The 26 largest monthly rises in the series are the 26 Septembers

Dark navy briefing slide, masthead “Monthly production report · manufacture of textiles · Italy · NACE C13” on the left and “Production volume index, 2021 = 100 · 54 months · unadjusted, as published” on the right, headlined “Output nearly quadrupled in September.” A standfirst says there are fifty-four months, one line and every observation on the page; that the figures are the statistical office’s own, being the volume of production in Italian textile manufacturing published monthly as an index with 2021 = 100; that nothing is smoothed, interpolated, indexed to a base of the report’s own or rebased; that no month has been dropped, none averaged into a quarter and none is missing; that the months are plotted where they fall so a month of calendar occupies a month of chart, and the axis is linear, evenly ticked, zero-based and unbroken. It then reads off the chart: the index stood at 25.3 in August and 92.0 in September, a rise of 263.6% in a single month, the largest one-month gain in the window and the tenth largest in the twenty-six years the series has run — and concludes that the order book turned in September. The main panel, headed “Production in Italian textile manufacturing, January 2022 to June 2026” and subtitled “All 54 monthly observations · linear time axis, months plotted where they fall · zero-based index axis in even steps of 20 · unadjusted data, exactly as published”, is a gold line chart on an axis running from 0 to 140 in steps of 20, with a marked point at each of the fifty-four months and January of each year labelled along the bottom. The line is a violent saw-tooth: it runs between roughly 70 and 128 for most of the year and then plunges once a year to somewhere between 25 and 41 before climbing straight back, with a smaller dip each December. The last of the deep troughs, August 2025 at 25.3, is circled, joined by a dashed line that climbs its own column and turns right to the circled September point at 92.0, and labelled “August to September: +263.6% in one month” over the note “The largest one-month gain in the 54 months on the chart”. Six cards run down the right. “The rise, in one figure” prints 263.6% in large type over a note that this is the one-month gain between August and September 2025, from 25.3 to 92.0, the largest single-month rise in the 54 months on the chart and the tenth largest since the series began in 2000. “Read-out for the board” says the industry spent the summer at a standstill and came back in September at close to four times the August rate, and recommends re-forecasting the year off the September rate, releasing the held orders and bringing the second-shift restart forward a quarter. “The five largest monthly gains on record” tabulates September 2003 at +561.3%, September 2009 at +552.2%, September 2004 at +460.9%, September 2008 at +362.6% and September 2002 at +357.3%, and observes that textiles is an industry that moves in September, every one of the five largest one-month gains landing in that month, which is where the current gain belongs as well. “Why the axis starts at zero” explains that production volume has a real zero — no output at all — so the axis is drawn from it, nothing is truncated, broken or capped, and the distance between any two points is the change in the index between them. “The last twelve months, as published” lists all twelve values in two columns, with August and September 2025 picked out in gold. A “Basis of measurement” card credits Eurostat, Production in industry — monthly data (sts_inpr_m): Italy, NACE C13 manufacture of textiles, production volume, index 2021 = 100, unadjusted, dataset last updated 5 September 2026 and extracted 7 September 2026. A method line runs along the foot: official monthly index, production volume; 54 consecutive months, none omitted; linear time axis; zero-based index axis in even steps of 20; no smoothing, no interpolation, no rebasing, no exclusions. A footnote credits the source, says all 54 values are as published, and states that the report is a demonstration drawn by us in the manner of a monthly production briefing — the board, the read-out and the recommendation are invented, and every figure behind them is real. — Showing the misleading chart

