How premium and implied volatility behave as an expiry date approaches — the most mechanically grounded family, since time decay is a structural feature of an option contract, not a claimed behavioural edge.
7 ideas registered in this family: 4 confirmed, 2 rejected, 1 excluded.
Each bar is that one pattern's own measured change, labelled with what was measured — never a portfolio return — so only compare bars that share the same label; green rose, red fell.
decay final three days Confirmed
What was actually tested (written down before the result was seen)An ATM straddle loses value as expiry approaches. Measured over the dte=3-to-dte=1 window specifically (a single 2-session step, despite the name): the final session into expiry (dte=1 to dte=0) and expiry day itself are unmeasurable UNDER A CALENDAR-DAY CLOCK, where dte=0 gives ttm=0 and no ATM quote solves (null at dte=0, 527 of 527 rows). Under a trading-day clock 378 of those rows do solve; coverage there is 71.7% and correlates with larger, more liquid cycles, so the decay measured is conditional on the contract still quoting at settlement.
This repeats, and it kept repeating on data it was not found on.
Happened 361 times in 16 years — about 23 times a year.
It worked 166 times out of 166.
It held up on the years used to find it, and then held up again on later years it had never been tested against.
prior
lower
family
expiry
n (in-sample)
195
n (out-of-sample)
166
effect IS
-41.7%
effect OOS
-41.4%
baseline IS
0.0%
baseline OOS
0.0%
went predicted direction (IS)
99.5%
went predicted direction (OOS)
100.0%
p (IS)
<0.001
p (OOS)
<0.001
q-value (BH)
q = <0.001
chain coverage
58.2%
95% CI (OOS, excess over baseline)
[-42.7%, -40.2%]
regime
n
effect
matched
R1
48
-42.2%
100.0%
R2
308
-41.5%
99.7%
R3
6
-39.2%
100.0%
verdict reason: significant in and out of sample; awaiting FDR
decay mid cycle Confirmed
What was actually tested (written down before the result was seen)Mid-cycle (10-20 days to expiry) an ATM straddle decays more slowly than near expiry.
This repeats, and it kept repeating on data it was not found on.
Happened 2,153 times in 16 years — about 138 times a year.
It worked 1,153 times out of 1,242.
It held up on the years used to find it, and then held up again on later years it had never been tested against.
Only 47.7% of the option contracts involved had a usable, screened price on the days this happened, so this rests on a thin slice of the market.
prior
lower
family
expiry
n (in-sample)
911
n (out-of-sample)
1,242
effect IS
-11.8%
effect OOS
-10.5%
baseline IS
-0.0%
baseline OOS
-0.0%
went predicted direction (IS)
92.8%
went predicted direction (OOS)
92.8%
p (IS)
<0.001
p (OOS)
<0.001
q-value (BH)
q = <0.001
chain coverage
47.7%
95% CI (OOS, excess over baseline)
[-10.9%, -10.0%]
regime
n
effect
matched
R1
428
-11.9%
93.2%
R2
1,573
-10.8%
92.8%
R3
154
-10.8%
91.6%
verdict reason: significant in and out of sample; awaiting FDR
monthly final two days Confirmed
What was actually tested (written down before the result was seen)Monthly expiries also decay in their final measurable session -- the same dte=2-to-dte=1 window, for the same reason (no ATM quote is computed on expiry day; atm_straddle is null at dte=0, 527 of 527 rows) -- and the comparison with weeklies is meaningful only if both are measured the same way.
This repeats, and it kept repeating on data it was not found on.
Happened 149 times in 16 years — about 10 times a year.
It worked 55 times out of 55.
It held up on the years used to find it, and then held up again on later years it had never been tested against.
prior
lower
family
expiry
n (in-sample)
94
n (out-of-sample)
55
effect IS
-27.6%
effect OOS
-27.1%
baseline IS
-0.0%
baseline OOS
-0.0%
went predicted direction (IS)
100.0%
went predicted direction (OOS)
100.0%
p (IS)
<0.001
p (OOS)
<0.001
q-value (BH)
q = <0.001
chain coverage
57.2%
95% CI (OOS, excess over baseline)
[-29.0%, -25.1%]
regime
n
effect
matched
R1
52
-27.9%
100.0%
R2
90
-27.0%
100.0%
R3
7
-27.8%
100.0%
verdict reason: significant in and out of sample; awaiting FDR
weekly final two days Confirmed
What was actually tested (written down before the result was seen)Weekly expiries decay in their final measurable session. Measured over the dte=2-to-dte=1 window specifically (a single 1-session step, not two, despite the name): dte=1-to-expiry and expiry day itself are unmeasurable UNDER A CALENDAR-DAY CLOCK (ttm=0 there, so nothing solves); a trading-day clock does measure them. The comparison with monthlies is meaningful only if both are measured the same way.
