The 21 DTE Rule Explained: When and Why to Close

The 21 DTE rule says you close or roll a short option once it has 21 days left to expiration, whatever its profit or loss. The reasoning: you keep most of the time decay and skip the stretch where an at-the-money option reacts hardest to moves in the stock. In our own simulated backtest of one wheel recipe on 40 tickers, managing at 21 DTE lowered total return on all 40, while it cut drawdowns on most of them. It cut assignments on all of them by construction: a put closed at 21 DTE can't be assigned.

The test covers one wheel recipe, and a large share of its option prices are modelled.

I built the 21-Day Rule research on this site to test exactly this. The methodology and new runs are published on the 21-Day Rule research page. This article explains the rule, why people use it, and what our numbers say.

What the 21 DTE rule is

DTE is days to expiration. Open a short put with 45 days left and it reaches 21 DTE after 24 days. At that point the rule says act: take the profit, cut the loss, or roll to a later expiration.

The rule is meant for short options: cash-secured puts, covered calls, credit spreads, iron condors and naked options. It applies whether the trade is winning or losing, and the decision is made in advance, so you're not arguing with yourself in expiration week.

Why traders use it

The rule is widely associated with tastytrade, which popularised managing short options at 21 DTE. The reasoning behind it holds up on its own.

Gamma rises near expiry. Gamma is how fast an option's delta changes when the stock moves. For an at-the-money option, with IV and the stock price held fixed, gamma scales with 1 / √(time left). That gives a simple way to see where the acceleration happens:

Gamma at 21 DTE vs 35 DTE ≈ √(35 / 21) ≈ 1.29×
Gamma at 7 DTE  vs 35 DTE ≈ √(35 / 7)  ≈ 2.24×
Gamma at 1 DTE  vs 35 DTE ≈ √(35 / 1)  ≈ 5.9×

At 21 DTE, gamma is only a bit higher than at entry. The steep part comes in the last week or two. Out-of-the-money options behave differently: their gamma stays low until the stock gets close to the strike, then jumps. Either way, a short option near expiry can go from comfortable to tested in a day or two. More on this in gamma risk near expiration and the Greeks cheat sheet.

It frees capital. Once most of the premium is in, the capital tied up in the position earns less per day than a fresh trade would. Closing early lets you redeploy it.

It avoids assignment. A put closed before expiry can't be assigned. If you don't want the shares, that alone is a reason to use the rule.

One date, no judgment call. A date-based exit is easy to follow and easy to check afterwards.

All four reasons are about risk and convenience. Whether the rule also costs return is something a backtest can check, so I ran one.

What our backtest shows

Our research runs one simulated wheel recipe two ways on the same tickers, dates and inputs:

• Hold to expiry: sell a 35 DTE put 2% out of the money. If assigned, sell 35 DTE calls 2% out of the money until the shares are called away.
• Manage at 21 DTE: the same entries, but each short put is bought back at 21 DTE and a new one is opened.

The run below covers 20 September 2024 to 18 September 2026. 45 tickers ran, and 40 passed the publication thresholds (enough cycles, history and data quality). Every one of the 40 used some modelled pricing, and a large share of option prices were modelled (a Black-Scholes fallback where historical option prices were missing). Treat all of it as a simulated backtest, not live fills.

21 DTE rule vs hold to expiry (40 published tickers, run as of 20 Sep 2026)	Result
Tickers where the rule had a higher total return	0
Tickers where the rule had a lower total return	40
Median change in total return	−20.7 pp
Median change in max drawdown	−8.2 pp
Median change in assignment rate	−32 pp

"pp" is percentage points: rule-arm value minus hold-arm value. Drawdown fell on 38 of the 40 tickers. On XLE and XLU it rose slightly. Assignment rate fell on all 40 by construction: a put closed at 21 DTE can't be assigned.

Some individual tickers show the range (all modelled):

Ticker	Hold: total return	Rule: total return	Hold: max drawdown	Rule: max drawdown
NVDA	94.2%	23.5%	38.0%	32.4%
QQQ	21.9%	15.8%	26.1%	12.3%
SPY	19.1%	10.6%	14.7%	10.0%
XLE	−11.4%	−46.0%	47.9%	50.9%

NVDA had the biggest gap in the run: the rule gave up 70.7 points of return for 5.6 points less drawdown. SPY and QQQ had smaller return gaps (8.5 and 6.1 points) and lower drawdowns with the rule.

Both measures come from the engine. Total return counts realized P&L only. Max drawdown is the largest drop in realized P&L, as a percentage of 100 × the ticker's starting share price, with open options not marked to market. Stocks that rose a lot over the period look worse on this drawdown measure.

Why the rule arm returned less

Here's what the engine does. A put bought back at 21 DTE is never assigned. So the rule arm recorded zero assignments and zero call-aways on all 40 tickers. It never owned the shares and never sold a covered call. It also traded far more often: on SPY, 48 cycles against 16 for holding. Each early close buys back a put that still has three weeks of time value in it.

