CALC-01 · THE FULL CALCULATOR
Fair split calculator: is the setup working for everyone? For 2 to 8 people
This tells you, in plain words, whether the way you share the work right now can keep going for everyone. It never scores a person or the relationship.
A place to start a conversation, not a verdict. A low score doesn't blame anyone, and a high score doesn't mean anyone should stop speaking up. Only have totals? Try the quick read.
More about this calculator
It works for any group that shares a load: partners, families, roommates, co-carers, a small team at work. The idea is borrowed from accounting: can the setup carry its own weight?
CALC-01 puts Pillar I (“See the whole load”) and Pillar II (“Fix the setup, not the person”) to work. Step Zero uses Pillar III, Read your state first, and the retuning count uses Pillar IV, Tune how you send and receive. In you: you see the whole load you carry, and check your state before you read the number. Between you and others: you get one fair read of the setup, never a score on anyone.
The math comes from Chapter II. To follow it week by week, use REPORT-01.
Step zero
Check your state
Before you read a number about how you share the load, notice how you feel right now. The same number lands differently when you're calm than when you're tired or wound up. This is a check on yourself only. It never changes the score; it changes whether today is the day to act on it.
Set up
Who's sharing the load?
List the people who share this set of jobs and can each fill in their own WP-02 (How much are you carrying?). Two to eight people. Names stay on this page.
Or load a worked example to see how it works:
You keep a Workpaper Suite draft on this device:
Two scores, two questions. The setup score uses balance, ownership and stress, and asks: can this week's load last as it's set up? The overall score adds retuning and asks a wider question: is the whole system working, including repair after rough moments? They can disagree, and the gap is useful.
Input 1 of 4 · Balance
How the hours split
Workload balance
The week's hours
From WP-01, Who did whatEnter this week's rough hours for each job, for each person. Rough is fine: the split is what matters. Count the invisible work too (planning, remembering, keeping track).
Swipe the table sideways to see everyone.
Each person's agreed share of the home load, in percent. They must add up to 100. A read against an agreed split can't be compared with a standard read, so the summary says which one you used.
In plain words: balance asks how many of the week's hours would have to change hands to reach an even split (or your agreed split), compared with the most that could ever need to move, which is when one person does it all. None to move scores 1.00. One person doing everything scores 0.00. With two people this is exactly the book's 1 − |pctA − pctB| ÷ 100. It doesn't matter who is carrying more, only that the load isn't shared the way you meant it to be.
Input 2 of 4 · Ownership
Does each job have a name?
Ownership clarity
The job list
From WP-03, One owner per jobFor each regular job, pick its owner: the one person who does it and sees it through. A helper (someone who pitches in or notices if it slips) is optional. A job counts as owned when it has an owner.
| Job | Owner | Helper or backup (optional) | Owned? | Remove |
|---|
Don't raise this one by shortening the list. Deleting the jobs nobody owns raises the number and changes nothing at home. A real job with no owner stays on the list. Bringing that row to light is the whole point of this score.
"Everyone together" counts as a name, as the book allows. But a job that everyone is accountable for is often a job nobody notices (Pillar V, Notice the quiet incentives). The read flags those.
Input 3 of 4 · Stress
How much each person is carrying
Average load (AS in the formulas)
Everyone's load score
From WP-02, How much are you carrying? (higher = heavier)Each person fills in their own card, about themselves, for the last 24 to 48 hours. 0 = not at all, 4 = very true. Guessing for someone else turns a reflection tool into a way of watching them, so the read waits until every card is complete.
Stress is the only input that counts against the score, as 1 − stress. A way of sharing the work that only holds up when everyone is rested isn't one that can last. This is a self-check, not a health measure.
Input 4 of 4 · Retuning · overall score only
Repair after rough moments
Retuning (RF in the formulas)
Counting repairs
From WP-09, Say it so it landsCount the week's rough moments (when something landed badly enough to notice). Then count how many you retuned: talked it through calmly before answering, using the WP-09 steps (fact, feeling, ask). This measures what you did about friction, not how much there was.
Output
The read
Fill in the inputs above to see a read.
The math, step by step
Where it is strong, and where it is short
Each bar shows how much of that part is working. The biggest gap, not the total, is the place to start.
What would help most
- Available once there's a read.
Optional: push scores toward the ends of the scale
This runs the same total through an S-shaped curve centered at 0.5: 1 / (1 + e^(−6 × (raw − 0.5))). Middling reads move further from the middle. It doesn't make the score more accurate, and reads taken with it can't be compared with reads taken without it. It's off by default.
Advanced: change the weights
These weights are a judgment, not a research finding. They say an uneven load matters a little more than unclear ownership, and both matter more than everyday stress. If you change them, new reads can't be compared with old ones.
| Part | Setup score | Overall score |
|---|---|---|
| Balance | ||
| Ownership | ||
| 1 − stress | ||
| Retuning | n/a |
Over time
Trend log
One read tells you about one week. The trend tells you whether anything is changing. Log each week's read here.
| Week of | Balance | Ownership | Stress | Retune | Setup | Change | Overall | Remove |
|---|
Off by default. When it's on, your entries 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.
Reference
How it's scored
The two formulas
setup = (balance × 0.40) + (ownership × 0.35) + ((1 − stress) × 0.25) overall = (balance × 0.35) + (ownership × 0.30) + ((1 − stress) × 0.20) + (retuning × 0.15)
Both are simple weighted totals, not fitted models. Each set of weights adds up to 1.00 and every input runs from 0 to 1, so each score runs from 0 to 1 too. No weight was estimated from real-world data. The first is the setup score and the second the overall score. Neither is a probability, and neither predicts anything.
