We built a tarot reading platform. People used it. Now we have data that nobody else in the tarot space has — actual card draw frequencies, spread preferences, and behavioral patterns from over a thousand real readings.
Most tarot content is opinion. This is numbers.
The short version (updated May 2026): Across 1,370 readings and 4,110+ card draws between January and May 2026, Knight of Wands led the deck (78 draws), followed by The Hanged Man (76). The Emperor appeared least often. Major Arcana hit 28.4% of all draws — within 0.2 points of the mathematical expectation of 28.2%. Past-Present-Future remained the dominant spread at 78.5% of readings. Tuesday peaked again — +37% above average. And the questions people actually ask? 28% are about the future, only 13% about love.
The dataset
Between January and May 2026, aimag.me processed 1,370 AI-interpreted tarot readings containing more than 4,110 individual card draws across 69 unique users. The readings used cryptographically secure pseudorandom number generation — the same type of randomness used in online banking and encryption systems.
Some context on scale: 1,370 readings is not a massive dataset by social science standards. It is, however, one of the only publicly reported datasets of actual tarot card draw frequencies from a digital platform. We are publishing it because the tarot space has almost zero empirical data, and even a modest dataset beats no dataset.
This is the second update to our public dataset (originally 1,061 readings, April 2026). The patterns largely held. One thing flipped — see "Knight of Wands" below.
All data is fully anonymized. No questions, user identifiers, or personal information are included in this analysis.
The 10 most drawn tarot cards (updated May 2026)
| Rank | Card | Arcana | Times drawn | % of all draws |
|---|---|---|---|---|
| 1 | Knight of Wands | Minor | 78 | 1.90% |
| 2 | The Hanged Man | Major | 76 | 1.85% |
| 3 | Queen of Pentacles | Minor | 69 | 1.68% |
| 3 | Five of Wands | Minor | 69 | 1.68% |
| 3 | The Tower | Major | 69 | 1.68% |
| 6 | Ten of Swords | Minor | 68 | 1.65% |
| 6 | Six of Swords | Minor | 68 | 1.65% |
| 6 | Three of Wands | Minor | 68 | 1.65% |
| 6 | Four of Cups | Minor | 68 | 1.65% |
| 10 | The Hierophant | Major | 67 | 1.63% |
Expected frequency per card in a perfectly random system: 1.28% (1/78). The Knight of Wands at 1.90% represents a 48% deviation above expected — notable, but still within normal statistical variance for this sample size. A chi-squared test across all 78 cards would likely not reject the null hypothesis of uniform distribution.
Translation: the randomness is working as intended. No card is being systematically favored. What makes the list interesting is not the statistical significance — it is the psychological resonance.
The flip: Knight of Wands took the lead
In our April 2026 dataset (1,061 readings), The Hanged Man was the most drawn card. In our May 2026 dataset (1,370 readings), Knight of Wands moved into first place by a margin of two draws.
Statistically, this means almost nothing. The difference between 78 and 76 draws across thousands of card draws is exactly the kind of noise you would expect in a random system. Both cards are within 1.85–1.90% of total draws — basically interchangeable.
Symbolically, it is interesting. The Hanged Man is the card of waiting. Knight of Wands is the card of action — impulse, fire, charging forward. Over four months, the deck shifted from "pause" to "go." Whether that mirrors a shift in what people are bringing to their readings, or just statistical noise, is genuinely impossible to tell with this sample size.
We mention it because it is the kind of pattern most tarot blogs would treat as meaningful. We are treating it as random — and asking you to do the same until our dataset is large enough to test the hypothesis properly.
Why The Hanged Man still matters
Even at #2, The Hanged Man is the most-drawn Major Arcana card in our dataset. Voluntary suspension. Pause. Surrender. Seeing the world from an inverted perspective. The card of the moment before action — when you are hanging between two states, waiting for clarity.
People come to tarot during transitions. Uncertainty. The space between a question and its answer. The Hanged Man is that space, personified.
