# Managing Options Trades: 50% Profit & 21 DTE Tested
Source: https://theoptionsbench.com/managing-options-trades-backtest/

> Plain-text mirror for AI/LLM ingestion. Canonical HTML page: https://theoptionsbench.com/managing-options-trades-backtest/

---

Updated 3 September 2026 · by Theo Chen

"Take profits at 50%. Close at 21 days to expiration." It is the most-repeated advice in options
selling, and almost nobody has tested it at real prices. So we did - on 163-164
monthly cycles per strategy from real SPY option chains across **5 strategies**, entering at
~45 days to expiration exactly as the rule intends, and marking every position
**every single trading day** to find the moment it could actually be closed.

No managed rule beat simply holding to expiration by a statistically distinguishable margin. But the useful
findings are sharper than that, and two of them overturn common belief:

- The 50% profit target is not crash insurance. Its worst month matched holding's in 4 of 5 strategies, but it cut the Cash-Secured Put's from -$8,716 to -$3,860 - a 55.7% reduction - after closing a winner before the crash. A target can only close a winner; it cannot protect a loser that never reaches the target.
- Whether a time exit helps depends on your strategy. For undefined-risk trades a 21-DTE exit is the only in-trade tail cap. For defined-risk spreads the long wing already capped the worst month, so managing mostly just gives back return.
- The time exit's worst-month reduction depends on March 2020. Strip that cycle and both undefined-risk strategies trail holding: $8,755 behind on the Short Strangle (t = -1.93) and $16,176 behind on the Cash-Secured Put (t = -2.49). Only the Cash-Secured Put clears the two-sided threshold.
- Managing cost return in ordinary months for most scorecard strategies here. The Iron Condor's 25% and 50% targets were the exception: +$1,711 and +$1,948 versus holding outside March 2020, but neither difference was statistically distinguishable (t = 0.47 and 0.88). Every close still pays a spread holding never does - at the ask, managing the four-legged Iron Condor flips from +$391 to -$534.

## Why the 50% profit target is not crash insurance

This is the finding to sit with, because the 50% profit target is the single most-repeated management
rule in retail options selling. A target can _only_ close a position that is already up 50% - so a
trade that never trades up 50% is never touched, and rides to expiration exactly as if you had no rule at
all. Its crash loss is untouched **by construction.** The timing exception is a different trade
closing as a winner before the crash arrives. Here are the scorecard strategies; the Covered Call leg is discussed below:

| Strategy | Worst month, holding | Worst month, 50% target | Difference |
| Short Strangle | -$7,602 | -$7,602 | identical |
| Bull Put Spread | -$2,582 | -$2,582 | identical |
| Iron Condor | -$2,228 | -$2,228 | identical |
| Cash-Secured Put | -$8,716 | -$3,860 | +$4,856 |

4 of 5 are identical to the dollar. The exception is the Cash-Secured Put:
its worst month improved from -$8,716 to -$3,860, a
55.7% reduction. Its short Put happened to be winning on the market's rally _into_
February 2020, so the target closed it before the crash arrived - pure sequence luck. Had that crash begun
a week after entry, the Put would have ridden it down too. That timing effect was not confined to one event:
among holding losses of at least $500, the target beat holding by more than $100 in
5 Cash-Secured Put cycles and 3 Short Strangle cycles.
It closed a winner before a later drawdown; it did not protect a losing trade. **If
you "manage at 50%," you have bought income smoothing, not crash insurance.**

