Changelog

September 2026, page 2

Jaccard Index Calculator

Last reviewed 13 September 2026

Recorded that all three ignore joint absence, with the practical test for whether that matters: would you be surprised if both sets lacked an item?

5updates in all, on the calculator's own page

Herfindahl Index Calculator

Last reviewed 13 September 2026

Contrasted it with the Gini on the same page: a thousand identical firms have a Gini of 0 and an HHI of 10; two identical firms have a Gini of 0 and an HHI of 5,000. The Gini ignores how many players there are and the HHI does not, which is the whole difference for a competition question.

5updates in all, on the calculator's own page

Gini Coefficient Calculator

Last reviewed 13 September 2026

Recorded that the Gini is most sensitive to transfers in the MIDDLE of the distribution and least sensitive at the top — which is the opposite of what most commentary assumes, and the reason top-share measures have displaced it for concentration questions.

5updates in all, on the calculator's own page

Shannon Entropy Calculator

Last reviewed 13 September 2026

Recorded that the plug-in estimate is biased DOWNWARD when the data is a sample rather than a population, by roughly (k−1)/(2n ln 2) bits — the opposite direction from most estimation problems.

5updates in all, on the calculator's own page

Life Table Calculator

Last reviewed 13 September 2026

Set out the one situation where this method is the only option — data that arrives already grouped, with no individual times — and measured what grouping costs when it is not: 11.1440 percentage points on the shipped registry-style preset.

5updates in all, on the calculator's own page

Hazard Ratio Calculator

Last reviewed 13 September 2026

Verified that the interval is exactly symmetric on the log scale across 100 generated inputs, and that the standalone entry point agrees with the log-rank one to machine precision.

5updates in all, on the calculator's own page

Log-Rank Test Calculator

Last reviewed 13 September 2026

Verified that swapping the arms flips z, leaves χ² unchanged and inverts the hazard ratio, and that total expected equals total observed across both arms on every generated trial.

5updates in all, on the calculator's own page

Kaplan-Meier Calculator

Last reviewed 13 September 2026

Verified that with nothing censored every step equals the plain running proportion still alive, and that survival never rises, stays inside [0,1] and keeps its confidence limits there across 120 generated cohorts.

5updates in all, on the calculator's own page

Etsy Fee Calculator

Last reviewed 13 September 2026

Tested that the fixed payment component makes the effective fee rate rise as order value falls.

3updates in all, on the calculator's own page

Flipkart Seller Fee Calculator

Last reviewed 13 September 2026

Tested that raising the GST rate on fees reduces settlement without changing any underlying fee.

3updates in all, on the calculator's own page

PayPal Fee Calculator

Last reviewed 13 September 2026

Tested that adding the fee percentage to a target net understates the required gross, which is the error reverse mode exists to prevent.

3updates in all, on the calculator's own page

Crypto Profit Calculator

Last reviewed 13 September 2026

Tested that selling at the purchase price always loses money once both fees apply, and that break-even sits above the entry price.

3updates in all, on the calculator's own page

Bitcoin Return Calculator

Last reviewed 13 September 2026

Tested that the return is measured against the full amount invested including the fee, not the post-fee amount.

3updates in all, on the calculator's own page

Portfolio Value Calculator

Last reviewed 13 September 2026

Tested that the largest position never exceeds the total and that the holdings always sum back to it.

3updates in all, on the calculator's own page

Average Order Value Calculator

Last reviewed 13 September 2026

Tested that a target equal to current order value requires no uplift, and that the two uplift figures always agree with each other.

3updates in all, on the calculator's own page

Cart Abandonment Rate Calculator

Last reviewed 13 September 2026

Tested that completing every cart gives a zero rate, and that recoverable revenue scales with both order value and recovery rate.

3updates in all, on the calculator's own page

MER Calculator

Last reviewed 13 September 2026

Tested that the ratio and the percentage are exact reciprocals of each other at every level of spend.

3updates in all, on the calculator's own page

Ad Spend Calculator

Last reviewed 13 September 2026

Tested that the implied marketing share of revenue is always the reciprocal of the ROAS entered.

3updates in all, on the calculator's own page

Reorder Point Calculator

Last reviewed 13 September 2026

Tested that a zero lead time leaves only the safety buffer, and that the point scales linearly with both demand and lead time.

