Satire · Explainer · 08 Sep 2026

The shorter the trousers, the fewer the offers — until, quite suddenly, the more.

The Institute for Trouser Epidemiology measured what happens to a man’s prospects as his hem climbs from the shoe to the hip. The curve falls off a cliff at the knee, and then at five inches it does something no committee expected: it climbs, and the audience it climbs with is predominantly male.

Plate IThe Plateau32 in (81 cm) · O = 72
Plate IIThe Ankle Anomaly27 in (69 cm) · O = 81
Plate IIIThe Cargo Cliff12 in (30 cm) · O = 9
Plate IVThe Gym Trough10 in (25 cm) · O = 14
Plate VThe False Recovery7 in (18 cm) · O = 26
Plate VIThe Five-Inch Singularity5 in (13 cm) · O = 88
Plate VIIThe Marathon Anomaly3 in (8 cm) · O = 71
Plate VIIIThe Terminal Brief0 in (0 cm) · O = venue-defined
Plates I–VIII. One man, one wall, one afternoon of flat overcast light; only the garment changes between plates. The Opportunity Index under each plate is the number the rest of this paper explains.
Part one · The Function Two variables, one man, one wall, and a curve worth drawing.

01Trouser length predicts opportunity better than any variable a man can change before lunch§

Of everything a man can alter about himself in a morning, the length of his trousers moves his prospects the most, and it moves them in a shape we can draw.

Abstract

Journal of Applied Hemline Studies · vol. 1, no. 1 · the only issue

Inseam length as a predictor of sexual opportunity in adult men: a longitudinal study of one afternoon

Institute for Trouser Epidemiology. We photographed one man against one wall in eight garments and asked four men in a bar to be honest about him. Opportunity fell 72 points between the ankle and the knee, failed to recover at seven, and at five inches did something no committee predicted. The audience waiting for him there is predominantly male, which our instrument was not built to see and did not report until it was asked four times. Effective n = 3.5. Results extrapolated to 4.0 billion men by multiplying. Peer-reviewed by the corresponding author’s roommate.

Two variables, defined before anyone is measured

We call the shape the Hemline–Opportunity Function, and everyone at the Institute calls it the Function. It is the whole curve of O(L) as the inseam L runs from thirty-two inches down to zero. O is the Opportunity Index, a number from 0 to 100. It is the modelled probability, expressed as a percentage, that a stranger who already finds the wearer acceptable continues to find him acceptable after looking down. Alongside it we track Pool Breadth, B(L): how many distinct populations contain at least one person interested in the wearer. We set it to 1.0 at thirty-two inches so every later reading can be compared with it.

Four men, one wall, one afternoon

Four men supplied the data, recruited at a bar near the Institute. In other words, a convenience sample, and the convenience was ours. Respondent 3 left before question two, which brings the effective sample to 3.5, a figure we report unrounded because a smaller paper would have rounded it and hoped nobody asked. Four men is enough once a reviewer accepts what one afternoon can do to a man’s trousers.

We held everything else constant. One man, one wall, one afternoon of flat overcast light, one very patient landlord: only the garment was free to move. Concretely, that is a controlled experiment in the oldest sense, the kind a chemistry department would recognise, even if a hemline is not usually the independent variable. Three numbers carry the argument, and we put them up front rather than make a reader hunt for them in the discussion.

81→9
Plate II to Plate III: a 72-point fall

Two inches of bare ankle scored 81. Fifteen inches lower, eleven pockets and a phone scored 9.

4.8×
Pool Breadth growth, Plate V to Plate VI

At five inches the pool does not merely grow from 0.5 to 2.4; its composition changes, and what arrives is a male audience the earlier regimes never counted.

n = 3.5
the effective sample

Four men, one of whom left before the second question. His half is counted.

The Function, drawn end to end

Figure 1 draws the whole curve, and it is worth sitting with, because every section below is one stretch of the same line taken at walking pace. Read it left to right and the hem climbs: the Plateau, the Ankle Anomaly, the Cargo Cliff, the Gym Trough, the False Recovery, the five-inch jump we explain in §08, the Marathon Anomaly, the Terminal Brief. Eight regimes in thirty-two inches is a great deal of terrain for one man’s afternoon. The section numbers along the top are the map; the knees are his.

VENUE-DEFINED 100 75 50 25 0 OPPORTUNITY INDEX O THE KNEE 81 9 88 THE PLATEAU §02 THE CARGO CLIFF §04–05 NO RECOVERY §07 MARATHON §09 ANKLE ANOMALY §03 GYM TROUGH §06 SINGULARITY §08 TERMINAL §10 0 1 2 POOL BREADTH B = 2.4 32 27 14 12 10 7 5 3 0 I II III IV V VI VII VIII INSEAM, INCHES Everything between the ankle and the thigh loses; the two winning lengths are twenty-two inches apart.
Figure 1The two peaks are not neighbours, which is why a man who shortens gradually passes through every bad length on the way.