## Does the answer depend on your strategy? Yes - on whether it caps its own tail

Here is the decision most traders actually face: _I already trade strategy X - should I manage it or let
it expire?_ The answer splits cleanly on one feature: whether the strategy has a long wing that already
caps its worst month. Same 163-164 cycles per strategy, dollars per contract, held to
expiry versus the full tastytrade rule:

| Strategy | Risk | Let expire | Take 50% | Close 21 DTE | Worst month: hold -> 21 DTE | Does managing help? |
| Cash-Secured Put30-delta Put | Undefined | $19,522 | $17,011 | $9,719 | -$8,716 -> -$2,347 | Time exit caps the crash tail - at half the return |
| Short Strangle16-delta Put + Call | Undefined | $6,110 | $4,269 | $3,334 | -$7,602 -> -$1,623 | Time exit caps the crash tail - costs return |
| Bull Put Spread30/12-delta | Defined | $9,200 | $7,065 | $3,494 | -$2,582 -> -$1,224 | No proven return edge - 52.6% tail reduction |
| Iron Condor16-delta / 7-delta wings | Defined | $1,148 | $3,095 | $391 | -$2,228 -> -$613 | No proven return edge - 72.5% tail reduction |

Read the last two columns together. For the **undefined-risk** trades (Short Strangle,
Cash-Secured Put), holding exposes a brutal crash month (-$7,602, -$8,716)
and the 21-DTE exit is the only thing that caps it. For the **defined-risk** spreads the long
wing already did that job. On the Iron Condor, the 50% target left the worst month at
-$2,228, while the 21-DTE exit trimmed it from -$2,228
to -$613. The time exit still barely changes the total (dte21 vs hold
-$757, t = -0.22, statistically nothing). **The
wing bought the essential insurance; further management bought no proven return edge.**

One caveat kept honest: 6 rules across 2 strategies
finished above holding on total, but none was a statistically distinguishable return improvement. The Iron
Condor made $1,148 holding and up to $3,095 managed. The
Covered Call is different: holding lost $13,858, and managing only
lost less, with a best managed total of -$4,165. The strongest result stopped at
|t| = 1.88, and these rules posted higher monthly P/L only
17.2%-38.4% of the time. A few large months carried
the totals. That is noise, not edge.

## How we know: the Short Strangle, cycle by cycle

The scorecard above is a summary; this is the proof, on the strategy where managing looked most valuable.
Entering 2020-02-04, 45 days before the March-2020 expiry, we sold the 308 Put and
342 Call for a $303 credit with SPY near $333. SPY first
_rallied_ toward 338 - bleeding the short Call while the Put decayed, so the position never printed
+50% and the profit target never fired. It was still open when the crash came, and expired at SPY
$229 for -$7,602. The 21-DTE rule, a pure calendar, fired on 2020-02-28
at SPY $296 and closed for -$1,623. That difference explains the 21-DTE rule's
worst-month improvement on this trade; it is not a general profit-target finding:

| Close at 21 DTE vs holding | Short Strangle | Cash-Secured Put |
| March-2020 cycle alone | +$5,979 | +$6,373 |
| The other 163 cycles | -$8,755 | -$16,176 |
| All 164 cycles | -$2,776 | -$9,803 |

Outside that one month, managing is not a wash - it is a steady drag (-$8,755 and
-$16,176), and on the Cash-Secured Put that drag is statistically real
(t = -2.49). The full-sample "no measurable difference" is the crash month paying for
163 months of premium. That is exactly what insurance looks like on a P&L - and
exactly why you cannot call it free. Holding's second-worst month, for scale, was only -$2,396;
one catastrophe, then a long tail of small ones.

## The 2020 crash, counted honestly across strategies

A single worst-month figure flatters the time exit. Count all three cycles that spanned the crash instead,
and the defined-vs-undefined split gets even clearer:

| Cycles expiring Feb-Apr 2020 | Hold | Close 21 DTE |
| Short Strangle (undefined) | -$6,772 | -$2,814 |
| Cash-Secured Put (undefined) | -$7,612 | -$4,682 |
| Bull Put Spread (defined) | -$766 | -$2,135 |
| Iron Condor (defined) | -$826 | -$1,070 |

The undefined-risk trades bled through the crash, and the time exit cut it. But the defined-risk spreads
_barely noticed 2020_ - the wing capped them at -$766 and -$826 - and managing
them through the crash actually made it **worse**, closing capped positions at bad prices. For
a spread, there was no tail to insure, so the insurance only cost.