3updates in all, on the calculator's own page

Safety Stock Calculator

Last reviewed 13 September 2026

Tested that identical maximum and average figures produce no buffer at all, and that the result is never negative.

3updates in all, on the calculator's own page

Stock Coverage Calculator

Last reviewed 13 September 2026

Tested that weeks of cover is always days divided by seven, and that zero demand does not divide by zero.

3updates in all, on the calculator's own page

Stockout Cost Calculator

Last reviewed 13 September 2026

Tested that a selling price below unit cost floors the margin at zero rather than reporting a negative loss.

3updates in all, on the calculator's own page

Inventory Carrying Cost Calculator

Last reviewed 13 September 2026

Tested that monthly cost is always a twelfth of annual and that both scale linearly with the rate.

3updates in all, on the calculator's own page

Sell-Through Rate Calculator

Last reviewed 13 September 2026

Tested that selling everything received gives one hundred percent, and that remaining units reconcile with the two counts.

3updates in all, on the calculator's own page

Demand Forecast Calculator

Last reviewed 13 September 2026

Tested that a flat three-month history makes both methods agree, and that a zero growth rate leaves the average untouched.

3updates in all, on the calculator's own page

GMV Calculator

Last reviewed 13 September 2026

Tested that zero returns and cancellations makes net merchandise value equal gross, in both modes.

3updates in all, on the calculator's own page

Repeat Purchase Rate Calculator

Last reviewed 13 September 2026

Tested that repeat customers exceeding the total are capped rather than producing a rate above one hundred percent.

3updates in all, on the calculator's own page

Customer Retention Rate Calculator

Last reviewed 13 September 2026

Tested that retention and churn always sum to one hundred, and that a growing base cannot produce a rate above one hundred.

3updates in all, on the calculator's own page

Ecommerce Valuation Calculator

Last reviewed 13 September 2026

Tested that add-backs raise the valuation by exactly the multiple times the add-back, and that both valuations scale linearly.

3updates in all, on the calculator's own page

Inventory Investment Calculator

Last reviewed 13 September 2026

Tested that doubling the target days doubles the cycle stock investment while leaving the safety component untouched.

3updates in all, on the calculator's own page

Daily Sales Target Calculator

Last reviewed 13 September 2026

Tested that the sessions figure always reproduces the order count at the conversion rate entered.

3updates in all, on the calculator's own page

Order Defect Rate Calculator

Last reviewed 13 September 2026

Tested that the threshold flag flips exactly at one percent, and that defect-free orders always reconcile with the total.

3updates in all, on the calculator's own page

Return Loss Calculator

Last reviewed 13 September 2026

Tested that a zero return rate produces no loss at all, and that a higher damaged share always increases the total cost.

3updates in all, on the calculator's own page

Exchange Cost Calculator

Last reviewed 13 September 2026

Tested that the total is always the per-exchange figure times the count, and that each cost component contributes one for one.

3updates in all, on the calculator's own page

India GST Calculator

Last reviewed 13 September 2026

Tested that extracting tax from an inclusive price is division rather than a percentage of the total, and that the two supply types collect the same amount.

3updates in all, on the calculator's own page

UK VAT Calculator

Last reviewed 13 September 2026

Tested that removing VAT from a gross price uses the VAT fraction rather than the percentage, which differs by a fifth at the standard rate.

3updates in all, on the calculator's own page

Canada GST HST PST Calculator

Last reviewed 13 September 2026

Tested that a zero provincial rate reduces the result to federal GST alone, and that both rates apply to the base rather than compounding.

3updates in all, on the calculator's own page

Salary TDS Calculator

Last reviewed 13 September 2026

Tested that tax already deducted beyond the full estimate leaves nothing further to collect rather than a negative deduction.

3updates in all, on the calculator's own page

Contractor TDS Calculator

Last reviewed 13 September 2026

Tested that switching the recipient type changes the deduction by exactly the difference between the two rates.

3updates in all, on the calculator's own page

Tolerance Interval Calculator

Last reviewed 12 September 2026

Stated that the one-sided factor is a different quantity, not half a two-sided interval — which is a common and consequential error in specification setting.

