How to test a tarot reading: practical guide, examples and limitations

To test a tarot reading after 30 days, preserve the original record before events unfold. Write the question, cards and positions, the exact interpretation, prediction and advice as separate statements, two alternative explanations, evidence that would support or contradict each one, your own action, and a fixed review date. A month later, compare that document with observable events rather than with your current memory of the reading. Some claims may be confirmed, some contradicted, and some too vague to test. The purpose is not to prove that tarot is infallible. It is to learn whether the reading clarified a decision and whether its claims survived contact with reality.

I began taking reading journals more seriously after an embarrassingly convenient review. My note said, “A conversation will clarify the situation.” Several conversations happened over the following weeks, and I selected the one that fitted the sentence best. At the time I called it a confirmation. Later I reread the entry and saw the problem: I had not written who the conversation was with, what “clarity” meant, or when it was expected. The statement was broad enough for almost any event to count.

Why memory is not an audit trail

Once an outcome is known, the earlier reading looks different. We notice matching details more readily, combine unrelated episodes, forget contradictions and give a symbol a meaning that was never written down. This hindsight adjustment is not always deliberate. Memory naturally tries to turn scattered events into a coherent story.

A useful review therefore evaluates statements recorded in advance, not the impression that “it kind of came true”. A disciplined record allows three honest outcomes:

• confirmed;
• contradicted or not confirmed;
• impossible to test because the wording or evidence was inadequate.

Partial confirmation is also legitimate when the expected parts were defined beforehand.

What to record before events happen

Store the original text somewhere it cannot be quietly rewritten: a note with version history, an email to yourself, a saved situation or a screenshot. You can add later observations below it, but keep the initial wording intact.

Separate prediction from advice

This distinction prevents many false confirmations. A prediction describes an event expected to occur. Advice describes an action you may take. If the advice changes the situation, the result should not automatically be counted as an independent prediction coming true.

Advice can be valuable even when a prediction is inaccurate. A coincidence with a prediction can also occur while the underlying decision remains poor.

The 30-day review protocol

1. Record the question and starting facts. Include only information available at the time.
2. Translate the interpretation into ordinary language. Replace “energy is shifting” with an expected observable event.
3. Mark prediction, advice and symbolic theme separately. Do not let them merge.
4. Write two alternatives. What else could explain the same cards and facts?
5. Define supporting and contradicting evidence. Use events that can be observed.
6. Specify your own action. You need this to see how you influenced the result.
7. Set a precise date and do not move it retrospectively. A poor time frame is a result about the question itself.
8. Keep a short event log until that date. Record facts without fresh interpretation.
9. On review day, read the original note first. Then open the event log.
10. Assign a status to every statement. Confirmed, partial, contradicted, untestable, or changed by your action.
11. Record practical value. What did the reading help you decide, ask, do or notice?
12. Do not draw a new card to rescue the old one. A new spread belongs to a genuinely new question and a new starting point.

Two alternatives protect against the convenient story

One interpretation quickly becomes a narrative we want to prove. Two versions force us to name the evidence that would distinguish them.

Suppose the Hanged Man appears as the obstacle.

• Version A: the process depends on an external approval that cannot be accelerated directly.
• Version B: the user is maintaining the pause by waiting for perfect certainty and not asking for a next step.

The distinguishing fact is practical. Is there a named approver and a defined stage, or is the uncertainty continuing because no clear request has been made? If a request is sent and the organisation identifies a procedure and date, Version A gains support. If the request reveals that the decision could have been made by the user, Version B becomes more plausible.

Both may be partly true. The review should mark which evidence supported each one rather than choosing the more elegant story afterwards.

Worked example: an internal project waiting for a decision

This is a composite case. A small-team manager is waiting for approval of an internal project. Discussion has been postponed twice, and he wants to know whether the situation will move during the next month.

Starting facts recorded on 8 August:

• the proposal was sent two weeks earlier;
• the financial estimate was preliminary;
• the meeting had been postponed twice;
• one senior colleague expressed verbal support;
• the final decision owner had not been identified;
• the team had not started work;
• the manager could send a short summary but could not approve the budget.

Question: “How will the project discussion develop by 7 September, and what can I do to obtain a clear decision?”

A hypothetical spread:

1. The Hanged Man — current dynamic;
2. Justice — what will shape the decision;
3. Eight of Wands — possible event;
4. Four of Cups — risk;
5. The Magician — the user’s action;
6. Two of Swords — alternative scenario;
7. Three of Pentacles — result criterion.

Original interpretation saved before events

Prediction: Before 7 September, a formal step will occur: a meeting with a defined group or a written request for revisions. The reading does not promise final budget approval.

Advice: By 12 August, send a one-page summary covering purpose, cost, risks, decision owner and two proposed meeting times.