The four inputs
| Input | How it's worked out | From |
|---|---|---|
| Balance | 1 − moved ÷ most, where moved = ½ × Σ |share − target share| and most = 1 − the smallest target share. The target is an even share (1 ÷ people) unless you choose an agreed split. With two people and an even target, it equals 1 − |pctA − pctB| ÷ 100. | WP-01 |
| Ownership | jobs with an owner ÷ all jobs listed. Measures how clear ownership is, not how fair it is. | WP-03 |
| Stress | Each person: their five 0–4 answers added up ÷ 20. Then the average across everyone. Counted as 1 − stress. | WP-02 |
| Retuning | moments retuned ÷ rough moments, capped at 1.00. No value when there were no rough moments. Overall score only. | WP-09 |
Balance for more than two people, worked through
Three roommates log 12, 6 and 2 hours: 20 hours in all. An even share is a third each, about 6.67 hours. Jordan is 5.33 hours over, Priya is 0.67 hours under, Lee is 4.67 hours under. So 5.33 hours would have to change hands to make it even: that's moved = 5.33 ÷ 20 = 0.267. The most that could ever need to move, if one person did all 20 hours, is 13.33 hours, or most = 1 − 1/3 = 0.667. Balance = 1 − 0.267 ÷ 0.667 = 1 − 0.40 = 0.60.
With 6 of 8 jobs owned (0.75) and batteries of 8, 12 and 4 points (0.40, 0.60 and 0.20, average 0.40):
(0.60 × 0.40) + (0.75 × 0.35) + ((1 − 0.40) × 0.25) = 0.24 + 0.2625 + 0.15 = 0.6525 → 0.65, needs a look
A family of five logs 10, 8, 6, 4 and 2 hours (30 in all; an even share is 6 each). The gaps are +4, +2, 0, −2 and −4 hours, so 6 hours would have to move: moved = 0.20. most = 1 − 1/5 = 0.80. Balance = 1 − 0.20 ÷ 0.80 = 0.75. With 9 of 12 jobs owned (0.75) and an average load of 0.40, the read is 0.30 + 0.2625 + 0.15 = 0.7125 → 0.71, holding.
An agreed split. If two people agree on 60/40 (say one works longer hours outside the home) and log 6 and 4 hours, nothing needs to move, so balance is 1.00. Against an even split the same week would score 0.80. Use the agreed split only if everyone really agreed to it, and say so when you share the read.
Why these weights
| Part | Setup score | Overall score | Reasoning |
|---|---|---|---|
| Balance | 0.40 | 0.35 | Highest, because an unshared load keeps building up whether anyone notices or not. |
| Ownership | 0.35 | 0.30 | Close behind, because work with no owner falls to whoever notices first, every time. That turns a fixable imbalance into a repeating one. |
| 1 − stress | 0.25 | 0.20 | Lowest of the three, because it changes most from week to week. One week of bad sleep shouldn't dominate the read. |
| Retuning | — | 0.15 | Overall score only, and small, because it's a simple self-count that's easy to overstate. |
The bands
| Score | Reading | First move |
|---|---|---|
| 0.70 or more | Working well: the setup is carrying its own weight. | Don't change the setup. Keep the same rhythm of check-ins. |
| 0.40 up to 0.70 | Needs a look. | Start with the biggest shortfall. It's often ownership, but check rather than assume. |
| below 0.40 | Needs a rethink, together. | Rework the whole job list together, not another one-off patch. This is about the setup, not about anyone. |
The band is read from the score as shown, rounded to two decimals, so the number and the band always agree. The edges are round numbers chosen to be easy to read: a 0.69 and a 0.70 are really the same week.
Reading the two scores together
- Both high: the load is shared and owned, and rough moments get repaired. Keep going.
- Setup score high, overall score lower (by 0.08 or more): the setup is fine, but repair isn't happening. Work on WP-09, not the job list.
- Overall score high, setup score lower (by 0.08 or more): you repair well, but the setup keeps creating the friction. Work on the job list.
- Both below 0.40: the setup and the repair are both struggling. Another worksheet may not be what helps most; outside support, like a counselor or a trusted mediator, is worth considering.
Edge cases, defined
- No hours logged: balance has no value, rather than 1.00. An empty week isn't an even week.
- No jobs listed: ownership has no value, rather than 0.00.
- Anyone's battery incomplete: no read at all. A score is never worked out from some people's answers standing in for everyone's.
- No rough moments: retuning has no value. The overall score uses the other three weights divided by 0.85, so they still add up to 1.00, and the read says so.
- More retunes than rough moments: retuning stops at 1.00.
- Agreed shares that don't add up to 100%: no balance until they do.
Drift rule
If the overall score drops by more than 0.10 three weeks in a row, it's time for a planned review of the job list, not an emergency conversation. The trend log watches for this.
What this calculator will never use. Tone of voice, recordings, the mood of messages, how fast someone replies, or anything about a person that they didn't help enter themselves.
What the number is not
- Not a compatibility score, and not a verdict on who's the better partner, parent, roommate or coworker.
- Not a tested clinical or statistical tool. It predicts nothing.
- Not a reason to decide a relationship's future. It looks at how the load is set up, one week at a time.
- Not an argument-ender. "Well, the number says…" is a misuse. Its only job is to give people a calm place to start.