5updates in all, on the calculator's own page

Prediction Interval Calculator

Last reviewed 12 September 2026

Published the factor table generated by the same code that computes the interval, so the two can never drift apart.

5updates in all, on the calculator's own page

Permutation Test Calculator

Last reviewed 12 September 2026

Stated the one assumption that survives — exchangeability — as a claim about how the data was collected rather than about how it is distributed.

5updates in all, on the calculator's own page

Bootstrap Confidence Interval Calculator

Last reviewed 12 September 2026

Seeded the generator. A bootstrap interval that changes every time you look at it cannot be put in a paper, and the suite asserts both that the same seed reproduces it and that a different seed does not.

5updates in all, on the calculator's own page

Games-Howell Test Calculator

Last reviewed 12 September 2026

Verified that a pair is significant exactly when its confidence interval excludes zero, across 100 generated datasets, and that Welch degrees of freedom never exceed the pooled degrees of freedom they replace.

5updates in all, on the calculator's own page

Dunnett's Test Calculator

Last reviewed 12 September 2026

Made the whole thing fast enough to be a page rather than a batch job — the quadrature weights are built once per degrees of freedom and the root is found by the Illinois method rather than bisection, which took one critical value from 3.4 seconds to under 200 milliseconds.

5updates in all, on the calculator's own page

Omega Squared Calculator

Last reviewed 12 September 2026

Corrected an error this site had made in its own reasoning: omega squared is NOT always below epsilon squared. They share a numerator and differ only in denominator, so below F = 1 the ordering flips. The assertion suite caught it on 11 of 300 random tables, and the rule that actually holds — same sign, omega always nearer zero — is now what is asserted.

5updates in all, on the calculator's own page

Eta Squared Calculator

Last reviewed 12 September 2026

Verified the two-group identity that eta squared equals the squared point-biserial correlation, against the site's own point-biserial engine, agreeing to 1.4e-7.

5updates in all, on the calculator's own page

Repeated Measures ANOVA Calculator

Last reviewed 12 September 2026

Printed the subject row as a share of total variation, because that share — 99.25% on one preset — is the entire argument for the design, and a between-subjects analysis would have left all of it in the error term.

5updates in all, on the calculator's own page

Two-Way ANOVA Calculator

Last reviewed 12 September 2026

Verified that SS_A + SS_B + SS_AB + SS_error reproduces SS_total across 200 randomly generated designs, and that eta squared sums to exactly 1 across all four rows in 200 more.

5updates in all, on the calculator's own page

Standardized Mortality Ratio Calculator

Last reviewed 12 September 2026

Verified that the SMR interval is exactly the Poisson interval on the observed count rescaled, at six different observed counts, and that expected events scale exactly linearly with person-time.

5updates in all, on the calculator's own page

Incidence Rate Calculator

Last reviewed 12 September 2026

Verified that the reporting multiplier scales the rate and both limits by exactly the same factor, and that the interval always contains the point estimate and widens monotonically as events fall.

5updates in all, on the calculator's own page

Attributable Risk Calculator

Last reviewed 12 September 2026

Reported protective exposures with their signs rather than flipping them silently: RR = 0.25 gives an attributable fraction of -300%, which is the arithmetic saying the preventable fraction is the measure wanted.

5updates in all, on the calculator's own page

Likelihood Ratio Calculator

Last reviewed 12 September 2026

Verified the odds-form update against direct Bayes across 300 random tests, agreeing to within 1e-12, so the arithmetic shortcut the page teaches is checked rather than asserted.

5updates in all, on the calculator's own page

ROC Curve & AUC Calculator

Last reviewed 12 September 2026

Recorded that an AUC below 0.5 is a good marker pointing backwards — reversing the score gives exactly 1 - AUC, asserted across 120 random datasets — rather than a worthless one.

5updates in all, on the calculator's own page

Cramer's V Calculator

Last reviewed 12 September 2026

Recorded that Cohen's small/medium/large marks fall as degrees of freedom rise (0.10/0.30/0.50 at df = 1 but 0.06/0.17/0.29 at df = 3), so an unlabelled 'medium effect' is not checkable. Verified against 69 assertions.

