Every number on this site is cast by the same engine. The compass and the calendar will never disagree with each other — that is the whole point of building the tools instead of borrowing them. Six of the fourteen cast a chart; the other eight exist to measure the reading, and were built after an audit of our own reports found that we were not measuring it.
Six instruments that take a date, a name or four letters and return the chart itself. Every one of them loads the same reduction engine, so the compass and the calendar can never disagree.
Twenty-four yes/no questions, about two minutes, confidence 0.55 — printed on the page rather than hidden. It produces four letters and hands them straight to the compass. A tied axis is reported as held, not guessed.
Eight concentric rings — western zodiac, 干支, personal year, destiny, soul urge, birth day, life path — with the Jungian function stack burning at the centre. The single view that shows how the frameworks agree.
Your Personal Year and Personal Month on the birthday-to-birthday clock, plus the full twelve-month arc. Numerology's most usable output: not who you are, but when the door opens.
Real ephemeris math — ten bodies, the ascendant, the midheaven, whole-sign houses and the major aspects. Each figure carries its own confidence limit, stated rather than hidden.
Universal Day frequency across a month. Which combinations repeat, which fire only once, how dense the master days are, and what backdrop the 28th happens to wear.
The first instrument that takes two people. Two function stacks on one sheet — which axes they share, where the positions invert, and which of the misses are structural rather than personal. No compatibility score; the reason is stated on the page.
Three instruments that measure the reading rather than the person. One asks how many people share a value, one asks whether the same person gets the same answer twice, and one asks how two charts connect and how often strangers connect the same way.
Every coordinate comes back with the share of people who hold it. The engine could always compute that figure and no report printed it — which is how a sentence about three thrones survived for months while being true of one chart in four.
Twelve questions, three per axis, answered twice a fortnight apart. Before the second sitting it names the axis most likely to flip, then shows whether the prediction held. Nothing is sent anywhere; a seven-character code carries your answers.
Two birthdays, three layers, and every wire that runs between them — each printed with the share of random pairs who have the same thing. Strangers connect 97% of the time in the sky and 99% in the numbers, so a wire existing says nothing. How many, and whether they run straight across or diagonally, is the part that does.
Five instruments for the harder half: reading a function from behaviour rather than from a label. These have no confidence figure, because no survival rate has been measured for them.
Four axes, five items each, and one item on every axis carries a weight of zero. It is there to catch a reader who agrees with everything. A settled axis is reported as settled; a dead band is reported as undecided rather than rounded.
Positions five through eight are the attitude-flip of one through four, and they come out of the same engine table the compass reads. What the grip is, what it is not, and why the bottom of a stack is derived rather than observed.
Ten scenes, each one a piece of behaviour with the function left off. Two of the ten cannot be decided from what is given, and the instrument tells you so if you decide anyway. Wrong answers are classified by the kind of mistake, not just marked wrong.
The third function is an investment; the inferior is bounded exposure, not a muscle. Every inferior exercise here is capped — one, ten minutes, once a month — because an unbounded one stops being an exercise and becomes a trigger.
Forty scenes, twenty a round, built on the eight function pairs that never share a top four — the eight people confuse. Name the function, then state a confidence. It returns a calibration curve, a confusion matrix and a Brier score, because being right and knowing when you are right are different skills.