CALC-02 · THE FULL LOOK-BACK
Look back over 30 days The 30-Day Look-Back · for 2 to 8 people
Put four weeks side by side: who did what, last-minute rescues, how much each person is carrying, and whether each job has an owner. You get the running imbalance, what shifted, what didn't, and one next step.
A place to start a conversation, not a verdict. It reads the setup, never a person. Only have a rough "then and now"? Try the quick look-back.
More about this look-back
One week can be a fluke: someone was ill, a deadline landed, a visit took over. Thirty days is long enough to see whether what you changed is changing anything. It is the look-back that week four of the 6-week program asks for: who did what, how many last-minute rescues, and the running imbalance.
CALC-02 puts Pillar I (“See the whole load”) and Pillar II (“Fix the setup, not the person”) to work over a month. In you: you see whether what you tried is helping. Between you and others: one calm picture of the month, never a score on anyone.
Set up
Who's looking back, and from when?
Or load a made-up example to see how it works:
This device keeps some of your own entries from the last 30 days: .
The entries
The four weeks
Fill in what you tracked. Blank is fine: each part is read from the weeks that have it. Week 1 is the first 10 days, so the four weeks cover all 30.
- Hours: each person's hours that week, from Who did what (WP-01). Count planning and remembering too.
- Load: each person's own score from How much are you carrying? (WP-02), 0 to 1, higher is heavier. Everyone fills in their own.
- Rescues: times that week a job was caught at the last minute, or slipped. For the whole household, not by person.
- Jobs: the regular jobs on your One owner per job (WP-03) list, and how many have an owner.
Who did what, added up
The running imbalance
Each person's hours over or under their share of everyone's hours, added up week by week. A gap that keeps growing in the same direction is the setup leaning, not a bad week.
Add hours to at least one week to see it.
Output
The read
Fill in at least one part near the start and near the end to see what shifted.
What shifted
What didn't shift
Week by week
| Week | Balance | Rescues | Avg load | Ownership | Setup |
|---|
Start = the average of weeks 1 and 2 (the ones filled in). End = the average of weeks 3 and 4. Setup is CALC-01's setup score, when balance, ownership and everyone's load are all there.
Just you · optional
Your own days
If you save Today's Weather check-ins, or keep a daily check-in draft, on this device, this part reads them for the same 30 days. It's about how the days felt, so it sits beside the read and never changes it.
Nothing from this device has been read. Press “Fill in from this device” above, if it shows.
Off by default. When it's on, your names and the four weeks are stored in this browser only, never sent anywhere. On a shared computer, everyone who uses this browser could see them. Turning it off, or pressing Erase, deletes them. Reading the other worksheets above never changes them.
Reference
How it's worked out
Start and end
Each part is read twice. Start is the average of weeks 1 and 2, using whichever of them you filled in. End is the average of weeks 3 and 4, the same way. A part needs at least one week at each end, or it isn't read. Averaging two weeks keeps one odd week from deciding the read.
| Part | One week's value | Counts as shifted when | Still worth a look at the end when |
|---|---|---|---|
| How the hours split | CALC-01's balance, 1 − moved ÷ most, 0 to 1. Higher is more even. | it moves by 0.10 or more | under 0.70 |
| Last-minute rescues | the count for the week | it moves by 1 a week or more | 2 or more a week |
| How much each person is carrying | each person's WP-02 score, and their average (only when everyone's own score is there) | the average moves by 0.10 or more (each person's own line uses the same 0.10) | anyone at 0.60 or more (a high load on WP-02) |
| Jobs with an owner | jobs with an owner ÷ jobs on the list | it moves by 0.10 or more | under 0.70 |
The cut-offs are round numbers chosen to be easy to read, not research findings. A move of 0.09 and a move of 0.10 are really the same month.
The running imbalance
gap(person, after week k) = Σ over weeks 1..k of ( their hours − their share × everyone's hours ) hours to change hands = ½ × Σ |gap| over everyone
Their share is an even share (1 ÷ people), or the split you agreed on. Weeks with no hours are skipped, so a blank week never counts as an even one. With two people, the hours to change hands is simply how far one person is over.
Made-up example: Sam and Alex log 14 and 6 hours, then 13 and 7, then 11 and 9, then 10 and 10. Sam's running gap is +4, +7, +8, +8 hours. Over the month, 8 hours would have to change hands to even it out, and all of it built up in the first three weeks. The last week was even, which is the change working; the 8 hours already done stay done.
The one next step
- If a part moved the wrong way, the step goes to it. If two did, the one that moved furthest past its cut-off.
- Otherwise, the first part still worth a look at the end, in this order: jobs without an owner, uneven hours, rescues, a high load. Ownership comes first because it is usually the easiest thing to change, and it often settles the other three.
- If nothing needs a look, the step is to keep the same rhythm and look back again in 30 days.
Reading this device
“Fill in from this device” reads, without changing them: the long fair-split calculator's kept trend log, the Workpaper Suite draft and single worksheet drafts (Who did what, How much are you carrying?, the daily check-in), and the jobs your weekly check-in card was told slipped. It only fills boxes that are still empty, and says what it used. A “week beginning” date is placed by its middle day. Where two entries fall in the same week, the later one is used. Rescues taken from the weekly check-in are the jobs it was told slipped, which is close to, but not the same as, every rescue: check the number.
What this look-back will never use. Tone of voice, messages, how fast someone replies, or anything about a person that they didn't help enter themselves. A load score is never filled in for someone else.
What it is not
- Not a verdict on who is the better partner, parent or housemate.
- Not a tested clinical or statistical tool. It predicts nothing; it describes what you typed in.
- Not a reason to decide a relationship's future. “The look-back says…” is a misuse. Its job is to give you a calm place to start.