Data study
The Rarest Lover Type, and How 14,000+ People Actually Love
By An Nayal - Published August 18, 2026
We looked at 14,556 Type of Lovers quizzes taken between September 10, 2024 and August 2, 2026. Here is what the numbers say about how people actually love: which of the 16 lover types is rarest, which is most common, how many people finish the quiz, and when they take it. Every figure below traces to the same anonymous dataset.
Type distribution is a derived estimate, cross-validated, not a raw label. See the methodology for exactly how it was built and what it does and does not support.
21.6%
are the Ember
The most common lover type, more than one in five people.
under 1.6%
each for the three rarest
Jewel, Stone, and Sky sit alone at the bottom.
78.3%
finish the quiz
11,403 of 14,556 complete it, in about five minutes.
The rarest lover types
Three types share the bottom of the table, each claimed by under 1.6% of the 11,357 people we could place: the Jewel (1.2%), the Stone (1.3%), and the Sky (1.5%). Two are Earth types and one is Air, and all three lean practical and logical in how they show love.
We treat these three as a cluster rather than crowning one single winner. Two independent ways of reading the data both put the same three types last, but they order them slightly differently, so the honest finding is that the Jewel, the Stone, and the Sky are the rarest, each under 1.6%, and not that any one of them is precisely rarest of all.
The most common lover type
At the other end, one type stands clearly ahead. The Ember (warm, empathetic, and deeply devoted) accounts for 21.6% of results, 2,455 people, more than one in five and over eight points ahead of the next type. If there is a default way this quiz sees people loving, it is warm, devoted, and emotionally direct.
All 16 lover types, most to least common
The full distribution across the 11,357 completed quizzes that resolved to a single clear type. Tap a name to read the full type.
| Rank | Lover type | Element | Respondents | Share |
|---|---|---|---|---|
| 1 | EmberEDFC | Fire | 2,455 | 21.6% |
| 2 | WaveEPFA | Water | 1,474 | 13.0% |
| 3 | SparkEDFA | Fire | 1,393 | 12.3% |
| 4 | StreamEPFC | Water | 1,303 | 11.5% |
| 5 | GemLDFC | Earth | 741 | 6.5% |
| 6 | AuraLPFA | Air | 665 | 5.9% |
| 7 | TideEPRA | Water | 559 | 4.9% |
| 8 | BreezeLPFC | Air | 526 | 4.6% |
| 9 | CrystalLDFA | Earth | 428 | 3.8% |
| 10 | BlazeEDRC | Fire | 381 | 3.4% |
| 11 | InfernoEDRA | Fire | 350 | 3.1% |
| 12 | WhirlwindLPRA | Air | 313 | 2.8% |
| 13 | CurrentEPRC | Water | 301 | 2.7% |
| 14 | SkyLPRC | Air | 174 | 1.5% |
| 15 | StoneLDRA | Earth | 153 | 1.3% |
| 16 | JewelLDRC | Earth | 141 | 1.2% |
N = 11,357 completed quizzes with a single clear type. Shares are of that set and total about 100% after rounding. Type is derived from each respondent's raw per-type scores, not a stored label; see the methodology.
How the world loves
The same quiz does not come back the same everywhere. Grouping the 11,357 typed quizzes by the country their connection was in, and reading each of the four traits as a simple lean, clear regional patterns show up in how heart-led, how direct, and how go-with-the-flow people are. One thing never moves: the Ember is the single most common type in every high-volume country, so the interesting story is never the top type, it is the leans and the runner-up.
90%
go with the flow in the Philippines
The most easygoing country in the data: 90% lean Flexible over strong-willed, against 79% overall.
82%
lead with the heart in India
The most heart-led and most direct of the big samples: 82% Emotional and 58% Direct, both the highest of any country with 500+ quizzes.
58%
keep their independence in Japan
East Asia wants the most space: Japan leans 58% Autonomous and Taiwan 54%, against 47% overall.
52%
keep it measured in Singapore
The most reserved of the high-volume countries: 52% lean Passive rather than Direct, just ahead of Taiwan, while most of the world tilts Direct (53% overall).
