Windows analysis: nothing in the record changed and the count did

Type which party drove each day, then declare a window. One six day record counts 0 to 1 windows on the strict reading and 0 to 3 on the other.

Live result

Two readings of one record count 1 and 3 complete windows on the grid you declared.

Window length2 days
Same day, across the family0 to 1
Within window, across the family0 to 3

Moving the grid start or the window length moves both counts.

Try a scenario

The declared grid

Window length2 days

The chart and both counts move with this one dominant control.

Record start

Civil date from 1970 through 2100. No time zone conversion is made.

Grid start

Civil date from 1970 through 2100. No time zone conversion is made.

Shortest family lengthday

Longest family lengthdays

Every whole length and every offset in this inclusive family is checked.

One scope and information cut

Every daily entry must describe the same scope and information cut before a count is shown.

1: A complete window counts only when at least one typed day says both parties on that same day.
3: A complete window counts when the first party appears on at least one day and the second appears on at least one day in that window.

The daily driver record

6 / 24

Day 1

Day 2

Day 3

Day 4

Day 5

Day 6

Up to 24 daily entries.

Declared grid coverage

Complete windows: 3
Covered days: 6
Uncovered edge days: 0
Grid offset: 0

Exact family range

Strict minimum and maximum: 0 (length 4, offset 3) to 1 (length 1, offset 0)
Loose minimum and maximum: 0 (length 4, offset 3) to 3 (length 2, offset 0)
Length and offset combinations: 21

Exact family sweep

At each length, the plotted point uses the typed grid start. The range readout checks every reachable offset.

First checked length change2 days: 1, 3

It counts complete windows under two declared readings over the typed daily trace, without establishing concurrency, deciding cause, allocating delay, awarding days, or deciding entitlement.

Here is the maths

The formula, in words

Reduce the civil date difference between the grid start and record start modulo the window length. Starting at that offset, count only equal windows that fit wholly inside the record. For the family range, repeat this for every whole length and every offset from zero through length minus one.

O=\operatorname{positiveMod}(G-R,L)

where

O

is the offset,

G

the grid start,

R

the record start, and

L

the window length.

C=\sum_{L=a}^{b}L

where

C

is the number of combinations,

a

the shortest window length, and

b

the longest window length.

A worked example

The six typed days are A, B, neither, both, A, B. The grid starts with the record and the window length is 2.
The three windows are A with B, neither with both, and A with B. The strict count is 1 and the loose count is 3.
Across lengths 1 through 6, the number of length and offset combinations is $1 + 2 + 3 + 4 + 5 + 6 = 21$ . Strict ranges from 0 to 1 and loose ranges from 0 to 3.

What this does not model

The daily entries are the whole observation domain. Nothing is inferred between them or beyond them. The calculation compares neither methods nor records and does not turn either count into responsibility or entitlement.

Check it yourself

At length 1 both readings equal the number of days typed both. At the whole record length an aligned grid has one complete window, while a shifted grid has none.

Source basis

Published scheduling mathematics and delay analysis guidance on declared analysis intervals. The source class supplies structural context only. The predicates and the typed record on this page are declared here, and no source wording is reproduced.

The whole record prints. Your browser can save it as a PDF instead of printing it.

This instrument does not decide whether concurrency exists, allocate delay, apply a named test, award days, assign responsibility, or decide entitlement. Requests to decide concurrency or allocate delay are refused.

Window grid count record