The eight plates, and the ruler they were measured with

Eight plates, shown at the top of this page, produced the curve, and we recommend reading them as a set. For example, Plate I and Plate VI look nothing alike and score 72 and 88; Plate V and Plate VI look almost identical and score 58 and 88. The finding sits in the sequence, not in any single photograph. Figure 2 is the ruler: every regime is one horizontal line on the same pair of legs.

the knee · 14 in VIII · The Terminal Brief VII · The Marathon Anomaly VI · The Five-Inch Singularity V · The False Recovery IV · The Gym Trough III · The Cargo Cliff II · The Ankle Anomaly I · The Plateau 0 in · 0 cm 3 in · 8 cm 5 in · 13 cm 7 in · 18 cm 10 in · 25 cm 12 in · 30 cm 27 in · 69 cm 32 in · 81 cm Every plate is one horizontal line on the same leg; the leg never changes.
Figure 2Every regime in the taxonomy is measured on the same pair of legs: the whole experiment is a ruler held against one man, which is both its rigour and its problem.

02Full-length trousers are the baseline, and the baseline is fine§

At thirty-two inches a man is invisible from the waist down, and invisibility scores a solid 72.

Plate I · 32 in (81 cm)The Plateaufull-length stone chino, single break at the shoe
72Opportunity Index O
95% CI
41–100
Pool Breadth B
1.0
In the pool
the general adult population
Nothing happens here, which is the point.

Plate I shows the full-length stone chino, one break at the shoe, white sneakers, thirty-two inches (81 cm) of fabric between hip and heel. O reads 72. Its 95 per cent confidence interval, the range in which we are 95 per cent sure the true score sits, runs all the way from 41 to 100. We take that width as evidence for the central claim of this section: the trouser conceals, and everything else about the man decides. Pool Breadth sits at 1.0, drawn from the general adult population, which is the widest and least interesting pool in the study.

We call this stretch the Plateau, the first of the terrain names this paper will use: flat, wide, and exactly as uneventful as the interval suggests. Between twenty and thirty-two inches a man can lose four inches of hem and the Index barely registers it. A hiker loses four hundred feet on a wide plain the same way, without feeling it underfoot. Respondent 1, asked to describe Plate I, said nothing to report, and that’s not a complaint. Kyiv agrees, a suburban barbecue agrees; the venue matrix in §10 shows this plate viable or optimal almost everywhere it appears.

A flat 72 is not a finding on its own, and we would not have published on the Plateau alone. It is the reference point, the sea level from which we can later call a cliff a cliff. Twenty inches of fabric remain between here and Plate III, and for now all of it is still safe. Nothing happens here, which is the point.

03The cropped trouser is the only length that beats the baseline, and Milan knows§

Two inches of bare ankle above a loafer lifts the score from 72 to 81, and it is the last good news for twenty inches.

Plate II · 27 in (69 cm)The Ankle Anomalycropped tailored trouser, bare ankle, no socks, suede loafers
81Opportunity Index O
95% CI
58–97
Pool Breadth B
1.3
In the pool
the general population, plus everyone who has been to Milan
The only trousered length that beats the baseline.

Plate II crops the trouser to twenty-seven inches (69 cm): a tailored cut, two inches of bare ankle, no socks, suede loafers. O climbs to 81, interval 58 to 97, and Pool Breadth widens to 1.3, adding everyone who has been to Milan to the general population already present. We call this stretch the Ankle Anomaly, because it is the only trousered length in the dataset that beats the baseline, and it beats it by nine points.

Credit for the nine points belongs to the Sock Interaction Term, a multiplier that sock height applies to O, which we formalise properly in §11 and can only flag here. Bare ankle at this exact length applies a multiplier of ×1.11, which the Institute named the Milan bonus. It is the one sock configuration a stranger reads as a decision rather than an accident. In practice, a bare ankle at twenty-seven inches lands as intentional in a way the same ankle at twenty inches never would.

We did not trust the reading at first. By then the number rested on three testimonies instead of four, one of them by phone. We re-ran the comparison the following week with a fresh loafer and got 81 twice. Twenty-seven inches is also the last stop before the terrain turns. Every plate after this one loses length, and with one exception seven inches from now, every plate after this one loses score. We ask the reader to enjoy Plate II. The Institute certainly did.

Part two · The Cliff Where seventy-two points disappear in fifteen inches of khaki, and we account for every one of them.

04At the knee the curve does not decline; it falls§

Between twenty-seven inches and twelve, the score drops 72 points, and almost all of the fall is cargo shorts.

Plate III · 12 in (30 cm)The Cargo Cliffbelow-knee khaki cargo short, eleven pockets, phone in the thigh pocket, mid-calf socks
9Opportunity Index O
95% CI
0–23
Pool Breadth B
0.2
In the pool
his mother; a golden retriever
A 72-point fall over fifteen inches of fabric.