The same months on a cream background, kicked “The same index · the same 54 months · three published versions · one shared zero-based axis” and headlined “August is not a downturn. August is August.” A standfirst says nothing on the report was cropped, truncated, smoothed or rescaled and every figure printed on it is correct; that what sets the picture is that this series has a season in it larger than anything else it contains, and the report is drawn from the version with the season left in; that Italian textile plants close for the summer holidays, so the index falls to about a third of its underlying level every August and climbs back every September; that August is the lowest month of the year in all twenty-six complete years on record, the twenty-six largest one-month rises in the whole 317-step series are the twenty-six Septembers and the twenty-six largest falls are the twenty-six Augusts, the first move that is not a factory holiday being the reopening after the spring 2020 shutdown at +131.4%, which ranks twenty-seventh; and that the same statistical office publishes the same months with the season taken out, in the same table, where August to September 2025 is −1.2% and the four and a half years the report draws as a saw-tooth are a decline of 22.6% once the season is out. The upper panel, “One series, three published versions”, subtitled “Same 54 months, same zero-based axis, 2021 = 100 · unadjusted · calendar adjusted · seasonally and calendar adjusted · every observation plotted where it falls · dotted verticals mark each August”, draws three lines on one axis from 0 to 140: the unadjusted series in olive, saw-toothed, plunging to its four deep August troughs; the calendar-adjusted series in a dashed blue that tracks it closely, showing that working days account for very little of the swing; and the seasonally and calendar adjusted series in a thick gold line that runs almost flat, drifting downward from a marked 100.1 at January 2022 to 77.5 at June 2026. Each line is labelled at its right-hand end with its final value — unadjusted 83.2, calendar adjusted 82.6, seasonally and calendar adjusted 77.5 — and a gold caption across the top reads “Seasonally and calendar adjusted: 100.1 → 77.5, −22.6% over the 54 months”, over the note that the same office publishes all three versions of these months, side by side, in the same table. Four faint dotted verticals labelled Aug mark the four Augusts. Three panels sit below. “What each calendar month is worth” plots the median of unadjusted divided by seasonally adjusted for each calendar month across the 26 complete years, as twelve bars against a dashed reference line at 1.00: ten of them stand between 0.93 and 1.18, while August stands at 0.32 and December, four bars along at the right-hand end, at 0.78. “Month-on-month change, one shared axis” draws all 53 steps in the window as bars from zero on a scale running from −80% to beyond +200%, Augusts and Septembers in gold and every other month in olive; a pair of gold hairlines close on either side of zero marks the whole range of the adjusted series, −6.8% to +3.8%, captioned as a pencil line on this scale, and the four gold spikes above and the four below dwarf everything else on the panel — the tallest, September 2025, carrying its value, +263.6%, printed above it. “The same three months, three ways” tabulates July, August and September 2025 in each of the three published versions — 98.3, 25.3 and 92.0 unadjusted; 94.7, 33.3 and 89.9 calendar adjusted; 81.8, 82.4 and 81.4 seasonally and calendar adjusted — with the August-to-September change on each: +263.6%, +170.0% and −1.2%. A note explains that the published calendar adjustment lifts August from 25.3 to 33.3 — the largest lift any August in the record has been given, and more than the working-day count alone would buy, at 20 against a normal month’s 21 or 22 — and that the shutdown is the rest, since +170.0% survives the correction. A second table under it, headed “September against September, the cheap fix”, gives the year-on-year change on each version: unadjusted 88.7 to 92.0, +3.72%; calendar adjusted 92.0 to 89.9, −2.28%; seasonally and calendar adjusted 83.8 to 81.4, −2.86%. Six cards run down the right. “What the report was actually showing” says that taking the season out turns the saw-tooth into one slow line running from 100.1 in January 2022 to a peak of 106.4 in May 2022 and then down to 77.5 in June 2026, −22.6% across the window, that the first half of 2026 runs between 80.4 and 77.5, and that the June figure is the lowest reading anywhere in the 54 months — the news on the page, and news the raw chart had no way to show. “Twenty-six Septembers, twenty-six Augusts” records that of the 317 month-on-month changes since January 2000 the twenty-six largest rises are the twenty-six Septembers and the twenty-six largest falls are the twenty-six Augusts, the smallest September rise on record being +154.2% and the smallest August fall −61.8%, with the twenty-seventh rise May 2020 at +131.4% and the twenty-seventh fall April 2020 at −50.7% — so a pandemic ranks below every ordinary summer holiday. “Year on year, and its own catch” says that comparing September with September removes the season for nothing and is the right first move, but that here it disagrees with itself: unadjusted, September 2025 is +3.7% on September 2024, and calendar adjusted it is −2.3%, because September 2024 had 21 working days and September 2025 had 22 — one extra day at the same daily rate being about 4.8% more output, enough on its own to flip the sign — and that year on year also answers a year late and inherits whatever happened in the month it is dividing by. “Three series, one table” explains that unadjusted is what the machines made, calendar adjusted takes out working days, weekends and public holidays since a month with an extra Monday is not a better month, and seasonally and calendar adjusted takes out the repeating within-year pattern as well, leaving trend and news; national offices estimate it with X-13ARIMA-SEATS or TRAMO-SEATS and publish it beside the raw figures. “When the raw series is the right one” says nobody ships a seasonally adjusted bale — shifts, lorries, electricity, warehousing and the cash that pays for them all run on the unadjusted line, and a plant that makes a quarter of its usual output in August has to be staffed for exactly that — so the raw series answers how much and the adjusted series answers how are we doing, the failure being not using one but using one to answer the other’s question. A “Basis of measurement” card credits Eurostat sts_inpr_m for Italy, NACE C13, production volume, index 2021 = 100, in all three adjustments, 318 monthly observations from January 2000 to June 2026 with the charts showing January 2022 to June 2026, dataset last updated 5 September 2026 and extracted 7 September 2026, working days counted by us. Five notes run along the foot: “Every observation plotted”, “Months where they fall”, “One zero-based axis for all three series”, “Nothing smoothed or interpolated” and “The season named, not quietly removed”. A closing footnote credits the source, notes that monthly index values are revised as later returns arrive and that these are as published on 5 September 2026, and states that both drawings are ours, that the board, the read-out and the recommendation on the first are invented, and that every measurement on both is real. — Showing the misleading chart