This repeats, and it kept repeating on data it was not found on.
Happened 302 times in 16 years — about 19 times a year.
It worked 174 times out of 176.
It held up on the years used to find it, and then held up again on later years it had never been tested against.
Almost all of the evidence for this (300 of 303 times) comes from the era when weekly BankNifty expiries existed. Weekly expiries were abolished, so this cannot be used today.
prior
lower
family
expiry
n (in-sample)
126
n (out-of-sample)
176
effect IS
-30.7%
effect OOS
-28.3%
baseline IS
-0.0%
baseline OOS
-0.0%
went predicted direction (IS)
98.4%
went predicted direction (OOS)
98.9%
p (IS)
<0.001
p (OOS)
<0.001
q-value (BH)
q = <0.001
chain coverage
59.3%
95% CI (OOS, excess over baseline)
[-29.6%, -27.0%]
regime
n
effect
matched
R1
1
-30.1%
100.0%
R2
300
-29.3%
98.7%
R3
2
-23.8%
100.0%
verdict reason: significant in and out of sample; awaiting FDR
forward drifts toward max pain Rejected
What was actually tested (written down before the result was seen)When the forward sits more than 1% above max pain in the last week of a cycle, it drifts back down toward it.
This does not repeat.
Happened 155 times in 16 years — about 10 times a year.
It went the expected way 22 times out of 53.
There is no evidence this repeats. It did not pass: not significant in sample (p=0.500)
The range of plausible values includes zero, which means the effect could be real or could be nothing at all.
prior
down
family
expiry
n (in-sample)
102
n (out-of-sample)
53
effect IS
0.2%
effect OOS
0.5%
baseline IS
0.0%
baseline OOS
0.2%
went predicted direction (IS)
48.0%
went predicted direction (OOS)
41.5%
p (IS)
0.4997
p (OOS)
0.3987
q-value (BH)
q = 0.3987
chain coverage
53.2%
95% CI (OOS, excess over baseline)
[-0.8%, 1.3%]
regime
n
effect
matched
R1
68
0.1%
51.5%
R2
76
0.5%
39.5%
R3
11
0.2%
54.5%
verdict reason: not significant in sample (p=0.500)
forward drifts up to max pain Rejected
What was actually tested (written down before the result was seen)The mirror case: a forward more than 1% below max pain in the last week drifts back up toward it.
This does not repeat.
Happened 128 times in 16 years — about 8 times a year.
It went the expected way 33 times out of 53.
There is no evidence this repeats. It did not pass: not significant in sample (p=0.083)
The range of plausible values includes zero, which means the effect could be real or could be nothing at all.
prior
up
family
expiry
n (in-sample)
75
n (out-of-sample)
53
effect IS
-0.3%
effect OOS
0.6%
baseline IS
0.1%
baseline OOS
0.2%
went predicted direction (IS)
42.7%
went predicted direction (OOS)
62.3%
p (IS)
0.0833
p (OOS)
0.2901
q-value (BH)
q = 0.3481
chain coverage
54.9%
95% CI (OOS, excess over baseline)
[-1.1%, 1.8%]
regime
n
effect
matched
R1
50
-1.1%
34.0%
R2
67
0.8%
59.7%
R3
11
0.7%
72.7%
verdict reason: not significant in sample (p=0.083)
iv rises into expiry week Excluded
What was actually tested (written down before the result was seen)Implied volatility rises as expiry approaches, because a shrinking time base inflates the volatility implied by a given premium.
This was not tested.
Not tested. This is not a rejection — the statistics needed to test it honestly are not available, so it was never run against the data. Reason on record: measures implied volatility, which is not directly observed but computed from an option's price using an assumed clock; this analysis counts calendar days while the market trades on business days (dte=7 is 4.76 trading sessions, dte=3 is 2.88); a zero-structure simulation containing no market behaviour at all -- constant true volatility, exact Black-76 prices on a random walk, no weekday or dte effect planted -- reproduces this pattern's published effect almost exactly (simulated +14.16% vs published +14.14%) using only that calendar-day clock; and the effect weakens to insignificance (+14.14% to +0.23%, p 0.0001 to 0.58) and reverses sign under a trading-day clock, on the real data. Not tested -- not rejected on evidence.
No hit rate, frequency or range is given for this row because none was ever computed — treat every number here as absent, never as zero.
prior (as registered)
higher
family
expiry
No effect size, hit rate, p-value or confidence interval
is shown here because none was ever computed — this hypothesis was not
run, not run-and-rejected. Every number that would normally sit in this
box is absent, never zero.