## What this means for how you trade

- Do not expect a 50% profit target to protect you. It cannot touch a trade that never went your way - which is precisely the trade that hurts. Use it to free capital and book winners early if you like, but not as risk management.
- If you trade defined-risk spreads, the manage-vs-expire question is low-stakes. The wing already capped your worst month; a time exit trims a small tail further and gives back return. Fine to manage for capital efficiency, but it is not buying you safety you do not already have.
- If you trade undefined risk, a time exit is your only in-trade tail cap - and it trailed holding in 120 of 163 Short Strangle cycles and 145 of 163 Cash-Secured Put cycles, for a worst-month reduction that depended on March 2020 here. Worth it only if a single uncapped -$7,602 to -$8,716 month is a risk to your account, not just your year. A sum of P/L cannot show path-to-ruin.
- Count the friction. Every close crosses a spread holding never pays; four-legged trades pay it four times. Model your own trade in the short strangle, cash-secured put or iron condor calculator.

**A note on covered calls.** We tested the short-Call leg on its own (the study holds no
stock). Held to expiry it lost $13,858 over 14 years -
but do _not_ read that as "covered calls lose money." A short Call is structurally short the trend,
and this sample was mostly a rising market, so that figure is **forgone upside on the option leg,
not an account loss** - a real covered call also holds the stock, which rose several-fold. Its worst
month was a 2023 _rally_, not a crash; a real covered call's tail is the crash, via the stock, which
managing the Call does not touch. That is why it is not in the scorecard: the manage-vs-expire question for
a covered call is really a question about the short Call alone, and over a mostly rising sample that
leg simply gave back upside.

## Caveats - read these

- The time-exit tail result depends on March 2020. The 21-DTE worst-month reductions on the undefined-risk trades depend on that crash's timing - the exit fired near the pre-crash peak, weeks before the bottom. A crash gapping through the 21-DTE mark changes the number completely. Treat direction as suggestive, magnitude as unknown.
- Mid and ask are a bracket. Holding pays no exit spread; every managed rule pays to close, and four-legged trades pay it four times - at the ask the Iron Condor's 21-DTE exit flips from +$391 to -$534. Reality sits between.
- Freed capital sits idle. After an early exit we do not re-deploy until the next monthly entry, which understates management's return - the one bias in holding's favour.
- Held to expiration means European-style intrinsic settlement, no commissions, no early assignment; a few cycles were dropped where the chain data had gaps. Undefined worst months are the 2020 crash; defined worst months are a 2022 grind.
- One window, one underlying, monthly SPY, at fixed deltas (16- and 30-delta shorts, 7- and 12-delta wings). Past performance is not predictive. Educational, not advice.

Method: real OptionsDX end-of-day SPY chains, 2010-01-05 to 2023-12-15. One position per
monthly expiry, entered on the trading day nearest 45 DTE, strikes by each option's real delta,
net credit at the bid/ask mid. Every position is marked on every trading day to locate the first day it
could be closed at the target or at 21 DTE; closes are priced at both the mid and the ask, and settled at
intrinsic (defined-risk losses capped by the long wing). All rules trade identical entries, so each is
compared to holding by the **paired** difference. Numbers were independently re-derived from
the raw chains by a separate script and by five per-strategy verification agents - the Short Strangle and
Cash-Secured Put reproduce this site's two-strategy management study to the dollar, and the four-legged Iron
Condor's worst cycle reproduces to the cent - and the findings were adversarially reviewed, which is how the
profit-target finding - timing, not crash protection - and the Covered Call framing were pinned down. Every figure
regenerates from the data.

Cite this study

Theo Chen. Managing Options Trades: 50% Profit Target vs 21 DTE. The Options Bench, 2026.
https://theoptionsbench.com/managing-options-trades-backtest/. Licensed under CC BY 4.0.