The 5 least drawn cards
| Rank | Card | Arcana | Times drawn |
|---|---|---|---|
| 74 | Two of Wands | Minor | 33 |
| 75 | The Devil | Major | 31 |
| 76 | Nine of Wands | Minor | 30 |
| 77 | Two of Swords | Minor | 29 |
| 78 | The Emperor | Major | 29 |
The Emperor — the archetype of structure, authority, and control — was drawn least often. Make of that what you will. We are not going to claim the universe is sending a message. But there is something almost poetic about the card of rigid order appearing least frequently in a dataset of people seeking guidance during uncertainty.
Major vs Minor Arcana: the randomness check
This is the number that matters most for validating our system's randomness.
A standard 78-card deck contains 22 Major Arcana (28.2%) and 56 Minor Arcana (71.8%). If our draw system is truly random, the actual distribution should approximate these percentages.
| Arcana | Cards in deck | Expected % | Actual draws | Actual % |
|---|---|---|---|---|
| Major | 22 | 28.2% | 1,167 | 28.4% |
| Minor | 56 | 71.8% | 2,943 | 71.6% |
28.2% expected. 28.4% observed. Across 4,110+ draws. The match remains within statistical noise — a reminder that randomness, given enough trials, converges on its theoretical distribution with mathematical inevitability.
This is the honest math behind aimag.me. We can prove our system is fair. What we cannot prove — and what we do not claim — is that any individual card is "meant for you" in some cosmic sense. The meaning emerges from your interpretation, not from the algorithm.
Suit distribution
| Suit | Element | Times drawn | % of minor draws |
|---|---|---|---|
| Wands | Fire | 644 | 26.0% |
| Swords | Air | 627 | 25.3% |
| Pentacles | Earth | 623 | 25.2% |
| Cups | Water | 582 | 23.5% |
Expected: 25% each. Observed range: 23.5%–26.0%. Cups is slightly underrepresented — the suit of emotions, relationships, and intuition drew 6% below expected. In a larger dataset this would likely equalize. With 2,476 minor draws, the deviation is not statistically significant.
The near-uniform distribution across suits further confirms what the Major/Minor split already demonstrated: the randomness engine is doing its job. Cards are not being weighted, favored, or manipulated. The patterns people find in their readings emerge from their own meaning-making, not from the distribution.
How people read: spread preferences
| Spread | Readings | % of total |
|---|---|---|
| Past-Present-Future (3 cards) | 832 | 78.4% |
| Love: Me, Them, Us (3 cards) | 58 | 5.5% |
| Single Card | 47 | 4.4% |
| Dual Card | 22 | 2.1% |
| Cross (5 cards) | 17 | 1.6% |
| Love Spread (6 cards) | 17 | 1.6% |
| Celtic Cross (10 cards) | 16 | 1.5% |
Past-Present-Future dominates with 78% of all readings. This three-card spread is the default for new users, which partially explains its prevalence. But even among users who tried other spreads, many returned to it. Three cards is the sweet spot: enough structure to tell a story, not so many that the reading becomes overwhelming.
Average cards per reading: 3.2 — confirming the three-card dominance.
The Celtic Cross, tarot's most famous spread, accounts for only 1.5% of readings. Ten cards is a commitment. Most people want clarity, not complexity.
When people read tarot (updated)
| Day | Readings | vs average |
|---|---|---|
| Monday | 198 | +1% |
| Tuesday | 271 | +37% |
| Wednesday | 182 | −7% |
| Thursday | 210 | +7% |
| Friday | 185 | −6% |
| Saturday | 167 | −15% |
| Sunday | 157 | −20% |
Tuesday is the peak tarot day. Not Monday (the stereotypical day of anxiety). Not Sunday (the day of reflection). Tuesday — the day when the week's reality has set in, the initial Monday momentum has faded, and people start confronting the actual shape of what they are dealing with.
The Tuesday spike persisted across both data updates. It is the single most stable weekly pattern we observe — and the most counterintuitive.
Weekend readings (Sat+Sun = 23.7% of all readings) are consistently lower than weekdays (76.3%), suggesting tarot on our platform functions primarily as a work-life reflection tool rather than a leisure activity.
Peak hour: 1 PM UTC — morning on the US East Coast, early afternoon in Europe. The lunch-break reading. The stolen moment of clarity between meetings.
