Model Drift Score Calculator

Work out model drift score instantly with clear inputs, formula shown and shareable results.

About Model Drift Score Calculator

Prediction drift is monitored by comparing the current output distribution against a stored baseline. Expressing the mean shift in baseline standard deviations makes the comparison scale-free and directly usable as an alert threshold: two standard deviations flags roughly the most extreme 5 percent of normal variation, three flags 0.3 percent. This catches upstream data changes before labels arrive to reveal accuracy loss.

Formula

Standardised drift

z = (current mean - baseline mean) / baseline standard deviation; alert when |z| >= threshold

Worked example

With the default values below, Model Drift Score Calculator returns Drift z score of 1.000. Change any input above and every figure here is recalculated in your browser.

Inputs used

Baseline prediction mean: 0.32
Current prediction mean: 0.41
Baseline standard deviationminimum 0: 0.09
Alert threshold in standard deviationsaccepts 0.5–6: 2

Results

Drift z score: 1.000
Absolute shift in prediction mean: 0.0900
Relative shift: 28.13%
Alert status: Within normal variation

Step by step

1Values used
2Standardised drift
3Drift z score= 1.000
4Absolute shift in prediction mean= 0.0900
5Relative shift= 28.13
6Alert status= Within normal variation

Assumptions and limits

The result is derived only from the 4 inputs listed above. Anything not among them is not part of the calculation.
In particular, this calculator does not take inflation, tax, fees or charges or compounding frequency into account.
Figures are rounded for display to 4 decimal places, so totals built from several rounded parts can differ slightly from the unrounded arithmetic.

How these calculators are built and tested is described on the Methodology page. If a figure looks wrong, please report it.

Frequently Asked Questions

Why monitor predictions rather than accuracy?

Labels usually arrive days or weeks later. The prediction distribution is available immediately, so it is the earliest available signal that inputs have changed.

Is prediction drift always a problem?

No. It can reflect a genuine change in the population, such as a seasonal shift. Drift is a trigger to investigate, not proof that the model has degraded.