This changes how a number is computed. Your charts can move without your data moving. Each change is named below with the reason for it.
Metric: Suggested weight, and the hours per weighted point beside it, on the Story Point Distribution card (Velocity).
What changed: one weighted point now means the time of one typical 1-point task, measured over your Estimation baseline. A 1 keeps a suggested weight of 1, and every other estimate is suggested as the number of 1-point tasks its time equals. If fewer than ten 1-point tasks were timed, the smallest estimate that has ten takes this role instead and keeps its own value. Estimates with too few timed tasks to measure are suggested at the time per point of the nearest measured estimate, marked ≈ in the weights editor, so applying the suggestions keeps every estimate in the same unit.
Before, the suggestions were scaled so that the baseline's weighted total matched its story points. That kept the total stable but pushed small estimates above their face value, which was hard to explain: a 1-point task could be suggested at 1.4.
Which way the numbers move: when your larger estimates take less time per point than a 1 (the common case), every suggested weight goes down. For example, if a 1-point task takes 10 hours and a 13-point task 45 hours, a 13 is worth four and a half 1-point tasks: where the suggestions used to read 1.4 and 6.3, they now read 1 and 4.5. The ratio between estimates is exactly the same as before; only the unit changed. An estimate that takes more time per point than a 1 is still suggested above its face value.
Why: "a 13 is worth four and a half 1-point tasks" is something you can say in a planning meeting or to a board. "A 1 is worth 1.4 points" is not.
What happens to your numbers: nothing until an admin acts. Saved weights, and the weighted work they produce, are untouched. If you apply the new suggestions, weighted work will usually come out lower than your story points, because larger estimates no longer count at face value.
What to do: before applying, check which statuses your View counts as work (the card's gear). Waiting stages counted as work make small tasks look more expensive than they are.