Position Trading Strategy: Holding for Weeks and Months

What position trading is

Position trading holds a directional position for weeks to months, aiming to capture a substantial portion of a major move. It differs from investing because the thesis is about price behaviour rather than asset value, and because there is a defined exit rule rather than an indefinite holding period.

Mechanically, it is trend following or momentum applied at a slow tempo, usually on weekly bars or daily bars with long lookbacks. The defining characteristics are wide stops, small position sizes relative to account, and a review cadence measured in weeks.

A complete position trading rule set

Universe

20 to 40 liquid ETFs across equity regions, sectors, bonds, commodities, and currencies, or a broad list of large-cap stocks.

Timeframe

Weekly bars, evaluated on the Friday close. Orders placed for Monday.

Entry

Weekly close above the 40-week moving average AND above the highest weekly close of the past 26 weeks.

Initial stop

Below the 40-week moving average, or entry minus 3 x weekly ATR(14), whichever is further away.

Trailing exit

Exit when the weekly close falls below the 40-week moving average, or below the lowest weekly close of the past 13 weeks.

Position size

Risk 1 percent of equity per position. Maximum 8 positions. Maximum 25 percent of equity in any one position.

Portfolio filter

If more than half the universe is below its 40-week average, halve all new position sizes.

Review

Once weekly, after the Friday close. No intraweek intervention except for a stop that is already resting.

Carry check

Before entry, note dividend yield, futures roll characteristics, and financing costs, since these accumulate meaningfully over months.

Worked example over four months

Equity 80,000 USD, risk 1 percent, so 800 USD per position. A commodity ETF closes the week at 34.20, above its 40-week average at 30.50 and at a 26-week closing high. Weekly ATR(14) is 1.45.

Notice the trade lasted four and a half months and produced 1.56R. Position trading does not generate large R multiples frequently; it generates modest multiples on a small number of trades while consuming almost no time and very little in costs. Annual returns come from having several such positions running concurrently across uncorrelated markets.

Carry and holding costs over months

Futures roll deserves particular attention. In a market in contango, a long position loses value on each roll even if spot prices are unchanged, which is why long-term long positions in some commodity futures and in volatility products decay steadily. See futures strategies.

The psychological profile it demands

• Tolerance for open profit giving back. A position up 40 percent may return to up 15 percent before the exit triggers. That is the design, not a failure.
• Comfort with inactivity. Many weeks contain no action at all. Traders who need engagement will invent trades, which is the main way this style fails.
• Indifference to short-term news. At a four-month horizon, most headlines are noise. Following them daily produces anxiety without producing information.
• Patience with validation. With 15 trades a year, judging the strategy takes years. You must derive confidence from the backtest and the mechanism rather than from recent results.
• Willingness to hold through drawdowns. Positions will be underwater for weeks. The stop, not your comfort, defines when the thesis is wrong.

How to validate a slow strategy

1. 1

   Test across many instruments, not many years alone

   Forty markets over twenty years produces far more independent observations than one market over eighty. Breadth is how slow strategies achieve statistical significance.

2. 2

   Test across regimes explicitly

   Separate results for rising-rate and falling-rate periods, for high and low inflation, and for bull and bear equity markets. A strategy that only worked in one regime will surprise you.

3. 3

   Use walk-forward rather than a single split

   With few trades, one out-of-sample period is a coin flip. Rolling walk-forward analysis uses the data more efficiently.

4. 4

   Model carry explicitly

   Include dividends, financing, and roll. For multi-month holds these can change the sign of the result.

5. 5

   Accept that live confirmation takes years

   Plan monitoring around rule adherence and whether results fall within the backtest distribution, not around whether the current year is profitable.