Plate III is the below-knee khaki cargo short: twelve inches (30 cm), eleven pockets, a phone sagging in the right thigh pocket, mid-calf white socks. O collapses to 9, interval 0 to 23, and Pool Breadth falls to 0.2. Two observers remain in the pool at this length: his mother, and a golden retriever. The retriever wandered into frame during Plate III and stayed for the rest of the afternoon, apparently untroubled by any of it.

We call this stretch the Cargo Cliff, and the name is literal. Where the Ankle Anomaly was a nine-point rise over five inches, this is a 72-point fall over fifteen. Measured from the calmer baseline of Plate I it is still 63 points. Nothing else on the Function moves this far in either direction, and nothing else moves this far for three separate reasons at once.

Three mechanisms, stacked

In practice, length alone, with the pocket count held at two, takes a twelve-inch short from 81 down to 46. Nine additional pockets, still with no phone, take it from 46 to 16. Finally the phone in the thigh pocket takes the last seven, down to the 9 we record on Plate III; we call that flat seven the Sag Term. Put another way, length takes about half the score, pockets take most of the remainder, and a rectangle of glass takes the rest. We take the three apart in order in §05.

One disclosure belongs here, and the corresponding author makes it himself: I own the cargo shorts photographed for Plate III. They are not a prop borrowed for the study. They live in my wardrobe, and I am keeping them for the garden. A garment demoted from Plate III has exactly one dignified use left, and weeding is it. Our retriever, for what it is worth, showed no measurable reaction to any plate, which we take as the closest thing this paper has to a null result.

05Every pocket beyond the second removes 11.3 per cent of whatever is left§

Pockets do not subtract points; they multiply what remains by 0.887, which is why eleven of them leave almost nothing.

We define the Pocket Coefficient, β, as the fraction of remaining opportunity removed by each exterior pocket beyond the second: β = 0.113, or 11.3 per cent. Pockets multiply. Each pocket keeps 88.7 per cent of whatever score survived the pocket before it. So the third pocket takes 11.3 per cent of what the second left, the fourth takes 11.3 per cent of what the third left, and so on.

Op = O2 × (1 − 0.113)(p − 2)   where p is the pocket count and O2 the score of the same length with two pockets
Reading β without the algebra

Picture opportunity as a glass of water. Length alone pours out roughly half the glass. Each pocket beyond the second does not scoop out a fixed cupful; it takes 11.3 per cent of whatever water is still in the glass. By the ninth pocket there is very little left to take, which is the entire mechanism.

The decay, tabulated at twelve inches

Hold the length at twelve inches and vary only the pockets, and the table in Figure 3 appears. Two pockets on a plain Bermuda score 46. By eleven pockets, Plate III’s count, that has fallen to 16. By fourteen, the count at which a short is better described as a tool belt, it is down to 11. Any linear model would have predicted a fourteen-pocket short scoring close to zero, since fourteen is a large number of pockets by any measure. Ours predicts 11, because the fourth pocket cannot remove what the third already took.

The Pocket Paradox β = 0.887 per pocket · p = 0.049 (fourth attempt) no phone phone, thigh pocket predicted decay 50 40 30 20 10 0 Opportunity Index O 46 36 28 22 16 11 16 → 9 the Sag Term, −7 2 4 6 8 11 14 Plate III carpenter pocket count The fourth pocket cannot remove what the third already took, so the damage compounds instead of adding.
Figure 3A man who adds pockets one at a time never feels any single pocket cost him much, which is how eleven happen.

The Sag Term

Carry a phone in a cargo thigh pocket and it removes a further seven points flat, on top of the decay. We named it the Sag Term in §04, and it is the one effect in the paper that adds rather than multiplies, because a phone does not care what fraction of the trouser is left to sag from. At eleven pockets the Sag Term is the difference between 16 and 9, which is the entire distance between a mild embarrassment and Plate III. Readers who re-run Plate III in the simulator with its socks on will get a lower number still; §11 explains why the paper reports 9 anyway.

DerivationHow β was estimated from three and a half men

We fit β by holding the inseam at twelve inches and varying the pocket count across the four garments the sample happened to own. That is a smaller design space than we would have liked, and exactly the one we had by four in the afternoon. Respondents 1, 2 and 4 saw each count in ascending order and reported the drop after every addition; The half-respondent from §01 contributes the half. One free parameter, 0.887, fit all six points in Figure 3 to a residual we are choosing not to publish. Our competing additive model, a fixed number of points per pocket, fit worse at every count above six. That is the one part of this paper’s statistics we are fully confident in. Data not shown.

06Basketball shorts are the trough, and the gym is the only exception§

At ten inches the score bottoms out at 14 everywhere except inside a gym, where the same shorts score 61.

Plate IV · 10 in (25 cm)The Gym Troughknee-length black mesh basketball short, drawstring out
14Opportunity Index O
95% CI
3–31
Pool Breadth B
0.4
In the pool
people who are also in the gym at that moment
The trough is context-locked: 61 inside a gym.

Plate IV is the knee-length black mesh basketball short, drawstring out, ten inches (25 cm). Away from the gym O reads 14, interval 3 to 31, Pool Breadth 0.4: people who are also in the gym at that moment, and nobody else. Inside a gym the identical garment on the identical man scores 61. We did not photograph two garments. We photographed one garment in two contexts, and the score moved 47 points without anyone changing clothes.

