Every dating platform gets asked about star signs, and most answer with content: Gemini compatibility guides, best matches for a Scorpio, what to text a Leo. We decided to answer with data instead. We took 70,759 first likes between SoulMatcher users whose birth dates we know, sorted every pair by the classic rules of zodiac compatibility, and checked whether the stars had any opinion about who liked whom back.
They did not. Astrologically compatible pairs got a reciprocal like 2.91% of the time. Incompatible pairs: 2.97%. Same sign: 2.96% against 2.94% for everyone else. Whatever decides whether your like comes back, it is not written in your birth chart.
Below are the numbers, including the one table where astrology fans might briefly feel hopeful, the significance tests behind each comparison, and what our data says actually moves the odds.
What we tested
Classic Western astrology groups the twelve signs into four elements: fire (Aries, Leo, Sagittarius), earth (Taurus, Virgo, Capricorn), air (Gemini, Libra, Aquarius) and water (Cancer, Scorpio, Pisces). The traditional compatibility rules are simple: same element gets along, fire pairs with air, earth pairs with water, and the rest is friction.
Read also: The Compatibility Factors That Actually Predict Success
One clarification before the numbers. We are testing the popular sun-sign compatibility rules used in dating content, the version built on those four elements, not full natal-chart synastry. Astrology systems that bring in birth time, Moon and Venus placements, aspects and houses make broader claims than a birth date alone can test. What follows is a test of the astrology people actually meet in horoscope columns and compatibility quizzes, applied to real swipes.
Those rules split our 70,759 first likes almost exactly in half: 35,488 pairs the stars approve of, 35,271 they warn against. That near-even split gives us two similarly sized comparison groups, so neither result is being driven by a much smaller sample. If astrology mattered, the two halves should perform differently.
The result
| Pair type | First likes | Got a reciprocal like within 30 days |
|---|---|---|
| Astrologically compatible | 35,488 | 2.91% |
| Astrologically incompatible | 35,271 | 2.97% |
The gap is 0.06 percentage points, and it points, very slightly, the wrong way for astrology: the incompatible pairs did marginally better. It is not a real difference. A two-proportion test returns p = 0.67, and the 95% confidence interval for the gap runs from about minus 0.30 to plus 0.19 percentage points, comfortably straddling zero. In plain terms, if astrology had switched off entirely the day we pulled the data, this is exactly the sort of non-difference we would expect to see.
Read also: Shared Sense of Humor as a Stronger Predictor of Compatibility Than Shared Hobbies
Same story for the strongest version of the claim, that like attracts like:
| Pair type | First likes | Reciprocated within 30 days |
|---|---|---|
| Same zodiac sign | 5,846 | 2.96% |
| Different signs | 64,913 | 2.94% |
A 0.02-point gap, p = 0.92. If you have been holding out for a fellow Pisces, the data will not stop you, but it will not help you either.
The element table astrologers will not enjoy
Maybe the pass-or-fail split is too blunt, and some element pairings are quietly special. So we ranked all ten element combinations by reciprocation rate. Astrological compatibility is symmetric, fire with air is the same claim whoever swiped first, so each pairing is one row. If the textbook were right, the compatible pairings (same element, fire with air, earth with water) should crowd the top.
Read also: Do Personality Tests Actually Predict Compatibility?
| Element pair | First likes | Reciprocated | Textbook verdict |
|---|---|---|---|
| Water + water | 4,649 | 3.38% | Compatible |
| Fire + earth | 8,545 | 3.11% | Clash |
| Earth + water | 8,804 | 3.11% | Compatible |
| Earth + earth | 3,840 | 3.10% | Compatible |
| Fire + fire | 4,672 | 3.06% | Compatible |
| Fire + water | 9,330 | 3.02% | Clash |
| Air + water | 9,172 | 2.97% | Clash |
| Earth + air | 8,224 | 2.75% | Clash |
| Fire + air | 9,174 | 2.67% | Compatible |
| Air + air | 4,349 | 2.18% | Compatible |
Look at the top and the bottom. The best-performing pairing, water with water, is one astrology blesses. The worst, air with air, is also one astrology blesses. The textbook-perfect fire-air match sits second from the bottom. The compatible pairings are scattered from the top of the table to the floor, which is not what a real effect looks like.
