Measurement methodology

How the result is produced on your own repository.

The result is based on delivery time, delivery scope, repository artifacts, and quality assessment.

Result model

Input data

  • RREF / EXT scope
  • TREF / EXT time
  • QEXT quality

Session result

WPO

Operational advantage coefficient

5

tasks per session

1–3

task weights

90

minutes of delivery

30

minutes of review

WPO

final result

Comparison model

The same scope. Two work modes.

REF shows the current work model. EXT shows the Nexir workflow.

Shared conditions

The same weights 1–3The same T0 start

REF · current work model

Current work model

The organization developer works in REF mode.

Reference point

REF creates the baseline for comparing the result.

EXT · Nexir workflow

Nexir workflow

The Nexir developer works in EXT mode.

Extended result

EXT shows the result after quality correction.

Comparison artifacts

Conditions

Four measurement rules.

The rules define scope, time, environment, and the measurement basis.

The same scope

The same tasks, weights, and time window.

Repository artifacts

Commits, pull requests, and the scope of changes in the repository.

Controlled environment

An isolated fork and defined session rules.

Work model

The comparison covers delivery modes.

Time model

90 minutes of delivery, 30 minutes of review.

Time T comes from the delivery phase. Correction Q comes from the quality assessment.

Task delivery
Review

Task delivery · 90 min

Quality review · 30 min

Delivery

Task execution and recording changes in the repository.

Artifacts

Commits and pull requests as the basis of the report.

Quality review

The EXT pull request is rated against four criteria.

Time T in the formula refers to the 90 min phase. Q comes from a separate quality assessment.

Result components

Delivery, time, and quality.

Three components form one WPO coefficient.

Delivery scope

Weighted scope of completed tasks.

/ weight × completion degree

Delivery time

Time required to complete the tasks.

/

Quality correction

Quality multiplier for the EXT result.

WPO formula

Operational advantage coefficient.

WPO compares EXT throughput with REF throughput after quality correction.

Operational advantage coefficient

= ( / × ) / ( / )

R = ( × )

where

WPO
The ratio of the extended mode throughput, with quality adjustment, to the reference mode throughput.
R_EXT
Weighted sum of task delivery degrees in the extended mode.
T_EXT
Task delivery time in the extended mode, in minutes.
Q_EXT
Adjustment of the extended mode result after the solution quality assessment. Range: 0.9–1.1.
R_REF
Weighted sum of task delivery degrees in the reference mode.
T_REF
Task delivery time in the reference mode, in minutes.

WPO = 1.0

REF and EXT have similar throughput.

WPO = 2.0

EXT has twice the throughput.

WPO < 1.0

REF has higher throughput in the session.

The notation is consistent with the methodology modal and the PDF report.

Quality correction

Quality corrects the EXT result.

Four criteria on a 1–10 scale form O_EXT and the Q_EXT multiplier.

Review criteria

Q_EXT multiplier

= 0.9 + (( 1) / 9) × 0.2

Correction

Q_EXT
Multiplier adjusting the extended mode result. Range: 0.9–1.1.
O_EXT
Average of four partial scores on a 1–10 scale: delivery completeness, technical correctness, alignment with the architecture, and scope of changes relative to the task.

Multiplier range

The average score forms the Q_EXT multiplier.

The 0.9–1.1 multiplier limits how much quality affects the result.

Artifacts

The repository is the basis of the report.

The session runs on an isolated fork. REF and EXT record work in separate branches.

Repository flow

RepositoryIsolated forkREF / EXTReport

Commits

Change history and commit times for both modes.

Pull requests

Separate change sets for REF and EXT.

Change metrics

Files, lines of code, and time to first artifact.

AI cost

Cost of models used in EXT mode.

Report

From metrics to report.

The result, quality, repository activity, and interpretation are included in one report.

Path

TasksDeliveryArtifactsReviewWPOPDF report

WPO result

WPO coefficient with values substituted.

REF vs EXT

Comparison of modes on the same scope.

Quality review

Review criteria and the Q_EXT quality multiplier.

Artifacts and cost

Commits, pull requests, change metrics, and AI cost.

Status and limits

Result status and edge cases.

Decision report

Preview of the result report.

The sample document shows the layout of the result, metrics, and interpretation.

Session report preview

Wynik przewagi

5/5 zadań

MetrykaREFEXTΔ
Czas pracy90 min65 min-28%

Interpretation

Result status and scope of interpretation.

The result is described together with edge cases and the scope of reading.

Edge cases

REF result equals 0

WPO is not presented as a numeric value.

Zero delivery scope

Average task time is not calculated.

No artifact

Time to first artifact is not calculated.

Partial task

Completion degree is counted on a 0–1 scale.

Interpretation limits

Reference point

The result supports the decision on the next stage.

Context dependence

Metrics depend on tasks and architecture.

Decision scope

The report does not replace a deployment assessment.

See the methodology on a sample report.

One session, one WPO result, and a report with a full description of how it is calculated.