The hold arm did get assigned, held the stock and sold calls against it. 32 of the 40 tickers rose on buy-and-hold over the period, so owning the shares after assignment paid. My read is that this explains a good part of the gap. The research doesn't isolate it, though. The rule also lost on all 8 tickers that fell over the period, with a median gap of −17.4 points against −23.6 for the 32 that rose, so owning shares doesn't explain the whole gap. XLE had the largest gap of the 8: it fell 27.5% on buy-and-hold, and the rule arm still did worse than holding (−46.0% vs −11.4%).

What this result doesn't tell you

• It's one recipe. 35 DTE entries, 2% out of the money, puts and calls. A 45 DTE entry, delta-based strikes or a 50% profit target could come out differently.
• It's a wheel, not a spread. Credit spreads and iron condors have capped losses and no assignment into shares. This backtest says nothing about the rule on those. For spreads, see best DTE for credit spreads and iron condor DTE optimization.
• It's one two-year window, and most of these stocks rose in it. A long falling market could change the picture.
• It's modelled. A large share of option prices were modelled rather than taken from historical quotes, and every ticker had some.

Later runs can differ from the dated numbers here.

So should you close at 21 DTE?

Decide what you want the rule to do.

If your priority is return on a wheel like this one, holding to expiry did better in our backtest, on every ticker we published. If you're happy to take the shares on assignment, the rule removes the share-owning half of the wheel, which likely did much of the work in this run.

If your priority is smaller drawdowns and fewer assignments, the rule delivered that. Against holding the full wheel, in this run, the median cost was 20.7 percentage points of total return over two years, and on some tickers it was much more.

Either choice is defensible. What isn't supported, at least by our data, is the claim that the 21 DTE rule improves returns across the board.

21 DTE vs a 50% profit target

Many sellers pair the 21 DTE rule with a profit target, most often closing once half the maximum profit is in. The two rules interact like this:

• You hit the profit target before 21 DTE. Close and redeploy. The 21 DTE rule never comes into play.
• You reach 21 DTE below the target. The date rule takes over and you close or roll anyway, even with less profit than you wanted.
• You're losing at 21 DTE. Close or roll before gamma picks up, rather than hoping for a recovery in the last weeks.

Our research tested the date rule alone. It didn't test a profit target, so I can't tell you which combination does better.

When holding past 21 DTE can make sense

• You want the shares. On a wheel, assignment is part of the plan. Our backtest suggests that's where much of the hold arm's return came from.
• The short strike is far out of the money. Gamma on a far out-of-the-money option stays low until the stock gets close. Some traders let these run and close only if the stock approaches the strike.
• The risk is defined. On a spread the maximum loss is fixed at entry. Gamma still moves the position, but the worst case doesn't grow.

Holding past 21 DTE on a tested, undefined-risk position is the case the rule is built for. That's where closing or rolling earns its keep. The roll vs close framework covers that decision.

How to apply it

Put the rule on a schedule rather than in your head. Check open short options once a week and group them by DTE, so anything approaching 21 days shows up before expiration week.

Decide in advance what 21 DTE means for each position type: close and reopen at your usual entry DTE, roll to the next cycle at the same strike, or roll out and adjust the strike if the stock has moved.

Track your own results as well. Note which positions you closed at 21 DTE and which you held, so after a few months you can compare the two on your own trades. The options portfolio tracker keeps that history in one place.

To run this on your own holdings, open your portfolio view, or look at the demo portfolio first.

For a single position, the calculator below shows how assignment risk changes with the stock price, strike and days left:

A few mistakes to avoid:

• Rolling forever. Rolling a tested position at 21 DTE again and again can hide a loss rather than manage it. If the same trade keeps getting tested, the strike or position size is probably the problem.
• Applying it to long options. The rule is for short premium. A long option buyer wants the gamma the rule tries to avoid.
• Assuming it's free. Closing early means buying back time value you would otherwise have kept. In our backtest that cost showed up in every ticker's total return.
• Ignoring trading costs. The rule has you close and reopen more often, so you pay more commissions and cross the bid-ask spread more times.

The bottom line

The 21 DTE rule trades return for risk. In our simulated wheel backtest it cut drawdowns and removed assignments, and it lowered total return on every published ticker. Use it if smaller swings and no assignment matter more to you than the last three weeks of premium and the upside of owning the shares. The methodology and new runs are published on the 21-Day Rule research page.

Related Articles

• Gamma Risk Near Expiration: What Every Options Seller Must Know
• When to Roll Options vs Close: A Decision Framework for Tested Positions
• Best DTE for Credit Spreads: A Data-Driven Comparison of 30, 45, and 60-Day Trades
• Iron Condor DTE Optimization
• Wheel Options Trading Strategy Explained
• Theta Decay in Options: DTE Curves, Strategies & Time Value Optimization
• Pin Risk in Options: Managing Expiration Uncertainty by DTE Phase
• Vanna Options Trading Explained
• Options Greeks Cheat Sheet