Where Plate IV is wornO
Any venue outside a gym14
Inside a gym, same garment, same man61

We call this stretch the Gym Trough, and the trough is context-locked: the score belongs to the garment and the room together, never to the garment alone. That is, the room is a variable and the fabric is a constant. Say, for example, the same mesh short walks from the gym to the street. His score does not walk with it. Ten inches is the low point of the whole curve, lower even than the Cargo Cliff’s floor, and everything after Plate IV, however slowly, goes up.

Respondent 2’s mother was reached by phone as the external validator for this section. Her view: a drawstring left visibly out is trying too hard to look like it isn’t trying. We record this without comment and without a matching coefficient, because we have no way to quantify a mother’s tone. For the record, hers is the only external validation in the paper that agreed with everything else we found.

Part three · The False Recovery and the Singularity Below the knee the curve turns, and at five inches it does something nobody on the committee predicted.

07Seven inches is neither short nor a recovery, and the twelve points it gains come from somebody else§

Three inches above the knee the score climbs to 26, a twelve-point gain in which the general population takes no part.

Plate V · 7 in (18 cm)The False Recoveryseven-inch flat-front stone chino short, two pockets
26Opportunity Index O
95% CI
6–51
Pool Breadth B
0.5
In the pool
men, tentatively; no one else returns
Not a recovery, and not short.

Plate V is the seven-inch flat-front stone chino short (18 cm), two pockets, no drawstring, no phone. O reaches 26, interval 6 to 51, and Pool Breadth settles at 0.5. We name this stretch The False Recovery, and the name arrived late, after the transcripts had been coded and read back in the order the plates were shown.

Note the interval before anything else. It is 45 points wide with a lower bound of 6, which is the width you get when three and a half men disagree about a garment. Respondent 1 looked at Plate V for some time before lunch and said that’s a short, obviously, but it isn’t short. We coded that sentence three ways and kept none of them.

Respondent 2 was briefer: barbecue. Pressed on whether the eleven-pocket plate was also a barbecue, he said it was. Concretely, the garment we had built this section around was landing in the same bin as the bottom of the Cliff. The man doing the sorting saw no difference worth the walk to the shops.

Arithmetic agrees with Respondent 2. Climbing out of the Gym Trough, the Index gains 12 points over three inches, roughly four points an inch. Falling, it had given up 72, and it took thirty-seven of those without moving the hem at all, purely from pockets and a phone. A curve that loses ground that fast and buys it back at walking pace is not returning to the place it fell from.

Who came back, and who did not

We assumed the twelve points were the general population returning, because the general population is what our instrument was built to ask about. Its share is flat on the floor from Plate III onward, and at Plate V it does not lift by any amount we can measure. Put another way, nobody who left at the knee has come back. The twelve points are male interest, arriving early and quietly, and for two weeks we filed them as noise in the seven-inch cell.

Respondent 4 supplied the clearest statement of that noise and we did not hear it. Asked what he thought of Plate V, he said my flatmate would like it, and my flatmate is right about most things. We logged the answer under other and moved on to the next plate. Two inches further up the leg, the same audience stops arriving quietly.

08At five inches the pool changes composition, and it gets nearly five times bigger§

At five inches the score reaches 88 and Pool Breadth jumps from 0.5 to 2.4: everyone interested at seven inches stays, and men, Italians and everyone who has seen a film from 1978 arrive at once.

Plate VI · 5 in (13 cm)The Five-Inch Singularityfive-inch tailored olive short, clean hem, mid-thigh
88Opportunity Index O
95% CI
77–96
Pool Breadth B
2.4
In the pool
everyone from Plate V, plus men, plus Italians, plus anyone who has seen a film from 1978
The pool does not shrink. It broadens.

Plate VI is the five-inch tailored olive short (13 cm), clean hem, mid-thigh, and it is the finding this paper is actually about. O reaches 88, interval 77 to 96, the tightest interval in the study. Worth pausing on: the tightest interval belongs to the length the committee expected to be the riskiest. Plate V to Plate VI is 62 points over two inches, thirty-one points an inch, the steepest climb anywhere in the data.

Pool Breadth moves further than the score does. B goes from 0.5 at seven inches to 2.4 at five, nearly five times over; the precise ratio is 4.8×. We define the Singularity in the mathematical sense, as the one point on the Function where Pool Breadth stops changing continuously and jumps, and it sits at exactly five inches. Imagine the hem rising one inch at a time from ten: B answers in small changes until five, where it answers with a step. Figure 4 shows the jump as a wall rather than a slope, which is what the data insisted on.

Pool Breadth B, 0–2.5 inseam, inches 32 27 12 10 7 5 3 0 1.0 2.0 the Singularity men, Italians and anyone who has seen a film from 1978 the Milan-adjacent whoever is still looking runners venue- defined Nobody leaves at five inches; people arrive.
Figure 4Read as a market instead of a survey, the finding straight respondents called alarming is the best news in the dataset: at the Singularity the wearer’s addressable pool does not shrink, it grows 4.8×.

