Kea & Rock and the Certain Sky
for grown-ups · Weather · Expedition One
The big idea
Your child has been setting the Sure-o-Meter since their first expedition. This episode is about what the meter actually says. Each setting is a promise about how often you are right: "really sure" promises nearly always, "pretty sure" promises most times, "just guessing" promises nothing and so can never break its word. You cannot check a promise against a feeling or a memory — only against a record. Kea, who can "always tell" tomorrow's weather, meets a season of Kea's own calls, kept in stone.
It is also the first expedition that runs across real days: forecast tonight, meet the sky tomorrow, come back and face the tally. The waiting is part of the lesson.
What happens (spoilers)
Kea calls rain for tomorrow — really sure, on the authority of feathers. Rock recalls the last "I can always tell": the rain came six days late and the tent drowned. ("'Tomorrow' is not 'eventually'.") A warm-up first: guess the stones in the fire ring, set the meter on the guess, then count — Kea's "two hundred, really sure" meets sixty-three, and the meter's real meaning lands in miniature. Your child claims (Kea can always tell / nobody can ever tell / you can tell a bit / not sure), tags a where-from, locks the meter, and makes their own first forecast: rain, no rain, or an honest "no call".
Then the story stops and asks them to come back when the sky has answered. (It trusts them — there is no clock lock, and an impatient child who returns too soon is gently sent back once, then allowed through.) Next day: before Rock opens the record, "what did you call yesterday?" — memory, then the record, which do not always agree. The child reports what the sky did; the tally begins, theirs and Kea's side by side. Signs are consulted for the second call: the southerly, the low-flying insects, the puddle (which reflects the whole sky and is very certain it contains it). Two signs point different ways — they usually do.
Day three brings the reveal: Rock has tallied a whole season of Kea's rain calls. Ten "really sure"s. The child locks a guess, then counts the ticks themselves: four. Kea's rescues arrive on schedule ("the sky changed its mind — all six times") and the child diagnoses the real problem: the needle is pinned. Wing-Flash: a "sure" is a promise, and a record is how a promise stays honest. Rock adds the sky's own habit — rain three days in ten — so four in ten is a little real skill wearing a badly broken promise. The child reads their own small tally ("small records whisper; they do not shout"), revisits their claim, re-sets the meter, and makes one last call that the story deliberately never checks: the story continues past the page, and Rock will be marking it.
The high route adds a signs duel (felt versus seen, and why the stronger where-from still is not certain) and the promise table, where the child shelves a row with four ticks in ten where it honestly belongs.
The thinking your child practiced
- The meter is a promise, not a mood. Each setting names how often it should come true.
- Promises are checked against records, never memories. Memory rounds toward how we feel now; the record does not. (The "what did you call yesterday?" beat plants a seed a later expedition will grow.)
- Signs add up, but never to certain. Two signs beat one; neither signs anything.
- An honest blank is an answer. Declining to call, said truthfully, is a kept promise — and it still earns its row in the tally.
- A pinned needle measures nothing. Kea's problem was never the forecasting; it was a needle that only had one setting.
- Small records whisper. Two days of tally is a hum, not a verdict; the promise sharpens as the rows grow.
Words your child may come home saying
| Word | What it means |
|---|---|
| a promise | what a meter setting says about how often you'll be right |
| Rock's Tally | the record of calls: the meter word, then a tick or a cross |
| an honest blank | a "no call", said truthfully — it gets a row too |
| a sign | a clue about tomorrow: a wind, low insects, a smug puddle |
| pinned | a needle stuck on one setting, measuring nothing |
| a rescue | a patch added to a theory so it survives a bad result ("the sky changed its mind") |
About the tally in the app, honestly
The child's checked calls appear as Rock's Tally in the Wonder Log — each row with its meter word and a tick or a cross, plus one line per setting ("really sure: kept its promise 1 of 2 times"). It grows across expeditions. Like everything else, it lives on your device only and travels in "pack up". The app never asks where you live and never checks the real weather: your child reports what the sky did, because your child is the observer. That is the point.
About the numbers, honestly
Forecast calibration is a real discipline. Professional forecasters are scored not on being right but on their stated confidence matching their hit rate — a "70% chance of rain" forecaster should be wrong about rain three times in ten. Kea's four-in-ten against a sky that rains three days in ten is a small real edge and a badly broken "nearly always"; that arithmetic is the episode's spine, and it is kept honest on stage. Children meet over-confidence daily; the meter gives it a name, and the tally gives it a mirror.
The home experiment
The fridge tally: a strip of paper. Every night, everyone in the family calls tomorrow — rain or no rain — and says the meter word out loud. Every morning, a tick or a cross beside each call. At the end of the week, don't ask who was right most: ask whose needle kept its promises. A "just guessing" that lands is a quiet win; a "really sure" that misses owes the fridge an apology. No materials beyond paper; the sky is outside everyone's window.
Talk about it afterwards
- "Kea was right four times in ten. Is Kea good at forecasting, or bad? What was actually broken?"
- "When would you honestly say 'really sure' about tomorrow? Is there anything you're that sure of?"
- "What did the record say that memory got wrong?"
For teachers
Nature of Science: prediction under uncertainty; confidence versus evidence; calibration against a record; why small samples "whisper". Statistics: tallying, base rates in child-sized form (the sky's own habit), hit rates as fractions of ten. The episode's forecast scores are invented; the reasoning — and the honesty about what four-in-ten does and does not show — is faithful to how forecast skill is really judged.
Every expedition ends the same way for a grown-up: a home experiment, and something to talk about. Nothing here is homework. If your child wants to argue with the card, that is the point.