Be the first to star this exhibit

A production report plots 54 months of Italian textile output on a zero-based, unbroken, evenly ticked axis, every observation where it falls, nothing smoothed. The index reads 25.3 in August and 92.0 in September — a rise of 263.6% in one month — and the board is told the order book has turned. Italian plants close for the summer: August has been the lowest month of the year in all 26 complete years on record, the 26 largest one-month rises in the whole 317-step series are the 26 Septembers, and the 26 largest falls are the 26 Augusts. The same office publishes the same months with the season taken out, in the same table. On that column, August to September is −1.2%, and the four and a half years the report draws as a saw-tooth are a decline of 22.6% once the season is out.

Eurostat, “Production in industry — monthly data” (sts_inpr_m): Italy, NACE C13 manufacture of textiles, production volume, index 2021 = 100, in all three published adjustments — unadjusted, calendar adjusted, and seasonally and calendar adjusted. 318 monthly observations, January 2000 to June 2026; the exhibit charts January 2022 to June 2026. Dataset last updated 5 September 2026; extracted 7 September 2026. Month-on-month and year-on-year changes, the monthly median of unadjusted ÷ seasonally adjusted, and the working-day counts are computed by us, 2026Added September 7, 2026

01The claim

Fifty-four months, one line, and every observation on the page: the volume of production in Italian textile manufacturing, NACE C13, published monthly by the statistical office as an index with 2021 = 100. Nothing is smoothed, interpolated, indexed to a base of our own or rebased. No month has been dropped, none averaged into a quarter, and none is missing. The months are plotted where they fall, so a month of calendar occupies a month of chart, and the axis is linear, evenly ticked and unbroken — and it starts at zero, because production volume has a real zero to start from, so the vertical distance between any two points is the change in the index between them. Read off the chart: the index stood at 25.3 in August and 92.0 in September, a rise of 263.6% in a single month. That is the largest one-month gain in the fifty-four months on the page and the tenth largest in the twenty-six years the series has run. The five largest of all are September 2003 at +561.3%, September 2009 at +552.2%, September 2004 at +460.9%, September 2008 at +362.6% and September 2002 at +357.3%. Textiles is an industry that moves in September. Read-out for the board: the order book turned. Re-forecast the year off the September rate, release the held orders, and bring the second-shift restart forward a quarter.