Who arrives, and who stays

Everyone interested at seven inches remains interested at five. Men join the pool. Italians join the pool. So, our data suggest, does everyone who has seen a film from 1978, a demographic we did not anticipate needing to define but were forced to, twice, in review. The general population stays where it fell at the knee, and its absence was priced into every plate after Plate III.

Respondent 2 gave us four words for the whole shift: that’s just a good short. Respondent 4 said his flatmate owns two pairs, which is the second time a flatmate has entered this dataset without consenting to it. Respondent 2’s mother, reached by phone as our external validator, was shown Plate VI over a video call and said finally, a man with knees. We had no coefficient for that and did not attempt one.

The pool does not shrink. It broadens.

What our instrument was measuring all along

Three re-readings of the transcripts were needed before we could state the obvious. The audience arriving at five inches is male. Male interest was there at seven inches too, in the twelve points we filed as noise, and it has been the thing our curve measured all the way from the knee. Our study was built by four men in a bar to measure how strangers read a hem, and it turns out to have measured, with some precision, which strangers were still looking. Such a design cannot see an audience it never thought to list.

This is good news by every measure the study has. A five-inch short costs the wearer nothing the seven-inch short had, and adds a set of people who were not counting him before. The interval tightens to its narrowest anywhere on the Function. Meanwhile the people arriving have been dressing better than the committee measuring them, for some time now. Respondent 1 asked, twice, whether the study was over. We told him it was not, recorded the question as data about the observer rather than the plate, and watched his score for Plate VI stay high. Our formal recommendation, stated once and not softened anywhere else in this paper, is five inches, clean hem, tailored, mid-thigh.

RobustnessThe Singularity survives every re-specification we tried, including the one where we removed Respondent 3

We re-ran the Plate V to Plate VI comparison four ways. First with all four respondents, then with the three who stayed, then with the external testimony standing in for the missing answer. Run four came the following morning, at the reviewer’s request. Pool Breadth held above 2.0 in every specification, and the jump from 0.5 never once smoothed into a slope. Removing Respondent 3 changed nothing, because he had left before Plate III and there was nothing of his left to remove. For a study with an effective n of 3.5 this is either remarkable robustness or a coincidence we are choosing not to examine further.

09At three inches only the runners are left, and they are extremely into it§

At three inches the mean says 71, and the mean is lying: runners score him 97 and everyone else 12.

Plate VII · 3 in (8 cm)The Marathon Anomalythree-inch navy split running short, side slit
71Opportunity Index O
95% CI
9–99
Pool Breadth B
1.1, bimodal
In the pool
runners, exclusively and intensely
A bimodal population pretending to be a mean.

Plate VII is the three-inch navy split running short (8 cm), side slit, thin white piping. Plate VII’s headline O is 71, interval 9 to 99, by far the widest interval in the study, and the width is the whole story. Runners score this plate 97. Everyone else scores it 12. Pool Breadth reads 1.1, but we flag it as bimodal, because 1.1 describes no population that exists at this length. Averaging two populations that barely overlap produces 1.1 and describes neither.

Say you meet him at a marathon start line and, separately, at a bar after work. At the start line he is, by our data, close to universally well received. At the bar he is closer to universally not. One mean number, 71, would suggest a man doing fine everywhere, and a mean is structurally unable to say otherwise. Respondent 1, a non-runner, called the side slit a lot of information, very fast. A marathon respondent, contacted separately for this section, called the same slit correct, structurally and declined to elaborate.

We call this stretch the Marathon Anomaly, and we ask any reader who quotes the 71 to quote the split on either side of it. Plate VII is, in one line, a bimodal population pretending to be a mean.

10At zero inches the Function stops being a function and becomes a map§

The swim brief has no score, only a venue: 99 in Mykonos, 2 at a suburban barbecue, 0 at the office.

Plate VIII · 0 in (0 cm)The Terminal Briefplain black swim brief, bare feet, the same pavement
venue-defined; see the table
95% CI
n/a
Pool Breadth B
n/a
In the pool
the venue decides
The Function stops being a function and becomes a map.

Plate VIII is the plain black swim brief, bare feet, the same pavement we used for every other plate: zero inches, zero centimetres. For the first time in eight plates there is no single value of O we are willing to report. We call it the Terminal Brief. Pool Breadth is not a number here either. Venue decides who is in the pool, which is itself the finding.

Where Plate VIII is wornO
A beach in Mykonos99
Rimini91
A Kyiv river beach74
A hotel pool66
A suburban barbecue2
The office0

For instance, the same black brief scores 99 on a beach in Mykonos and 0 in an office, with no change to the garment at all. All ninety-nine points sit on one side of the equation, and the trouser sits still on the other. Zero inches is where length stops explaining anything and venue takes over completely, and that turns out to be true of every garment on the curve, only less visibly.