5updates in all, on the calculator's own page

Phi Coefficient Calculator

Last reviewed 12 September 2026

Returned no result rather than zero for an empty row or column, since undefined and 'no association' are different findings. Verified against 69 assertions.

5updates in all, on the calculator's own page

Point-Biserial Correlation Calculator

Last reviewed 12 September 2026

Verified against 69 assertions including the Pearson identity, the r = t/sqrt(t^2+df) relation, and the constant-column and single-group cases that must return no result.

5updates in all, on the calculator's own page

Intraclass Correlation Calculator

Last reviewed 12 September 2026

Reported negative ICCs as computed, and verified the whole set against 69 assertions built on the shared ANOVA decomposition so the six forms can never disagree about the data underneath them.

5updates in all, on the calculator's own page

Cronbach's Alpha Calculator

Last reviewed 12 September 2026

Verified against 69 assertions covering alpha, the drop-one recomputation, the Spearman–Brown identity that generates the length table, and the degenerate matrices that must return no result.

5updates in all, on the calculator's own page

Friedman Test Calculator

Last reviewed 12 September 2026

Verified against 69 assertions covering midranks for ties, the omnibus statistic, W, and the ragged designs that must return no result rather than a number.

5updates in all, on the calculator's own page

Runs Test Calculator

Last reviewed 12 September 2026

Verified against the 69-assertion sequence suite, including the exact run-count distribution on small sequences where the normal approximation should not be trusted.

5updates in all, on the calculator's own page

Goodness of Fit Calculator

Last reviewed 12 September 2026

Verified against 69 assertions covering the statistic, the residuals, the calibration sweep and the boundary cases where a category has zero expected count.

5updates in all, on the calculator's own page

Bartlett's Test Calculator

Last reviewed 12 September 2026

Verified against 53 assertions shared with the F-test and Levene pages, including the calibration sweep that produces the 4.20% and 40.40% figures.

5updates in all, on the calculator's own page

Levene's Test Calculator

Last reviewed 12 September 2026

Verified against the same 53-assertion suite as the F-test and Bartlett pages, which is what lets the three pages quote each other's numbers without drift.

5updates in all, on the calculator's own page

F-Test Calculator

Last reviewed 12 September 2026

Verified against 53 assertions covering the statistic, the interval, the calibration sweep and the degenerate cases where a group has no variation at all.

5updates in all, on the calculator's own page

Laplace Distribution Calculator

Last reviewed 12 September 2026

Measured the tail: matched on standard deviation the Laplace tail is LIGHTER than the normal's inside about one sigma and heavier past two, with the ratio climbing from there. Excess kurtosis is exactly 3.

5updates in all, on the calculator's own page

Gamma Distribution Calculator

Last reviewed 12 September 2026

Shape 1 is verified to reproduce the exponential CDF and density exactly, and the skewness is reported as 2/sqrt(k) - the central limit theorem visible in a single parameter.

5updates in all, on the calculator's own page

Sensitivity and Specificity Calculator

Last reviewed 12 September 2026

Predictive values at an arbitrary prevalence are computed from Bayes' theorem rather than rescaled from the observed table, so the sample's own prevalence never leaks into the answer.

5updates in all, on the calculator's own page

Number Needed to Treat Calculator

Last reviewed 12 September 2026

The number needed to harm is the same arithmetic with the sign reversed, and the tool switches the label rather than reporting a negative NNT.

5updates in all, on the calculator's own page

Relative Risk Calculator

Last reviewed 12 September 2026

Verified that a relative risk is near-invariant when both event counts are scaled while the absolute reduction changes in every case - which is why both belong in a report.

5updates in all, on the calculator's own page

One-Proportion Z-Test Calculator

Last reviewed 12 September 2026

The score test's non-rejection region is exactly the Wilson interval, verified on all 39 interior outcomes at n = 40.

5updates in all, on the calculator's own page

Wilson Score Interval Calculator

Last reviewed 12 September 2026

The identity between the Wilson bounds and the score test is verified on every bound from 1 to 39 successes out of 40 - each returns p = 0.05 to six decimal places.

5updates in all, on the calculator's own page

Midrange Calculator

Last reviewed 12 September 2026

The simulation is reproducible by construction — identical seed, identical table — and the suite verifies that a different seed gives different numbers, so it is genuinely simulating rather than returning a constant.

