The slicer said forty grams and the spool is visibly lighter than that. Or a job estimated at six hours took nine, and the difference turned an evening print into an overnight one.
Both estimates are wrong in the same direction, and both are wrong for reasons you can find in under ten minutes.
The mass estimate assumes a density you supplied
A slicer computes the volume of plastic in the toolpaths and multiplies it by the density in the filament profile. That density is a number someone typed, and the default is usually the generic figure for the polymer family.
That is fine for plain PLA. It is not fine for the filled and blended grades that make up a growing share of what people actually buy. A mineral-loaded matte PLA runs nearer 1.35 g/cm³ than the 1.24 a default profile assumes — and on a part the slicer reports as 200 g, the real mass is 8.871% higher, which is nearly eighteen grams of filament that was never in the estimate.
Put your spool's real figure into the length and weight converter and see what it does to the metres. Glow-in-the-dark grades sit higher again because of the phosphor they carry; metal-filled grades are not in the same conversation at all, running several times the density of plain PLA, so a profile left on the default there is not slightly optimistic but wrong by a multiple.
Correcting that one field in the filament profile is the fastest useful thing on this page.
What the headline gram figure leaves out
Most slicers report the model's material and quietly exclude, or list separately in a place nobody reads:
- the prime line or purge blob at the start of every print
- brim and raft material
- support structures and their interfaces
- purge for colour or material changes, which on a multi-material print can exceed the part itself
- the length wasted at every filament change and at the end of a spool
None of these is large on its own. Together, on a supported part with a brim, they routinely add a tenth to the total.
The time estimate is a model of a machine that does not exist
Time estimates come from simulating the toolpath at the speeds in the profile. The simulation usually applies acceleration limits imperfectly, and it does not know about the machine's real behaviour: the flow ceiling that caps effective speed on fast profiles, the minimum layer time enforced on small layers, the pause while the bed heats, or the seconds lost to every retraction and travel.
The print time estimator on this site states an explicit ±30% band for that reason. Any tool that gives a print time to the minute without a band is expressing more confidence than the underlying model supports.
Learn your machine's error factor
The fix for all of this is one habit rather than a setting.
- Print something ordinary and note what the slicer predicted for mass and time.
- Weigh the finished part — including the brim and supports if you are honest about it — and note the real duration.
- Divide actual by predicted for both. Keep the two ratios.
After three or four prints you will have a personal correction factor for your machine, your materials and your profile, and it will be more accurate than any general advice. Most people find mass runs a few percent over and time runs considerably further over.
Then add the things that are not in any slicer
For work you are pricing rather than printing for yourself, the estimate needs three more terms that no slicer models at all: electricity, machine wear, and failures. A failure allowance is the one people leave out and the one that hurts — a job with a genuine one-in-ten failure rate costs materially more per good part than its nominal price, because the failed attempt consumed filament, power and hours.
The cost calculator takes all of them and shows each term separately, so you can see which one dominates rather than guessing. A short job in inexpensive filament is almost entirely a material cost; a job that runs for two days on a big plate is not, and the terms swap places.
The materials where estimates go worst
Anything filled or foamed. Matte, glow, metal-filled and wood-filled grades all differ from their base polymer's density, and foaming filaments actively change density during the print depending on temperature, so no single figure can describe one honestly. On those, weighing a test print is not a refinement; it is the only way to get a number worth quoting.