Theo Chen. Managing Options Trades Across Five Strategies. The Options Bench, 2026.
https://theoptionsbench.com/managing-options-trades-backtest/. Licensed under CC BY 4.0.

Download the underlying data: JSON CSV (two-strategy study) JSON CSV (five-strategy study)

The bottom line

Across 5 strategies with 163-164 monthly SPY cycles each, no managed rule beat simply holding to expiration by a statistically distinguishable margin. Two findings matter most. First, the famous 50% profit target is not crash insurance: its worst month matched holding's in 4 of 5 strategies, while the Cash-Secured Put improved from -$8,716 to -$3,860 after the target closed a winner before the crash. That exception is timing, not protection for a losing trade. Second, whether a 21-DTE time exit helps depends on your strategy - it is the only tail cap for undefined-risk trades (Short Strangle, Cash-Secured Put), but for defined-risk spreads the long wing already capped the worst month, so managing added no distinguishable return edge. The time exit's worst-month reduction depends on March 2020. Manage for a tail you cannot otherwise survive, never for return.

## Frequently asked questions
Does taking profits at 50% protect you in a crash?

Not reliably - and not as a mechanism. A profit target only closes a position that is already winning, so it cannot rescue a crash-bound trade that never reaches the target. In this sample, the 50% target's worst month matched holding to the dollar in 4 of 5 strategies. The Cash-Secured Put was the exception: its worst month improved from -$8,716 to -$3,860, a 55.7% reduction, because the target closed a winner before the later crash. That is sequence luck, not crash insurance.

Is it better to manage options trades or let them expire?

Managing is insurance, not edge. No managed rule beat holding by a statistically distinguishable margin in our test. If you trade a defined-risk spread, the long wing already caps your worst month, so managing is low-stakes and adds no proven return edge. On the undefined-risk Short Strangle and Cash-Secured Put, a 21-DTE time exit is your only in-trade tail cap, but its worst-month reduction in this sample depends on March 2020. Manage for a tail you cannot survive, not for return.

Does closing at 21 DTE reduce risk?

For an unhedged downside trade, yes - but the worst-month reduction here depends on the March-2020 exit, and it costs return. Closing the Short Strangle at 21 DTE cut the worst month from -$7,602 to -$1,623, but it beat holding in only 26.8% of months, and strip the March-2020 cycle and it trails holding (ex-March t = -1.93). For a defined-risk Iron Condor the wing already caps the tail. The 21-DTE exit materially improved this sample's worst month from -$2,228 to -$613, but total return fell by $757 versus holding (t = -0.22) - and at realistic fills it turns negative.

Is managing options trades free?

No. Over the full sample the cost looks statistically indistinguishable from zero, but that verdict rests on one cycle: March 2020 alone contributes +$5,979 to the 21-DTE rule's Short Strangle result and +$6,373 to the Cash-Secured Put's. Remove that month and only the Cash-Secured Put's drag clears the two-sided threshold (CSP t = -2.49; Short Strangle t = -1.93). And every close pays a spread that holding to expiration does not - at the ask, managing the four-legged Iron Condor flips from +$391 to -$534.

Should I close my options trades early or let them expire?

This one 14-year SPY sample cannot decide that for you, and anyone claiming a backtest settles it is overreading one crash. What it shows: no managed rule beat holding by a statistically distinguishable margin; a profit target is not crash insurance, although its timing spared the Cash-Secured Put's worst month here; and a time exit trades away ordinary-month return for a smaller worst month on undefined-risk trades. If a single uncapped losing month could end your account, that trade may be worth it - a sum of P/L cannot show path-to-ruin risk. Educational only, not advice.

## Related questions

- What does a short strangle actually return?
- Should I roll or take assignment?
- How does an iron condor cap both tails?
- All options data studies