| Country | Quizzes | Emotional | Direct | Flexible | Independent | Runner-up type |
|---|---|---|---|---|---|---|
| All countries | 11,357 | 77% | 53% | 79% | 47% | - |
| United States | 1,792 | 79% | 55% | 81% | 44% | Stream 13% |
| Malaysia | 1,149 | 72% | 50% | 79% | 49% | Spark 12% |
| Hong Kong | 724 | 77% | 50% | 79% | 48% | Wave 15% |
| India | 507 | 82% | 58% | 80% | 44% | Spark 15% |
| United Kingdom | 480 | 79% | 52% | 75% | 44% | Stream 13% |
| Indonesia | 464 | 72% | 57% | 86% | 45% | Spark 14% |
| Canada | 447 | 73% | 55% | 79% | 48% | Spark 13% |
| Germany | 429 | 77% | 55% | 76% | 47% | Wave 13% |
| Australia | 348 | 75% | 51% | 77% | 43% | Wave 14% |
| Singapore | 335 | 72% | 48% | 80% | 50% | Wave 14% |
| Taiwan | 332 | 79% | 49% | 78% | 54% | Wave 15% |
| France | 317 | 79% | 51% | 80% | 44% | Wave 14% |
| Philippines | 313 | 77% | 56% | 90% | 44% | Wave 16% |
| Italy | 208 | 80% | 52% | 72% | 50% | Wave 15% |
| The Netherlands | 174 | 77% | 55% | 75% | 51% | Wave 13% |
| Japan | 172 | 72% | 51% | 83% | 58% | Wave 19% |
| Spain | 163 | 83% | 58% | 77% | 52% | Spark 16% |
| Poland | 154 | 79% | 46% | 70% | 45% | Wave 14% |
Each percentage is the share of that country's typed quizzes leaning to that pole: Emotional (vs logical), Direct (vs reserved), Flexible (vs strong-willed), and Independent (needs space, vs togetherness). Only the 18 countries with at least 150 typed quizzes are shown, most quizzes first.
The Ember tops every country above, but the runner-up moves around: in the United States and United Kingdom it is the Stream, in the Philippines and Japan it is the Wave, and in India, Indonesia, and Spain it is the Spark.
A note on geography: the country comes from each quiz's IP address, so it reflects where the quiz was taken, not a verified nationality - a traveler or VPN sits in whatever country their connection is in. These are historical leads through August 2, 2026, and this data carries no city-level detail.
How people take the quiz
78.3% finish. Of the 14,556 quizzes recorded, 11,403 were completed. For a free quiz with no account required, better than three in four finishing is a strong rate.
About five minutes each. The typical quiz takes 4:50 (median), with the middle half of people finishing between 3:35 and 6:41. The mean is a bit higher at 5:45, pulled up by a handful of tabs left open for up to 118 minutes. The fastest finish of just 3 seconds is almost certainly a bot, not a reader.
October 2025 was the peak. 4,186 quizzes landed that month, 28.8% of everything in the dataset. The single busiest week was October 6 to 12, 2025 with 1,479 quizzes. Volume built slowly through 2024, climbed across 2025, spiked hard in the autumn, and has settled back down since.
Quizzes per month
Monthly quiz submissions, September 10, 2024 to August 2, 2026. The final month is partial (the export cut off in early August 2026).
Across the week the load is fairly even, with Tuesday the busiest day and Saturday the quietest, and a broad afternoon peak around 3 PM. The file does not label a timezone, so read the time of day as a ranking within the data, not a specific local clock, especially given how international the respondent base is.
Methodology
The data. 14,556 quiz submissions recorded between September 10, 2024 and August 2, 2026, plus 14 newer submissions through August 18, 2026 used only to validate the method, for 14,570 in total. These are anonymous aggregates. No names, emails, or individual rows appear anywhere on this page.
How the type was worked out. The export did not store a plain lover type for each person. Its result column held opaque legacy identifiers, one unique token per submission, that carry no type meaning on their own. So we reconstructed each person's type from the raw scores the quiz already records: every submission carries a match score against all 16 types, and we took each person's highest-scoring type. That covers 11,357 of the 11,403 completed quizzes; 46 exact ties were set aside rather than guessed.
How we checked it. We rebuilt the type a second, independent way, from the four underlying trait axes, and the two methods agree on about 84% of quizzes. Where they disagree it is almost always the first trait (heart-led versus head-led), which is a genuine near-tie for many people; the other three traits are reliable. We also checked the method against 14 quizzes that did carry a real type code, and 11 of 14 matched exactly, the three misses being rounding ties. This is why we report the rarest types as a cluster of three rather than ranking them to the decimal.
What this is, and is not. Type of Lovers is a fun, free relationship quiz, not a clinical or validated psychological instrument, and this is a look at how its own takers land across the 16 types, not a claim about the general population. The type distribution is a careful, cross-validated estimate, not a stored value. We did not report a traffic-source breakdown (99.7% of rows carry no source tag) or an email opt-in rate (the export measures it two different, non-comparable ways), because the data does not honestly support either.
What the quiz is based on
The lover types are framework-inspired, not a clinical instrument. The traits behind them draw on well-known ways of thinking about relationships, used here as inspiration rather than a claim of scientific equivalence:
- The five love languages (Gary Chapman) - how affection is given and received.
- Attachment theory (Bowlby and Ainsworth) - how people handle closeness, conflict, and reassurance.
- Four-letter typologies and trait models (Myers-Briggs, the Big Five) - the idea of a short, memorable code built from a few independent traits.
For the full breakdown of how the quiz works, see the about page.
Cite this study
You are free to cite or reference these findings, with a link back to this page (CC BY 4.0). Suggested citation:
Type of Lovers. (2026). The Rarest Lover Type, and How 14,000+ People Actually Love. https://www.typeoflovers.com/study
Canonical URL: https://www.typeoflovers.com/study
Download the data: lover-type distribution (CSV)
