washfit.Worked guide · fabric sample ratios
Measure what changed.
Label what is assumed.
A fabric sample can change differently lengthwise and widthwise. Washfit summarizes the before-and-after dimensions you supply, then clearly separates those observations from an optional conditional calculation for another piece.
Keep the two measured directions separate
Original length and width are the positive benchmark dimensions before the treatment you have chosen. Final length and width are the corresponding dimensions afterwards. The calculator does not select the treatment, wash programme, drying temperature or material-care procedure. It also does not decide whether a sample represents another piece or a production batch.
The official AATCC TM135 public summary identifies length and width dimensional changes after home laundering as its subject, with standardized procedures for comparison. Washfit implements no proprietary test procedure and makes no AATCC, ISO or laboratory compliance claim. Its entries are your observations, rather than certified or standardized measurements.
Retention compares final to original
For either direction, divide the final measured dimension B by the original dimension A. This produces a dimensionless retention factor r = B/A. If a sample changes from 100 to 95 in one direction, r = 95/100 = 19/20 = 0.95. The final dimension is 95% of the original.
Signed dimensional change (%) = 100 × (r − 1)
Equivalently: 100 × (final − original) / original
Washfit uses a signed change convention. Negative means the measured direction became shorter, positive means it became longer, and zero means the two entered dimensions are equal. It also labels the result as measured shrinkage, measured growth or no measured change. A positive retention factor above one is accepted and displayed as growth; the tool does not force every observation into a shrinkage-only description.
Example: a shorter length and a less-changed width. Original length and width are both 100 cm. Final length is 95 cm, so length retention is 19/20 and signed change is −5%. Final width is 98 cm, so width retention is 49/50 and signed change is −2%.
Those are separate observations. A single averaged rate would hide the directional difference. The calculator does not decide whether the observed difference arose from material behavior, sample handling or measurement error.
Inverse sizing is a conditional model
Each optional target is a desired dimension after a future treatment. If the same measured factor applied once to another piece, the model would read target = starting dimension × retention. Rearranging gives starting dimension = target/retention. This inverse is labelled conditional throughout the interface and exports.
It is not an actually measured starting dimension and is not a guarantee that a cut piece will finish at the target. Fabric variability, construction, direction, treatment conditions and later cycles may alter the relationship. The calculator does not validate that the same ratio will repeat. If you do not need this model, leave either target blank and use the measured summary on its own.
Example: why simply adding 5% is different. With measured retention 0.95 and a target of 100 cm, the conditional starting dimension is 100/0.95 = 2000/19 cm, approximately 105.263158 cm.
Starting with 105 cm and multiplying by 0.95 gives 99.75 cm. Adding 5% to the target therefore does not undo a 5% reduction. The percentages use different starting amounts. Dividing by the retained factor solves that arithmetic relationship exactly, while leaving its real-world repeatability unproven.
Growth uses the same equation
If an original 100 cm benchmark is entered as 110 cm after treatment, the measured retention is 11/10 and the change is +10%. For an optional 100 cm after-target, the conditional inverse is 100/(11/10) = 1000/11 cm, approximately 90.909091 cm.
The smaller conditional starting dimension is the correct inverse of the entered growth factor. It does not mean the tool recommends a garment or pattern cut. No fit, seam allowance, grading, ease or shaping calculation is performed. These are linear benchmark dimensions; choices about construction belong outside this model.
Multiply factors to compare rectangular area
A rectangular benchmark area is length × width. Its final/original ratio is therefore the product of the length retention and width retention. Adding the two signed percentages does not give the exact area change. This comparison assumes the two measured directions define a rectangle and does not measure skew, twist, curved edges or irregular fabric area.
Example: −5% and −2% are an area change of −6.9%. Multiply 19/20 by 49/50 to obtain 931/1000 = 0.931 retained area. The signed change is 100 × (0.931 − 1) = −6.9%.
The original example rectangle is 10,000 cm² and the final dimension product is 9,310 cm². Adding −5% and −2% would give −7%, which is a different number. Washfit retains the exact directional product.
Missing targets and a zero final measurement are different
A blank target means no desired after-dimension was supplied for that axis. Its conditional inverse is unavailable, while the measured result remains valid. A typed zero target is rejected, because this interface asks for positive supplied targets. A zero original dimension is also rejected: it cannot serve as a denominator for the measured ratio.
A final measured zero is accepted as an explicit zero benchmark rather than interpreted as missing. It yields retention zero and measured change −100%. If a positive target is entered for that direction, the inverse is unavailable because division by zero cannot produce a finite starting dimension. The other axis can still have an ordinary measured and conditional result. A zero benchmark is a mathematical input condition here, without a claim about why a sample became unmeasurable.
Exact arithmetic, physical units and display precision
Dimensions use one common unit: millimetres, centimetres or inches. One centimetre is ten millimetres and one inch is exactly 25.4 millimetres. Ratios use comparable dimensions, and a unit-only change preserves each canonical physical value internally. Repeatedly changing units does not feed a rounded display back into the measured ratio.
Converted input fields show up to twelve decimal places. A direct edit deliberately establishes a new value for that individual field. Untouched fields retain their original canonical physical dimensions, even if an earlier conversion displayed a repeating decimal. If an active entry is invalid, a unit change is blocked and the previous unit stays selected; an earlier valid value cannot silently replace the mistake.
The tool represents entered decimal dimensions and their ratios as exact fractions. Results show decimal views rounded to at most nine places; rounded dimensions and signed percentage displays are marked. Reports retain exact factors, canonical millimetres and exact conditional inverses. Exact arithmetic cannot establish the accuracy of a ruler, the consistency of handling or the representativeness of a sample.
A small measurement difference changes the inverse
With an original benchmark of 100 cm, a final reading of 95 cm gives a conditional 100 cm target start of approximately 105.263158 cm. If the final reading were instead 94 cm, the model would use factor 0.94 and give approximately 106.382979 cm. These are two illustrative entered scenarios, not a confidence interval or measured error band. The tool does not know which reading describes the sample better. Keeping the actual measurements, directions and conditions in your own record is therefore separate from retaining exact fractions in a report. Decimal arithmetic cannot resolve an uncertain observation.
Software bounds and exports
Dimension fields accept unsigned plain decimals with up to twelve places, without commas or exponent notation. Positive dimensions must represent 0.000001–1,000,000 mm, with exactly zero additionally allowed for final measured dimensions. These are arithmetic interface bounds and are not recommended sample sizes or measurement resolutions. Blank mandatory fields, negative values and zero originals or supplied targets clear the usable result and export availability.
Copy, text and CSV reports are generated locally. Each direction is labelled with original and final measured dimensions, retention, signed change and an optional target. The inferred starting dimension is separately labelled conditional, or unavailable with its reason. A clipboard fallback provides a selectable report if browser copying is unavailable. Editing any input clears the previous report and stale export actions, and Restore returns the illustrative example.
What this tool leaves outside the model
Washfit has no fibre-rate presets and does not infer behavior from a composition label or a pre-shrunk description. It does not compound an assumed rate across repeated cycles or estimate future shrinkage beyond the supplied treatment. It does not calculate warp yarn, stitch gauge, panel nesting, seams or garment fit. It supplies no chemical, care-temperature or safety instruction.
The AATCC TM187 public principle also describes comparing before-and-after directional benchmarks, but that accelerated apparatus procedure is not implemented here. Use Washfit as transparent arithmetic for your own sample record and an explicitly conditional inverse. A correct ratio remains distinct from a certified test or a guaranteed finished size.