5updates in all, on the calculator's own page

Exponential Smoothing Calculator

Last reviewed 12 September 2026

Errors are one-step-ahead throughout, so a fitted value never uses the observation it is compared against, and MAE divides by n - 1 because the first period has no forecast.

5updates in all, on the calculator's own page

Multiple Regression Calculator

Last reviewed 12 September 2026

Each VIF is computed by actually regressing that predictor on the others and cross-checked against an independent fit, rather than read off a correlation matrix.

5updates in all, on the calculator's own page

Sign Test Calculator

Last reviewed 12 September 2026

Calibration checked on 4,000 null studies: the realised rejection rate came out at 3.33%, at or below the nominal 5% as an exact test must be.

5updates in all, on the calculator's own page

Normality Test Calculator

Last reviewed 12 September 2026

Power against a mildly skewed alternative was measured, not asserted: a lognormal with skewness 0.72 is detected 9.6% of the time at n = 10 and 100% at n = 1,000.

5updates in all, on the calculator's own page

Conditional Probability Calculator

Last reviewed 12 September 2026

The suite confirms Bayes' theorem and the law of total probability across thousands of randomly generated feasible triples rather than on the worked example alone.

5updates in all, on the calculator's own page

Cpk Calculator

Last reviewed 12 September 2026

Prints the observed defect rate beside the predicted one, because the PPM conversion is a normal tail probability and lives four standard deviations out — exactly where the normality assumption is least testable and least likely to hold, and where a bounded or skewed characteristic breaks it by orders of magnitude.

4updates in all, on the calculator's own page

Equivalence Test Calculator (TOST)

Last reviewed 12 September 2026

Establishes a conditional law the suite found rather than assumed: more noise raises the TOST p-value on all 2,728 cases where the observed difference lies inside the bounds and LOWERS it on all 272 where it lies outside, because there the extra noise weakens evidence against equivalence.

4updates in all, on the calculator's own page

McNemar's Test Calculator

Last reviewed 12 September 2026

Separates the odds ratio for change, b/c, from the marginal shift — two quantities that are routinely conflated, since the first is conditional on having changed at all and the second is what a reader usually means by the effect.

4updates in all, on the calculator's own page

Trimmed Mean Calculator

Last reviewed 12 September 2026

States that trimming is symmetric regardless of where the extremes actually are, so on one-sided data it discards good observations from the clean side — the price of a method that does not have to decide which points are outliers.

4updates in all, on the calculator's own page

Moving Average Calculator

Last reviewed 12 September 2026

Detects crossovers between two windows and states that one cannot precede the turn that caused it — the gap is roughly the difference in the two lags, 75 periods for a 50/200 pair. The suite asserts that no crossover appears before the turn on a series built to turn at a known point.

4updates in all, on the calculator's own page

Bayes' Theorem Calculator

Last reviewed 12 September 2026

Reports likelihood ratios and handles repeated evidence in odds form, while stating that independence is usually false for a repeated test — whatever produced the first false positive is still there for the second — so the two- and three-positive figures are an upper bound rather than an expectation.

4updates in all, on the calculator's own page

Cohen's Kappa Calculator

Last reviewed 12 September 2026

Offers linear and quadratic weighting for ordinal categories and states the condition the usual belief rests on: quadratic >= linear >= unweighted held on all 600 ordinally-structured test tables and was close to a coin flip on 600 unstructured ones, so weighting categories that are not genuinely ordered can LOWER kappa. Also says the Landis-Koch interpretation labels were proposed as personal benchmarks with no empirical basis.

4updates in all, on the calculator's own page

Kendall Tau Calculator

Last reviewed 12 September 2026

States that the normal-approximation p-value assumes no ties and about ten observations, and that with heavy ties the variance formula is the untied one and the p-value comes out optimistic.

4updates in all, on the calculator's own page

Spearman Correlation Calculator

Last reviewed 12 September 2026

Plots the midranks rather than the raw values, because that is what Spearman actually correlates, and states that the t approximation for testing rho is unreliable at small n and ignores the tie pattern entirely.

