Cases

Everyone measured. It broke anyway.

One case per article. The claim first, the source beside it. Where a figure rests on a single organisation, that organisation is named. Where a figure could not be held up, it is listed as cut, with the reason. 9 figures are cut across these 5 articles.

1. Mars Climate Orbiter · 1999 · 14 Sept 2026 · 6 min

   Mars Climate Orbiter: a factor of 4.45, filed hundreds of times

   The software aboard the spacecraft was correct. A program on the ground wrote pound-seconds where the specification required newton-seconds, and every thruster firing entered the navigation model 4.45 times too weak.

   spaceflight · software · newton-second · NASA

   ×4.45 Video 00:25 1 source · 2 cut

2. Laufenburg bridge · 2004 · 13 Sept 2026 · 5 min

   Laufenburg: 27 centimetres, subtracted instead of added, became 54

   Germany measures height from the North Sea and Switzerland from the Mediterranean. The 27-centimetre offset between them was known, drawn correctly, and applied with the wrong sign.

   civil engineering · vertical datum · Rhine · surveying

   54 cm Video 02:08 1 source · 3 cut

3. SNCF platforms · 2014 · 12 Sept 2026 · 5 min

   SNCF: 1,300 platforms of 8,700, measured from the wrong thirty years

   The trains were built exactly to specification. The specification came from platform dimensions sampled only from platforms built in the previous three decades, and most French platforms are older than that.

   rail · sampling · specification · France

   1,300 Platforms requiring modification, out of 8,700 on the network. Video 03:49 2 sources · 2 cut

4. Vasa · 1628 · 11 Sept 2026 · 6 min

   The Vasa: two feet in one hull, one of 11 inches and one of 12

   Four carpenters' rules were recovered from the wreck, all slightly different. Two were marked in Swedish feet of twelve inches, two in Amsterdam feet of eleven. Every one of them was used correctly.

   shipbuilding · archaeology · customary units · Sweden

   14 mm 297 mm less 283 mm, per foot. The millimetre lengths come from one source and are attributed in the text. Video 05:24 3 sources · 1 cut

5. Hubble mirror · 1990 · 10 Sept 2026 · 6 min

   Hubble: NASA puts the spacing error at 1.3 millimetres

   The mirror was tested. The instrument built to test it had a lens in the wrong place, so the polishing machine stopped at the wrong moment and the test it had to pass was the test that was broken.

   optics · metrology · null corrector · NASA

   1.3 mm NASA's figure. No second organisation confirms it, so the article attributes it rather than stating it flatly. Video 07:07 2 sources · 1 cut