Tool

Astrological Prediction Journal

The Astrological Prediction Journal is a testing tool that lets you record the predictions you read before they come due, and score them coolly once their period is over.

Judging the accuracy of predictions from memory is misleading: the mind remembers the predictions that came true and quietly forgets the ones that did not. This is called selective memory, and it is one of the main sources of the feeling that “the readings always fit”. This journal moves the assessment out of memory and into a record: you save a prediction on the day you read it, and you come back to score it once the period it covers has run out. The method is not specific to astrology; it is an honest testing practice that works for every kind of forecast, from economics to sport.

Tool

Frequently asked questions

Why do I need to record a prediction before it comes true?

Assessing afterwards leaves room for the mind to remember the event in a way that fits the prediction; this is called retrospective interpretation. A prediction written down in advance with a defined period removes that flexibility and makes the assessment fair.

How should I score vague predictions?

The most honest route is to mark them as “could not be scored”; giving half a point to readings that partly fit quietly inflates the measurement. Answering the question “what will I count as fulfilment?” in writing at the moment of recording shrinks the problem from the outset.

Can I use this journal for predictions outside astrology?

Yes; the method does not distinguish between sources. You can test the accuracy of economic commentators, sports tipsters and your own forecasts with the same routine; the results are usually instructive in every field.

What does it mean if my accuracy rate comes out high?

First look at how specific the predictions were: general statements are expected to hold. Consistent accuracy above the level of chance on specific, pre-recorded predictions, however, is genuinely interesting and worth testing under stricter conditions.

Method and formula

A record is made with three pieces of information: the text of the prediction, the period it covers, and how clear you consider it to be. Where clarity is low, the tool encourages you to make the prediction concrete; you are asked to write down in advance which outcome you will count as fulfilment of a phrase such as “an important change may take place”. When the period ends, the journal reminds you of the record and asks you to give it a fulfilment score. Over time the accumulated entries form an accuracy summary: how many predictions held, how many did not, and how many were too vague to score. Your records are stored in your browser; they are not sent to a server.

How to read the result

In the summary table, the most instructive column is often the “could not be scored” column: if a prediction cannot be falsified by any outcome, its accuracy cannot be measured either — and that says more about the structure of the wording than about the prediction. When reading the accuracy rate, do not forget the base rate: the prediction “there may be some tension at work this month” is likely to hold without any astrology at all. A fair comparison asks whether the prediction was more accurate than chance and ordinary life probabilities would lead you to expect; keeping records makes that question answerable for the first time.

Limits

  • The fulfilment score is a subjective judgement; two people may score the same event differently.
  • An accuracy rate built on a small number of records is statistically unreliable; a meaningful pattern needs time and a larger number of entries.
  • The journal cannot calculate the base rate of a prediction — its probability of coming true without astrology; that assessment is left to you.
  • If your browser data is cleared, the records are deleted; bear this in mind for long-term tracking.

Calculator and tool methodology · Nothing you type is sent to a server; results stay in your browser.

First published: 2026-08-12Last reviewed: 2026-08-12Editorial status: working editionReport an error