4updates in all, on the calculator's own page

Bonferroni Correction Calculator

Last reviewed 12 September 2026

Says plainly that m is the number of tests RUN, not the number being reported. Correcting the subset that looked promising controls nothing, because the selection already happened — which is the most common way a correction is applied and still provides no guarantee.

4updates in all, on the calculator's own page

Tukey HSD Calculator

Last reviewed 12 September 2026

Applies the Tukey-Kramer modification automatically when group sizes differ, and refuses to print a single HSD threshold in that case, because with unequal sizes each pair has its own standard error and one quoted threshold applies to no pair. Simultaneous intervals are reported for every comparison.

4updates in all, on the calculator's own page

Frequency Distribution Calculator

Last reviewed 12 September 2026

Uses half-open bins with the maximum included in the last one, because otherwise the largest observation falls outside every bin and the counts sum to n - 1 — the classic off-by-one in a hand-built frequency table. The suite checks that counts sum to n and relative frequencies to 1 across 300 random datasets.

4updates in all, on the calculator's own page

Mean Absolute Deviation Calculator

Last reviewed 12 September 2026

The suite verifies the minimisation property by SWEEPING every candidate centre on a grid rather than by comparing two formulas, confirms the ratio lands on sqrt(2/pi) over 200,000 simulated normal draws, and checks that one wild value multiplies the SD by more than a thousand while moving the median absolute deviation by under 20%.

4updates in all, on the calculator's own page

F-Distribution Calculator

Last reviewed 12 September 2026

Takes any positive real degrees of freedom, so Welch's ANOVA and mixed-model denominators need no rounding, and states plainly that the distribution is not symmetric: F(2,27) at 95% is 3.354 and F(27,2) is 19.459.

4updates in all, on the calculator's own page

t-Distribution Calculator

Last reviewed 12 September 2026

Verified against the standard table values at df = 1, 2, 30 and 120, and against the identity F(1, v) = t(v) squared, which ties it to the F distribution and shows a two-group ANOVA and a two-sided t test are the same test.

4updates in all, on the calculator's own page

Fisher's Exact Test Calculator

Last reviewed 12 September 2026

Prints chi-square and Yates on the same table with the smallest expected count, and returns Fisher's own tea-tasting probability as exactly 1/70 — the design used four cups each way precisely so that only a perfect score would clear 5%.

4updates in all, on the calculator's own page

Statistical Power Calculator

Last reviewed 12 September 2026

Reports Cohen's h for proportions rather than the raw difference, because a proportion's variance depends on the proportion: the same five-point gap is a 2.513 times larger effect at a 2% baseline than at 50%. Verified against a 200,000-trial simulation of the actual t test at four designs, and against the requirement that power at a zero effect equals alpha exactly.

4updates in all, on the calculator's own page

Sum of Squares Calculator

Last reviewed 12 September 2026

Computes the corrected sum from the deviations in a second pass rather than from the textbook shortcut, which subtracts two nearly equal large numbers and loses a digit of precision for every digit they share — the classic route to a negative computed variance.

4updates in all, on the calculator's own page

Residual Calculator

Last reviewed 12 September 2026

Standardised residuals divide by s times the square root of 1 minus leverage rather than by s, because residual variance genuinely falls with leverage; studentised residuals exclude the point from its own variance estimate. Cook's distance is verified against a full delete-and-refit at every point rather than only against the algebraic shortcut.

4updates in all, on the calculator's own page

Power Regression Calculator

Last reviewed 12 September 2026

Shows the log-log R-squared beside the R-squared the curve achieves on the data, and counts rows dropped because x or y was not positive — a stricter requirement than the exponential or logarithmic fits, which each constrain only one variable.

4updates in all, on the calculator's own page

Logarithmic Regression Calculator

Last reviewed 12 September 2026

Rows with x at or below zero are excluded and counted, and the page states plainly that shifting x by a constant to dodge that changes every coefficient by an amount depending entirely on the constant chosen.

4updates in all, on the calculator's own page

Quadratic Regression Calculator

Last reviewed 12 September 2026

Shows the straight-line R-squared beside the quadratic one and explains that a higher figure is arithmetic rather than evidence, since R-squared cannot fall when a term is added. The speed-and-economy preset scores 0.068 as a line and 0.998 as a parabola, which is what a large gap looks like.

4updates in all, on the calculator's own page