-end exit rule |
| Swing trading | A shorter price swing over days or weeks | Review the setup as its local target or invalidation approaches |
| Position trading | A broader trend over weeks, months, or longer | Daily entry checks, scheduled broader reviews, and standing risk controls |
A frozen weekly-trend and daily-breakout rule set
This is a hypothetical long-only research specification for one liquid stock on standard daily candles, using regular-session prices and the exchange calendar. It is not a proven strategy or a recommendation. Use one open position, no additions, and a fixed initial stop. Record the symbol, price-adjustment setting, session, and timezone before testing.
- Allow at least 21 completed weekly bars for the weekly average comparison and 20 prior daily bars for the entry and stop windows
- The example holds through scheduled earnings; test that exposure explicitly instead of silently excluding losing event trades
- Keep the protective stop active after entry. If an entry gaps below the stop, the position needs an immediate exit; record the actual fill and do not pretend the planned stop capped the loss
- A holiday changes the first trading day of the week. Define the review by the exchange week, not an assumed Monday session
| Decision | Rule to implement |
|---|
| Weekly background | For each trading week, use the previous completed weekly close and its 20-week simple moving average. Permit entries only when that close is above the average and the average is above its value one completed week earlier. |
| Daily entry | While flat and the weekly filter is true, require a completed daily close strictly above the highest close of the preceding 20 daily bars, excluding the signal bar. Submit a market entry for the next regular-session open. |
| Initial invalidation | At the signal close, freeze the lowest low of the preceding 20 daily bars, excluding the signal bar, as the protective stop. Reject missing data, a stop at or above the signal close, or a position size below one share. |
| Position size | For this research example, use the smaller of whole shares allowed by a 1% equity risk budget divided by signal-close-to-stop distance and whole shares allowed by a 20% equity exposure cap. Fix quantity when submitting the entry; an opening gap can change actual risk and exposure. |
| Reassessment day | On the first daily close of each new exchange week, inspect the previous completed weekly close. If it is at or below its 20-week average, exit at the next regular-session open. Do not wait for this review if the protective stop triggers. |
| Time exit and priority | Count the entry bar as holding bar 1. If still open at daily close 60, submit an exit for the next open. The first filled exit ends the trade; cancel remaining exit orders. Do not create a new entry on a bar with an exit fill. |
A worked entry and an overnight-gap loss
Assume $50,000 equity, a $500 price-risk budget, and a $10,000 exposure cap. The completed weekly close is $99, its 20-week average is $95, and the prior weekly average is $94.50: the filter passes. A daily close at $100 exceeds the prior 20-day highest close of $98; the stop window low is $92. These are invented prices for arithmetic, not historical results.
- A stop-market order prioritizes getting out after triggering; it does not guarantee the stop price. A stop-limit order can remain unfilled
- TradingView documents that a price order crossed in a gap can fill at the next bar open in its emulator. Actual broker fills can differ again
- If the next-open entry were $103, the same 62 shares would have $682 of price risk to $92 before costs. Signal-time sizing cannot guarantee the intended $500 budget
- If entry and stop occur within the same daily candle, inspect the assumed order of prices and available lower-timeframe data before accepting the simulated fill
| Calculation | Result |
|---|
| Risk-based quantity: floor($500 / ($100 - $92)) | 62 shares |
| Exposure-based quantity: floor($10,000 / $100) | 100 shares; use the smaller quantity of 62 |
| Assume the next-open entry fills at $100 | $6,200 exposure; planned price loss at $92 is 62 × $8 = $496 |
| Assume a later session opens at $88 and the stop fills there | 62 × ($100 - $88) = $744 loss before costs, or 1.488% of starting equity |
| Extra loss beyond the $92 stop-price calculation | $744 - $496 = $248 |
Count holding costs over calendar days
The baseline example buys shares with cash and has no margin-loan interest. A financed variant needs its own cost ledger. For an illustrative constant $3,100 loan at 8% simple annual interest for 60 calendar days on an actual/365 basis, interest is $3,100 × 0.08 × 60 / 365 = $40.77. This is an assumed rate, not a broker quote; 60 trading bars generally span more than 60 calendar days.
- With a hypothetical $620 gross gain and $12 total commissions and execution costs, the financed example nets $620 - $12 - $40.77 = $567.23 before tax and distributions
- Use the actual loan balance, daily rates, accrual method, and broker netting policy. SEC guidance explains that these terms differ between brokers
- Keep dividends and price adjustments consistent so a distribution is not counted twice. Borrow fees, FX rollover, futures rolls, and perpetual funding need separate instrument-specific models
- Commission and slippage settings do not by themselves prove that a simulation includes every holding cost. Reconcile omitted amounts separately
Test a long enough span without borrowing future information
A daily chart can contain thousands of bars but only a few completed position trades. For an illustrative historical split, develop on 2015 through 2021, compare frozen parameter candidates on 2022 through 2023, and reserve 2024 through 2025 for a final check. This is a proposed research design, not a backtest we ran. Data availability and corporate histories can limit the usable span.
- Use only completed weekly values on daily bars. A developing weekly close can change; check that the same signals survive a chart reload
- Report trades, holding duration, exposure, drawdown, and cost-adjusted results for each period and market regime. More calendar years do not guarantee an adequate or independent sample
- At each split boundary, choose a flat-start test with earlier bars used only for warm-up, or a continuous portfolio test with carried positions. Label the choice and do not mix the results
- Fix order timing and costs before comparing parameters. Review opening gaps, halts, sideways whipsaws, and losses concentrated around earnings
- Include open-position exposure at the end of the test. A closed-trade report can miss the unrealized loss on a position that has not exited
Implement the specification, then compare a small parameter grid
Give the frozen rule table to Pine Script AI Coding Agent and inspect the generated strategy in TradingView, especially weekly data timing, stop activation, and the next-open orders. After that baseline matches the written rules, Strategy Optimizer can compare exposed inputs. For example, weekly averages of 20 or 30 bars and daily breakout windows of 20 or 40 bars produce four candidates; keep all other rules fixed.
- Select candidates using development and comparison data, then evaluate the reserved period separately without retuning
- Use Backtest Deep Report with a supported TradingView trade export to inspect trade count, profit factor, and drawdown computed from closed trades. Compare that view with TradingView open-position and intratrade risk
- Record the selected inputs, date ranges, costs, and rejected candidates. Parameter comparison does not establish a profitable edge or execute a broker order
Keep a core holding separate from tactical trades
Trading around a core position combines a longer holding with a separate trading portion. Give each portion its own size, invalidation, and records, then measure their combined exposure. A tactical loss does not justify moving its shares into the core account or increasing the core allocation. The single-position example above does not model this two-part portfolio.