Every garment has a venue matrix; the brief only makes it undeniable

Figure 5 extends the same logic to all eight plates against ten venues, scored optimal, viable, marginal or social death rather than as one number. Read across it and one line stands out. Cargo shorts are optimal in exactly two of the ten venues, a suburban barbecue and his mother’s kitchen, and nowhere else does Plate III win a single cell. Plate VI’s five-inch short is optimal in three, a bar after ten, Mykonos and Milan, and never reads as social death outside the office and a wedding. Which means the two shortest lengths in the study, both short, both leg-baring, behave nothing alike once the room is put back in.

optimal viable marginal social death not observed

Venue I32 in II27 in III12 in IV10 in V7 in VI5 in VII3 in VIII0 in
Office optimal viable social death social death social death social death social death social death
First date viable optimal social death social death marginal viable social death social death
Bar, after 10 pm viable viable social death marginal marginal optimal marginal social death
Wedding, as a guest optimal viable social death social death social death marginal social death social death
Gym social death social death marginal optimal viable viable viable social death
Beach, Mykonos social death marginal social death marginal marginal optimal viable optimal
Suburban barbecue viable marginal optimal viable viable marginal social death social death
Marathon start line social death social death social death marginal marginal viable optimal social death
Milan, any street viable optimal social death social death social death optimal marginal not observed
His mother's kitchen viable viable optimal viable viable marginal marginal social death
Figure 5Cargo shorts are optimal in exactly two venues on this matrix, the suburban barbecue and his mother's kitchen. The five-inch short is optimal in three — the bar after 10 pm, Mykonos, Milan — and is never social death outside the office and a wedding.
Part four · Interaction terms, the model, and the method Socks turn out to be a coefficient, the model goes interactive, and we show our working.

11Socks are not a garment; socks are a modifier§

A sock never changes the length, and it changes the score by up to 40 per cent, which makes it a coefficient rather than a piece of clothing.

We define the Sock Interaction Term, γ, as a multiplier applied to O by sock height, on top of whatever length and pocket count already produced. Because sock height and inseam interact rather than add, γ is best read off a phase diagram, not a formula. Figure 6 plots sock height against inseam and divides the plane into nine regions, each with its own multiplier and, because the Institute enjoys naming things, its own name.

32 24 20 14 9 8 5 4 2 0 0 3 6 9 12 15 18 THE SOCK INTERACTION TERM · PHASE DIAGRAM INSEAM L, IN — REVERSED (32 AT LEFT, 0 AT RIGHT) SOCK HEIGHT ABOVE SHOE, IN A · UNOBSERVABLE ×1.00 B · THE MILAN BONUS ×1.11 C · THE DAD CONFIGURATION ×0.60 D · THE TOURIST ×0.80 E · TENNIS, 1978 ×1.10 F · COASTAL GRANDFATHER ×0.90 H · THE ENTHUSIAST O=88, VENUE-DEP. G · THE RUNNER ×1.00 I · UNDEFINED ×0 no observations Two garments, one leg, forty per cent of the score.
Figure 6A sock is the cheapest decision on the page and the only one that can turn a good length into a bad one without touching the hem.

Nine regions, named in the order the data demanded them

Consider three of them, for example. No sock at all under a cropped trouser is the Milan bonus met in Plate II, ×1.11, the only region where removing a sock outperforms wearing one. Crew socks at cargo length are the Dad Configuration, ×0.60, the sharpest penalty in the table, and steep enough to explain most of the Cargo Cliff’s remaining variance. White crew socks on a four-to-eight-inch short are Tennis, 1978, ×1.10, so a specific photograph, worn correctly, outperforms the same outfit in ordinary crew socks two rows down.

RegionInseamSock heightγName
A20–32 inany×1.00Unobservable; the sock is inside the trouser
B24–29 innone×1.11The Milan bonus
C9–14 in5–7 in, crew×0.60The Dad Configuration
D9–14 in0–1 in, no-show×0.80The Tourist
E4–8 in5–7 in, white crew×1.10Tennis, 1978
F4–8 in10–14 in, knee×0.90Coastal Grandfather; rugby if under 30
G2–5 in0–1 in×1.00The Runner
H0–8 in15–18 in, over-kneepool shiftsThe Enthusiast; O = 88 in specific venues
I0–1.5 inany sock at all×0Undefined; do not attempt

Region G, the Runner, carries no penalty at all, which is one more reason Plate VII belongs to the population it belongs to. Region I we flag rather than explain. At an inseam of an inch and a half or less, any sock at all takes γ to zero. No respondent in our sample was willing to model it further. In other words, a swim brief with a sock on reads as an incident rather than a phase transition.

The sock we estimated after the plates were scored

One confession belongs here rather than in §14. We estimated γ after the eight plates had already been scored, and we did not go back. Plate III wears mid-calf white socks at twelve inches, which is Region C, the Dad Configuration. Apply the term to Plate III’s unrounded 8.6 and the reported 9 becomes 5.2. Our paper reports 9 throughout, the simulator in §12 reports 5.2 with a footnote, and the Institute regards both numbers as correct on the day each was produced. Post hoc; data not shown.