Version A: The delay is procedural. Justice and the Hanged Man point towards a missing approval stage. Supporting evidence: a responsible person and decision process are named.

Version B: Interest is weaker than it appears. The Four of Cups and Two of Swords suggest avoidance. Supporting evidence: even after a clear request, nobody takes ownership and dates remain undefined.

Evidence supporting the prediction: a meeting with a decision-maker is scheduled by 7 September, or a written list of requested changes arrives.

Evidence contradicting it: the project is closed, a direct rejection arrives, or no formal step or written response occurs by the date.

Practical criterion: The team knows the next stage, owner and time frame even if the answer is no.

What happened during the 30 days

• the summary was sent on 11 August;
• on 15 August, a request came back for support-cost details;
• a meeting took place on 22 August, but nobody with budget authority attended;
• a second meeting was set for 10 September;
• no final decision existed by 7 September;
• the budget owner was named only after a direct question.

Honest evaluation

It would be inaccurate to say the whole reading “came true”. It correctly focused attention on process and produced a useful action, but the timing was optimistic and complete clarity did not arrive. That is more informative than an attractive accuracy score.

A five-status evaluation matrix

“Changed by action” does not make the reading less useful. It shows practical value. It simply must not be confused with an independent forecast.

Rules against hindsight fitting

Do not broaden the event

If the note says “written decision”, an informal hallway comment does not count. If it names a meeting with a particular person, any routine call is not enough.

Do not move the deadline silently

An event that occurs two months later may still be interesting, but a 30-day time claim was not confirmed. Record the later event separately.

Do not add details that were absent

After the result, it is easy to say, “I always knew this was about budget.” If the original note never mentioned budget, that detail does not belong in the evaluation.

Account for your own action

Advice may have changed the course of events. That is useful, but it is not proof of an independent prediction. Record the chain: action, response, outcome.

Keep the contradictions visible

If one part matched and two did not, the review must include all three. The complete original entry is the unit being tested.

Do not repeat the spread to obtain a new date

Multiple readings create several competing versions, making it easy to select the convenient one later. A new spread is appropriate only when the actual starting conditions or question have changed.

What a non-confirmed reading can teach

There is no need to defend the cards. A claim may fail because:

• the question was too broad;
• the interpretation contained no observable event;
• the time frame was arbitrary;
• no alternative was written;
• starting conditions changed;
• a coincidence was treated as a pattern;
• the symbol was useful as advice but weak as prediction;
• the interpretation was simply inaccurate.

The final possibility must remain available. Intellectual honesty begins when every reading does not need to be rescued.

Practical value matters more than a “hit rate”

At the 30-day review, ask more than “What matched?”

• Did the reading separate facts from assumptions?
• Did it produce one action that I actually completed?
• Did my question to another person—or to myself—become more precise?
• What would I have done without the spread, and how did the actual decision differ?

Sometimes the best outcome is not a predicted event. It is a timely conversation, a clearer boundary, the end of unproductive waiting, or an honest record that uncertainty remained.

A reusable journal template

Reading date: 8 August
Review date: 7 September
Question: …
Starting facts: …
Cards and positions: …
Prediction: …
Advice: …
Version A: …
Version B: …
Supporting evidence: …
Contradicting evidence: …
My action: …
Events before review: …
Status of each claim: …
Practical value: …
What I will change in the next question: …

Common mistakes

• writing the interpretation only after the event;
• using “soon”, “change” or “important conversation” without a date or criterion;
• treating advice as a prediction;
• recording one version with no alternative;
• remembering only matching details;
• explaining contradictions through “wrong energy”;
• extending the time frame until a suitable event appears;
• drawing a new spread to explain the old one;
• rating accuracy without accounting for personal action;
• turning one coincidence into universal proof of the method.

Questions and answers

Must the review be exactly 30 days later?

No. The period depends on the question, but it must be chosen in advance. Thirty days works for medium-range events; another question may need a week or three months.

What counts as partial confirmation?

Only a part defined beforehand. If the note expected a meeting and a decision, and only the meeting occurred, one element was confirmed and the other was not.

Can individual cards be scored?

Yes, when a specific claim was written for each card in advance. Otherwise, evaluate the statements in the interpretation rather than assigning meanings retrospectively.

What if the original note is too vague?

Mark it untestable and use that as information for the next question. Do not invent criteria after the outcome.

Should I do another spread after the review?

Only when a new question exists or the starting conditions have materially changed. Complete the old evaluation first and write what you learned.

In Reflecta, the original wording, action, criterion and review date can remain inside one saved situation. After the real event, add an observation without rewriting the earlier meaning. The practice becomes less dramatic and more useful: a coincidence stays a coincidence, a contradiction stays visible, and the decision returns to the person making it.

— Daniel Aster, official Reflecta author