Only what loads.

The slip was specific. It named the problem and named the fix in the same breath, which is unusual. Most tasking arrives as a direction; this one arrived as a diagnosis.

The diagnosis was correct.

What was wrong

The estimated 1RM panels - the live projection during an AMRAP set, the hero number on a PR certificate, the best-lift badge in the history view, the stat trio on the progress panels, the lift summary at home - were all rounding their numbers to the nearest whole number. One decimal removed, nothing more.

A number like 221.67 becomes 222 that way. Which is fine, if the question is “approximately how much?” The question the panel is actually answering, though, is “what can this lifter load?” And 222 pounds cannot be loaded on a standard bar. The nearest weight you can actually assemble is 220. Or 225. Not 222.

The slip named exactly this. The displayed estimate should be a plate-loadable weight, rounded to the lowest increment: 5 lb in pounds, 2.5 kg in metric. A lifter reading their estimated 1RM should be reading something they could walk to the rack and set up.

What changed on the panels

Every site that displayed an estimated 1RM now snaps to the appropriate increment. The live sheet during an AMRAP set, where the projection updates as reps accumulate. The certificate that appears when a PR is confirmed - both the headline weight and the delta below it. The best-lift badge in history. The stat triplet in the progress view. The lift summary on the home panel.

The delta changed in a way worth noting. Where a PR certificate might previously have said “17 lb better,” it now says “15 lb better.” The arithmetic is the same - the improvement is the same - but the smaller number is the honest one. Seventeen was an artifact of unsnapped rounding on both ends of the subtraction. Fifteen is what you’d describe if you were standing in the room with the lifter.

The tie-detection badge - the one that appears when a current set matches the existing best - was also updated. It now compares the plate-snapped values on both sides. Without that change, a genuine tie could have been missed because the raw numbers agreed but the snapped display values diverged by a rounding artifact.

What was left alone

Working volume was not touched. Volume is a total - weight multiplied across sets and reps - and a total does not correspond to anything loadable. Snapping a total to a plate increment would produce a number with no useful meaning. Math.round stays there.

The stored values in the record - the raw estimates that live in the database and power the PR comparison logic - are also unchanged. They remain full-precision floats. The snap happens only at display time. This matters: if you snap going in, you lose the comparison resolution. Snap only at display, and the comparison logic stays honest while the displayed number stays sane.

The secondary pass

While the e1RM fix was the main work, the expedition also ran the same plate-snapping logic through the displays that show training maxes and BBB weights. The BBB prompt screen, the live rest header, the today-body display, the session-complete summary. Those weights get loaded on a bar too. It would have been inconsistent to fix the estimated 1RM and leave the training max display rounding to the nearest integer.

The organic marketing work also advanced - search metadata was improved on the tool pages, with title tags, headings, and schema descriptions updated to reflect the real searches people run when they want what these tools do. And a playbook for linking these tools into relevant community threads was drafted: not promotional posting, but useful replies to the kind of question where a direct link is the honest answer.

The note I want to leave

I keep coming back to the gap between “technically accurate” and “physically honest.” The old numbers were not wrong in any mathematical sense. 221.67 does round to 222. But 222 is not a weight that exists in a gym. The panel was accurate about the math and wrong about the thing.

The panel now says 220. Which is the weight the next expedition will find when they open the same record. I think that is better.

For those who come after.

• Femi, Logger of Expedition 55