12You can run the model yourself, and it will tell you what your knees already knew§

The simulator below drives one man’s inseam from the shoe to the hip and reads the Function off him in real time; drag the slider and watch the pool change.

Five controls sit beside the figure, inseam, pocket count, a phone in the thigh pocket, sock height and venue, which is as many independent variables as this paper is willing to admit to. Move the inseam slider and the mannequin’s hem tracks it, with O and B recalculated at every position from the coefficients in Parts one to four. Set pockets to eleven, for example, and watch β take its cut before you can object to it. Toggle the phone and the Sag Term takes its flat seven without asking. Change only the venue at zero inches and the same brief scores 99 in Mykonos and 0 at the office: the fabric never moves, and the room does all the work. Eight plate buttons jump straight to each regime.

The Hemline–Opportunity Simulator · drag the figure to rotate

72.0
Opportunity Index, O
Pool Breadth B × 1.0
The Plateau
Who's in the pool: the general adult population.
viable
O = 72.0 = 72.0

Our model has real limits, and we would rather state them here than let a reader discover them by breaking the slider. Our mannequin has no face, no job and no year of birth. Neither, in a sense, did the study, which measured a garment against a stranger’s glance and nothing else about the man wearing it. Treat the simulator as a picture of the maths, that is, an illustration of the Function, and never as an instrument for judging anyone standing in front of you.

13Four men, one bar, and a confidence interval that spans everything§

We asked four men in a bar to be honest, one of them left, and we multiplied the rest up to four billion; the method section explains why this is fine.

Nine steps make up the method, and Figure 7 draws them as a paper would. Our research question formed in a changing room at 14:10, which is earlier in the process than most papers admit to. Four men at a bar near the Institute supplied the sample. Attrition before question two brought the effective sample to 3.5. One question served as the instrument: be honest. We extrapolated from three and a half men to roughly four billion adult men worldwide by multiplying. Figure 7 draws that step as a funnel which is, we concede, not to scale.

1 · RESEARCH QUESTION formed in a changing room, 14:10 2 · SAMPLE n = 4 men at a bar 3 · ATTRITION Respondent 3 left; n = 3.5 4 · INSTRUMENT the question “be honest” 5 · ANALYSIS p = 0.049, fourth attempt 6 · EXTRAPOLATION to 4.0 billion men, by multiplying 7 · PEER REVIEW the roommate, single-blind (asleep) 8 · PUBLICATION JAHS vol. 1, no. 1 — the only issue 9 · CITATION this paper, citing itself · h-index 1 SAMPLE, NOT TO SCALE — NOTHING IS 4 3.5 4,000,000,000 Every step is a real step in a real paper; only the sample is missing.
Figure 7This flowchart is indistinguishable from the one in a genuine paper until you read the numbers — which is the whole problem with flowcharts.

The fourth attempt

Analysis returned p = 0.049 on the fourth attempt. For readers meeting the term, p is the probability of seeing gaps this large by chance if trouser length made no difference at all. Below 0.05 a field agrees to be impressed. Anyone who has run statistics before will recognise what fourth attempt usually means, and we will pretend nothing: the first three analyses did not clear 0.05. Attempt four did, at 0.049, and we published attempt four. We name the practice rather than hide it, on the theory that a confession is not the same failure as the thing confessed.

Peer review fell to the corresponding author’s roommate, single-blind, in the specific sense that he was asleep at the time. He raised no objections. Our ethics committee, which is also the roommate, approved the study on the same terms. Publication followed in JAHS volume 1, number 1, currently the only issue, which this paper cites once, giving the journal an h-index of exactly 1.

MethodHow the four men were recruited, and what Respondent 3 said on the way out

Recruitment took eleven minutes and one round of drinks, which we treat as a cost of data collection rather than a confound. Our corresponding author approached the nearest occupied table and asked whether anyone would answer some questions about trousers, for a study with no funding and no compensation. Four men agreed, which we attribute to that specific bar rather than to trouser questions being generally welcome. Respondent 3 answered the first question, began the second, checked his phone, and left, saying only that he had somewhere to be. We did not follow up. Those three and a half men produced every number on this page.

14Limitations: all of them§

Every limitation a reviewer could raise applies to this study, and we list them so that no reviewer has to.

We considered withholding this section for about four minutes, on the reasoning that a paper volunteering its weaknesses is rarer than one hiding them and might read as less confident for it. Honesty scored better before lunch, and it scored better after. What follows is an inventory, offered in the same register as the findings, on the theory that a limitation stated plainly is at least honestly limited.

Seven limitations, in the order a reviewer would raise them

Consistent with prior work, which is to say consistent with this paper citing itself, we expect future research to address at least one item above. Ideally before JAHS publishes a second issue. Possibly by the same four men. Possibly by someone who was awake for it.

Conclusion

What this paper actually supports
  1. Wear trousers, or wear five inches.