02The trick

Every figure on that slide is right, and the read-out it produces is wrong, because the largest thing in this series is the calendar. Italian textile plants close for the summer holidays, so the index falls to about a third of its underlying level every August and climbs back every September — and it does so with a regularity that is hard to overstate once you go and count. August is the lowest month of the year in all twenty-six complete years on record. Of the 317 month-on-month changes since January 2000, the twenty-six largest rises are the twenty-six Septembers and the twenty-six largest falls are the twenty-six Augusts; the smallest September rise in the record is +154.2% and the smallest August fall is −61.8%. The first move in the ranking that is not a factory holiday is the reopening after the spring 2020 shutdown, at +131.4%, and it comes twenty-seventh — so a pandemic ranks below every ordinary summer in the series. The slide’s own boast gives it away, too: all five of the largest gains it cites are Septembers, which is a fact about the month rather than about the industry. What makes this different from a bent axis is that the fix already exists and was published in the same table. The same office puts out these months in three versions, and the report is drawn from the first of them. Take the calendar out — the working days, the weekends and where the public holidays fall, which the office estimates on its own model rather than by counting days — and August rises from 25.3 to 33.3. That is the largest lift any August in the record has been given, and more than the working-day count on its own would buy, August 2025 having held 20 of them against a normal month’s 21 or 22. The August-to-September gain is still +170.0%: the calendar explains part of it, the shutdown the rest. Take the repeating within-year pattern out as well and the same two months read 82.4 and 81.4, a change of −1.2%. On that column the whole window is one slow line: 100.1 in January 2022, a peak of 106.4 that May, and 77.5 by June 2026 — a decline of 22.6% that has been running underneath the saw-tooth the entire time, invisible because the season is several times larger than the trend. The first half of 2026 runs between 80.4 and 77.5, and that June figure is the lowest reading anywhere in the 54 months — which is the only piece of news on the page, and the raw chart has no way to carry it. Even the cheap fix has a catch worth knowing, because it is the one most people reach for. Comparing September with September removes the season for nothing, and here it disagrees with itself: unadjusted, September 2025 is 3.7% up on September 2024; calendar adjusted it is 2.3% down. September 2024 had 21 working days and September 2025 had 22, and one extra working day at the same daily rate is about 4.8% more output — enough on its own to flip the sign. Year on year also answers a year late, and inherits whatever was odd about the month in the divisor. (Both drawings are ours; the board, the read-out and the recommendation on the first are invented, and every measurement on both is real.)

03The fix

Ask first whether the move repeats. A shape that appears at the same point in every year is a season, a season carries no news, and the check costs one glance at the same month last year. Then compare like with like: December against December, September against September, which holds the calendar constant without a model and is the cheapest correction available. Know its two costs before leaning on it — it answers a year late, so a turn in March arrives diluted for months, and it inherits whatever was strange about the period it divides by — and remember it does nothing about working days, which here were enough on their own to flip a sign. Where the series is published seasonally adjusted, take the published column. It is made by people with the whole history and a documented method, X-13ARIMA-SEATS or TRAMO-SEATS, estimating the repeating within-year pattern along with the trading-day and holiday effects; this is not a calculation to reproduce by eye from a chart, and the usual failure is not disagreeing with it but never noticing that the query came back with the raw column. Where no adjusted version exists, a twelve-month moving average is a blunt and honest stand-in for the trend, provided you say so and remember it ends six months short of today. Say which version you drew, in the caption, the way you print a unit: “seasonally adjusted” is two words that change how every gap on the chart should be read. Better still, draw both, since the distance between the raw line and the adjusted one is exactly the season, it costs one extra series, and it is worth seeing once — put them on one shared zero-based axis so the reader can measure the gap rather than take it on trust. And keep the raw series, because it is not the inferior version of anything. Nobody ships a seasonally adjusted bale: shifts, lorries, electricity, warehousing and the cash that pays for them all run on the unadjusted line, and a plant making a quarter of its usual output in August has to be staffed for exactly that. The raw series answers how much, the adjusted series answers how are we doing, and the failure is never using one of them — it is using one to answer the other’s question.