Calendar Lab | optionsdatabase.com

Three new calendar-spread strategies, honestly costed

These are three NEW calendar strategies, each backtested against this database's own option chains -- not the existing Big Movers or Fridays screeners. Every P/L below is quoted at a realistic fill (mid plus 35% of the quoted spread, both ways), with both the mid-price fantasy and the full-spread worst case shown beside it. The first version of strategy 1 looked like +$116k at the mid and was −$211k paying the real spread -- the entire "edge" was inside the bid/ask. That is why every table here carries a Worst Case column.

1. Mean Reversion

CONDITIONAL

Idea: a violent move pumps front-week IV, which SHRINKS the calendar debit. Buy it while it sits in the cheapest 15% of its own history, sell when it returns to normal.

Result: 377 trades, 61.5% wins, +$56,196 total, median +$107. Survives down to a fill quality of about 0.75 of the quoted spread, then dies.

Verdict: a real but execution-sensitive edge. It needs decent fills; it is not a market-order strategy.

2. Expiry-Day Theta

STRONGEST

Idea: on expiry day the 0DTE leg’s remaining premium is burned by the clock -- a mechanical certainty -- while the next expiry holds its value. Sell the 0DTE, buy the later expiry, same strike, a couple of strikes OTM.

Result: 415 trades on SPY and QQQ, 58.3% wins, +$38,706 total, median +$96, worst single trade −$726. Still +$24,289 paying the full quoted spread on every leg, and it holds up out-of-sample.

Verdict: the only one of the three that survives pessimistic execution — because the short 0DTE leg usually expires worthless, so there is no closing spread to pay on it. See the dedicated page.

3. Earnings IV Ramp

NOT TRADABLE YET

Idea: buy the calendar whose BACK expiry contains earnings and whose front does not; the back leg inflates into the print. Exit before the announcement, never through it.

Result: the mechanism is confirmed -- back-leg IV rises further the earlier you enter (+7 vol points at 5 days, +17 at 21 days). But P/L is only +$6,896 over 78 trades and goes negative at worst-case fills in every configuration tested.

Verdict: your observation is right, the calendar just is not the way to monetise it. Better as a filter than a trade.

Headline — Realistic Fills

Theta P/L

+$38,706

Theta Win %

58.3%

Theta Worst Case

+$24,289

Mean-Rev P/L

+$56,196

Mean-Rev Win %

61.5%

Earnings P/L

+$6,896

Execution Sensitivity

The single most important table here. fill_frac 0.00 means every fill lands at the mid-price, which never happens. 1.00 means you pay the entire quoted spread on both legs, both ways -- nobody market-orders a calendar, so this is the pessimistic bound. Where a strategy crosses zero tells you how good your execution has to be.

Expiry-Day Theta

Variant	Trades	Win %	Total P/L	Avg / Trade	Median	Avg Return	At Mid	Worst Case
fill_frac=0.00	415	60.0%	+$47,277	+$114	+$108	+12.3%	+$47,277	+$24,712
fill_frac=0.20	415	59.3%	+$42,376	+$102	+$104	+11.1%	+$46,995	+$24,482
fill_frac=0.35	415	58.3%	+$38,706	+$93	+$96	+10.2%	+$46,782	+$24,289
fill_frac=0.50	415	57.1%	+$35,304	+$85	+$89	+9.3%	+$46,836	+$24,353
fill_frac=0.65	415	56.9%	+$31,925	+$77	+$80	+8.3%	+$46,907	+$24,439
fill_frac=0.80	415	56.1%	+$28,518	+$69	+$76	+7.5%	+$46,838	+$24,521
fill_frac=1.00	415	55.9%	+$23,830	+$57	+$72	+6.2%	+$46,538	+$24,407

Mean Reversion

Variant	Trades	Win %	Total P/L	Avg / Trade	Median	Avg Return	At Mid	Worst Case
fill_frac=0.00	377	70.6%	+$115,681	+$307	+$224	+32.3%	+$115,681	−$38,323
fill_frac=0.20	377	66.3%	+$79,309	+$210	+$163	+22.4%	+$108,358	−$38,460
fill_frac=0.35	377	61.5%	+$56,196	+$149	+$107	+15.7%	+$104,830	−$35,656
fill_frac=0.50	377	56.8%	+$33,520	+$89	+$62	+9.5%	+$100,370	−$34,805
fill_frac=0.65	377	52.8%	+$12,799	+$34	+$18	+3.7%	+$97,256	−$34,143
fill_frac=0.80	377	46.7%	−$7,454	−$20	−$17	-1.8%	+$92,815	−$33,972
fill_frac=1.00	377	40.3%	−$30,286	−$80	−$68	-8.5%	+$89,702	−$31,660

Backtested on this database's own chains: 175 liquid tickers of nightly EOD ladders and 5-minute expiry-day chains, Nov 2025 – Aug 2026. $1,000 per position, whole contracts, no compounding, no commissions. EOD/5-min marks are not fills -- these are research results, not a track record. Scripts live in calendar_lab/; method notes in educational/CALENDAR_LAB.txt.