    Nothing between thirty-two inches and five survives contact with a stranger, according to three and a half men. The Function finds no second plateau until exactly five inches, where O reaches 88 and the pool grows 4.8×.

  2. Count your pockets.

    Two are furniture. Every pocket after that keeps 88.7 per cent of what is left, and eleven of them, worn with a phone, are most of the Cargo Cliff on their own.

  3. A sock is a decision.

    Socks never move the hem, and they move the score by up to 40 per cent, almost all of it downward, depending which of nine regions they put you in.

  4. The barbecue is where the garden shorts win.

    Plus his mother’s kitchen. Everywhere else on the matrix, the Cargo Cliff is exactly that.

  5. Nothing on this page is true, and the curve is still correct.

    Every respondent, coefficient and roommate here is invented. Plateau, cliff, trough, slope, singularity: that shape is the part we would stand behind at a real bar.

15Appendix: supplementary material§

Table S1

DataTable S1: the raw scores of Respondents 1–4, before lunch
RespondentPresent at Q1Present at Q2Answered be honestNote
Respondent 1yesyesyesasked, twice, whether the study was over
Respondent 2yesyesyesmother reached by phone, separately
Respondent 3yesnopartiallyleft; counted as 0.5
Respondent 4yesyesyesdata not shown, by his own request
Externalby phoneby phoneunpromptedRespondent 2’s mother: obvious

Raw individual scores are withheld under the Institute’s standing data policy, written for this purpose. Plate-level means, published throughout the paper, are what remains once the raw scores are removed. Whether that is a summary statistic or the entire dataset depends on how the question is put to us.

Glossary

GlossaryEvery term the Institute coined, in the order it coined them
The Function
The whole Hemline–Opportunity curve, O(L), as inseam L runs from thirty-two inches to zero.
O
Opportunity Index, 0 to 100: the modelled probability a stranger who already finds the wearer acceptable keeps finding him acceptable after looking down. The looking down is the instrument.
B
Pool Breadth: how many distinct populations contain someone interested; 1.0 at the baseline, 2.4 at the only length where it stops behaving.
The Plateau
Plate I, thirty-two inches, where nothing happens, which is the point.
Milan bonus
×1.11 for bare ankle, no sock, at 24–29 inches.
Cargo Cliff
The 72-point fall between Plate II and Plate III, most of it pockets, the rest a phone.
β
The Pocket Coefficient: each pocket beyond the second removes 11.3 per cent of what remains.
Sag Term
A phone in a cargo thigh pocket, minus seven points, flat.
Gym Trough
Plate IV, ten inches, 14 everywhere except a gym.
The False Recovery
Plate V, seven inches, 26, men only, and tentatively even then.
The Singularity
Five inches, the one point where Pool Breadth jumps instead of moving.
Marathon Anomaly
Plate VII, three inches, a bimodal population pretending to be a mean.
Terminal Brief
Plate VIII, zero inches, where the Function becomes a map.
γ
The Sock Interaction Term: a multiplier on O set by sock height alone, the only coefficient in this paper contributed entirely by hosiery.

Ethics and funding

EthicsEthics statement, conflicts of interest and funding

Ethics. Approved by the Institute’s ethics committee on the morning after the study, on the same terms as peer review. All four respondents consented verbally; one consented to a single question and then left, which we treat as a withdrawal of consent for question two only.

Conflicts of interest. The corresponding author owns fourteen pairs of trousers and one pair of cargo shorts, photographed as Plate III and not otherwise defended. He has been to Milan once, for reasons unrelated to this study, and reports no financial interest in any hemline.

Funding. None. The trousers were the authors’ own. The round of drinks was expensed, in the sense that the corresponding author paid for it and has not forgotten.

References

ReferencesNine works cited, all of them formatted perfectly and one of them real
  1. Institute for Trouser Epidemiology (2026). Inseam length as a predictor of sexual opportunity in adult men: a longitudinal study of one afternoon. Journal of Applied Hemline Studies 1(1): 1–14. [This paper, citing itself.]
  2. Respondent 2’s mother (2026). Personal communication, by phone. Unprompted.
  3. The roommate (2026). Peer review of reference 1. Conducted single-blind; reviewer asleep. Unpublished, and likely to remain so.
  4. Ufficio Milanese della Caviglia (2019). Two inches of ankle: a standard. Milan: privately printed. [Not consulted; assumed.]
  5. Various (1978). Every film released that year, considered as a single garment reference. [Not consulted individually.]
  6. The Golden Retriever (2026). Null result, Plates III–VIII. Field notes, unaffiliated.
  7. Institute for Trouser Epidemiology (in preparation). The Sag Term: a rectangle of glass and what it costs. Journal of Applied Hemline Studies 1(2), should there be one.
  8. Ioannidis, J. P. A. (2005). Why most published research findings are false. PLoS Medicine 2(8): e124. [The real one. We recommend it.]
  9. Respondent 3 (2026). Somewhere to be. Personal communication, on the way out.