There is some genuine unevenness here, and it would be dishonest to wave it away by eye. A chi-square test across the ten pairings does reject the idea that they are all identical (chi-square 17.9, 9 degrees of freedom, p = 0.04). But it does not run along astrology's axis. The spread tracks which single element is in the pairing, combinations involving air sit low and combinations involving water sit high, rather than whether the two elements are supposed to be compatible. And which element is only shorthand for which slice of birthdays we are looking at, each with its own mix of ages, genders and swiping habits. The effect is also stronger for the person sending the like than the one receiving it, which is another way of saying it is a fact about the swiper, not about the couple. Astrology's claim is about the pair. Tested on the pair, it is null.
After the match, the stars stay silent
Maybe astrology does not decide who likes whom, but shapes what happens next? We checked that too. Among mutual matches where both birth dates are known, astrologically incompatible pairs reached a real two-way conversation 37.2% of the time (387 of 1,040); compatible pairs, 36.2% (359 of 991). A one-point difference, again in the wrong direction for astrology, and again not significant (two-proportion test, p = 0.65). Whatever turns a match into a conversation, the birth chart is not in the room for that either.
Read also: The Way You Argue in the First Year Predicts Everything That Follows
What actually moves the odds
The frustrating and useful thing about our data is that the factors that do move the numbers are unglamorous. In our study of 417,785 swipes, the differences that dwarfed anything astrological had nothing to do with birthdays: how long a like had been sitting unanswered (the chance of a match falls from about 3.0% at the moment a like is sent to 0.5% after a silent week), who did the liking, and how selectively each side swiped. We did not build a predictive model, so these are associations rather than predictors, but they are large exactly where the zodiac effect is nothing. The one real pattern in the element table above, the swiper's own element counting for more than the pairing, is just another face of the same thing: who initiates. Chemistry, as we argued in why chemistry isn't always a good sign, is a famously poor screening tool; a birthday appears to be an even poorer one, with the added disadvantage of being fixed before you could walk.
This is also why SoulMatcher takes a different approach: its compatibility profile is built from a psychological assessment rather than a zodiac sign. Whether that score predicts real dating outcomes is a separate question, and one we can put to exactly the same test we just ran on the stars: same dataset, same reciprocity measure, with the compatibility score in the place the elements just failed to fill. That study is next.
Why zodiac compatibility feels true anyway
None of this will change what people believe, and it is worth separating two different reasons for that. Start with the evidence against the mechanism. In a well-known 1985 double-blind study published in Nature, the physicist Shawn Carlson tested whether astrologers could match natal charts to the personality profiles of people they had never met. Working from the charts alone, they were unable to pick the matching profile at a rate better than chance (Carlson, 1985).
Read also: Can You Guess Your Own Psychological Profile? We Checked 7,013 People
Belief survives that anyway, and the reason is less about astrology than about memory. Zodiac compatibility is easy to confirm after the fact: the hits are memorable, the misses are forgotten. A Taurus who clicked with a Virgo remembers the textbook said they would; a Taurus who did not simply never opens the textbook. Dating hands that selective memory thousands of fresh chances a year to feel confirmed, and the times it was wrong are never counted. Our 70,759 likes count them.
We wrote about the gentler side of this in Astrology and Dating: Harmless Fun or Genuinely Limiting? The data adds one number to that discussion: in 70,759 real decisions, the limiting part cost people nothing, because the stars were not steering anything to begin with. Filter your matches by sign if it makes swiping more fun. Just know that you are decorating the lottery ticket, not changing the draw.
Methodology and definitions
Sample: 70,759 first likes between SoulMatcher users where both people's dates of birth are on file, sent at least 30 days before measurement so every like has a full observation window. Data aggregated on 31 August 2026.
- Zodiac sign: derived from date of birth using standard tropical date ranges.
- First like: a unique like sent while no reciprocal like existed yet; repeated likes in one direction are collapsed to the earliest.
- Astrologically compatible: same element, or the traditional complementary pairs (fire with air, earth with water). This classification split the sample almost exactly in half, 50.2% to 49.8%.
- Reciprocated: a reciprocal like arrived within 30 days of the first like.
- Conversation: a mutual match in which both people sent at least one human message; automatic system messages are excluded. The conversation comparison covers 2,031 matches with both birth dates known.
- Significance: reciprocity and conversation comparisons use a two-proportion z-test; the element ranking uses a chi-square test across the ten unordered element pairings. Percentages are rounded for display, but every test runs on the raw counts.
All figures are group-level aggregates; no individual profiles were reviewed. Because this sample comes from users of a psychology-based dating app, its attitudes and behaviour may differ from those of dating-app users generally: this is one platform's data, not a claim about every couple on Earth.
This article is informational and reports aggregate, de-identified patterns of user behaviour. It is not advice about any individual person or relationship.
These articles are for information only. They are not a substitute for consulting a psychologist, psychotherapist or other qualified specialist.