What people actually ask (new in May 2026)
We analyzed 1,261 readings where users typed a question of more than 10 characters. Most tarot content assumes love is the dominant question. The numbers say otherwise.
| Category | Readings | % of asked questions |
|---|---|---|
| Future / uncertainty ("when", "will", "future") | 354 | 28.1% |
| Love / relationships | 169 | 13.4% |
| Career / work | 129 | 10.2% |
| Money / finance | 13 | 1.0% |
| Health | 12 | 1.0% |
| Family | 9 | 0.7% |
| Other / mixed | 575 | 45.6% |
Future-oriented questions are the single largest category, more than double love queries. People are not coming to tarot for romance prediction. They are coming for time anxiety — the question "when?" appearing in 28% of asked readings.
Love is real, but smaller than the tarot industry tells you. Money and health are barely present (1% each), suggesting people self-screen these questions to other tools — finance apps, doctors, search engines.
Career questions outpace money questions 10:1. People ask tarot about what kind of work they should do. Rarely about how much they will earn.
Love readings: the irony
In readings using love-specific spreads, the most frequently drawn card was:
The Hanged Man (10 draws out of 75 love readings)
The card of waiting. Of surrendering control. Of accepting that you cannot force a romantic outcome. People come to love spreads wanting answers about relationships — and the most common card they receive literally means "the answer will come when you stop demanding it."
The second most common love card was Temperance (9 draws) — patience, balance, the slow blending of two elements. Not exactly the dramatic romantic confirmation most people hope for.
Two of Cups — the classic romance card — appeared 6 times. Present, but not dominant.
The questions people ask
Average question length: 64 characters, 12 words. People are concise when they sit down with tarot. They already know what they are asking about. The tarot reading is not where the question forms. It is where the question gets answered.
Our platform serves readers in 7 languages. English dominates (90%), followed by Polish (4%) and Portuguese (3%). The remaining readings split across French, Spanish, German, and Italian.
What this data does and does not tell us
It does tell us:
- Digital tarot systems produce genuinely random distributions (confirmed by the 28.2% Major Arcana match)
- People prefer simple spreads overwhelmingly (78% three-card)
- Tarot reading peaks on Tuesdays during work hours
- The most frequently drawn card during our observation period was The Hanged Man
It does not tell us:
- Whether any card is "luckier" or more significant than others
- Whether the frequency of a card has any predictive or spiritual meaning
- Anything about the quality or accuracy of readings
The value of tarot is in the interpretation, not the draw. As we have argued in our analysis of randomness and meaning-making, the randomness is the feature that makes the system work for self-reflection. This data confirms the randomness is real.
Methodology and limitations
- Period: January 1 – May 2, 2026 (4 months)
- Sample: 1,370 readings, 4,110+ card draws, 69 unique users
- Randomness source: PHP
random_int()(CSPRNG) - Question categorization: keyword-based pattern matching (love/relationship/partner; work/job/career/business; money/finance/pay; etc.) on 1,261 questions of length >10 characters
- Limitations: Sample size is moderate. The dataset skews heavily toward three-card Past-Present-Future readings due to default spread selection. Love-spread subsample remains small. We have not yet accumulated enough data to identify statistically significant deviations in card frequency. Question categorization is heuristic, not NLP-grade.
We plan to publish updated statistics quarterly as our dataset grows. If you are a researcher, journalist, or content creator who wants to cite this data, you are welcome to do so — link back to this page as the source. We would also love to hear from you: contact [email protected] and tell us what you would like the next dataset to investigate.
Explore for yourself
Curious which cards will appear in your reading? Every draw on aimag.me uses the same cryptographically secure randomness analyzed in this article.
Start a free reading and see which cards the algorithm selects for you. Or try a focused single-card draw: money tarot for financial insight, career tarot for professional guidance, or love tarot for relationship clarity.
This article is updated quarterly as our dataset grows. Last updated: May 2, 2026. Dataset: 1,370 readings, 4,110+ draws, 69 unique users.
Related reading
- Is tarot real? What science actually says — the cognitive science behind why random cards produce meaningful insights
- All 78 tarot card meanings — upright, reversed, and psychological depth for every card in the Rider-Waite deck
- The Hanged Man tarot meaning — our most frequently drawn card, explored in full
