MyType.lol / Data & context

MyType.lol statistics: what do face choices reveal?

Curious about your MyType? Explore what a published face preference study reveals about the choices people make, then discover your own type with MyType.lol.

Find your type on MyType.lol
Illustrative AI-generated portraits from MyType.lol
Illustrative AI-generated portraits from MyType.lol
Illustrative portraits from MyType.lol. These faces are not the subjects of the statistics below.

01 / The sample

One week, millions of choices.

The reference study recorded activity in a self-selected online test. A run is a browser’s pass through the test on one UTC day, not necessarily a unique person. Its participants, face deck and recommendation system differ from MyType.lol.

4,502,629

taps in the reference study

48,356

test runs in the reference study

155

countries with a run in the reference study

43,673 runs reached a result; 1,084,545 grids were completed. Across all grids, 37.7% were passed without a tap. Those totals include personalized grids and replacement faces.

The week in numbers

ActivityTotal
Runs started48,356
Runs reaching a result (6+ grids)43,673 (90.3%)
Grids completed1,084,545
Faces shown, including recommendations17,518,035
Taps across all grids4,502,629
Grids passed with no tap408,784 (37.7%)
Runs on men’s faces25,540 (52.8%)
Runs on women’s faces16,993 (35.1%)
Runs on both decks5,823 (12.0%)
Countries represented155

The busiest day was November 13, with 16,016 runs. A result required at least six grids in this source test.

02 / Randomly shown faces

How selective were the choices?

The rates below use randomly dealt faces, excluding personalized recommendations. A randomly shown face was tapped 10.3% of the time for women’s faces and 10.1% for men’s faces. On 52.5% and 55.0% of the respective random grids, no face was tapped.

Women’s faces

10.3%

114,023 taps / 1,102,902 random showings

Men’s faces

10.1%

171,546 taps / 1,700,370 random showings

When one run tapped a randomly shown face, another run shown the same face tapped it 16.4% of the time for women’s faces and 15.7% for men’s. The deck labels describe the faces shown, not the identity of the people choosing them.

03 / Selectivity

More than half of random grids got no tap.

A grid of twelve faces sounds like plenty of choice. Yet in most randomly dealt grids, the run selected none of them. This is useful context for any MyType-style test: passing over a crowd can be part of a clear preference pattern. It does not mean that a participant has failed to make a choice.

How many of 12 random faces were tapped?
Women’s facesMen’s faces

No faces

52.5%
55.0%

1 face

19.3%
17.2%

2 faces

10.8%
9.6%

3 faces

6.5%
6.3%

4 faces

4.1%
4.3%

5 faces

2.7%
2.9%

6–12 faces

4.1%
4.6%
0%30%60%

Source: The Face Report, November 10–16, 2026. Rates refer to randomly shown AI faces in its test. Women’s deck: 75,968 grids; men’s deck: 127,300 grids. Published percentages are rounded.

View chart data
CategoryWomen’s facesMen’s faces
No faces52.5%55%
1 face19.3%17.2%
2 faces10.8%9.6%
3 faces6.5%6.3%
4 faces4.1%4.3%
5 faces2.7%2.9%
6–12 faces4.1%4.6%

Across the 43,673 runs that reached a result, the median was 3.47 taps per grid. That is much higher than the random-grid rate because later grids were mostly personalized and a tap brought a replacement face into view. One grid could therefore accumulate more than twelve taps. The lower quarter of runs averaged 1.46 taps or fewer; the upper quarter averaged 7.87 or more.

The spread across a whole run

DeckLower tenth ≤Lower quarter ≤MedianUpper quarter ≥Upper tenth ≥
All decks0.631.463.477.8715.75
Women’s faces0.671.503.457.6014.86
Men’s faces0.581.413.438.0016.21
Both decks0.691.633.708.2916.50

Taps per grid; personalized grids and replacement faces are included.

04 / Shared taste

People agree a little. Most picks still differ.

If someone likes a face, does that make another person more likely to like it too? In the reference study, yes. When the first run tapped a randomly shown face, the second run tapped it 16.4% of the time for women’s faces and 15.7% for men’s. When the first run skipped it, the corresponding rates were 9.9% and 10.3%.

Did the second run tap the face?
Women’s facesMen’s faces

When the first run tapped

16.4%
15.7%

When the first run skipped

9.9%
10.3%
0%10%20%

Pairs of different runs shown the same face at random; these are conditional tap rates.

View chart data
CategoryWomen’s facesMen’s faces
When the first run tapped16.4%15.7%
When the first run skipped9.9%10.3%
How much did their choices overlap?
Women’s facesMen’s faces

Observed shared choices

9.0%
8.5%

Equal-probability baseline

5.6%
5.8%
0%6%12%

Faces both runs tapped, divided by faces either run tapped. The baseline assumes equal tap probability across faces.

View chart data
CategoryWomen’s facesMen’s faces
Observed shared choices9%8.5%
Equal-probability baseline5.6%5.8%

The overlap measure answers a different question from the conditional rate: among faces selected by at least one run, what share did both select? The answer was 9.0% for women’s faces and 8.5% for men’s. That is above the study’s equal-probability baseline, but still leaves a large majority of choices unshared.

A small part of the deck received a large share of taps
Women’s facesMen’s faces

Top 1% of faces: share of taps

4.2%
3.7%

Top 10%: share of taps

29.4%
27.6%

Bottom 50%: share of taps

17.1%
17.2%
0%17.5%35%

Faces were ranked on half the runs and evaluated on the other half, in both directions, then averaged. Shares assume equal exposure.

View chart data
CategoryWomen’s facesMen’s faces
Top 1% of faces: share of taps4.2%3.7%
Top 10%: share of taps29.4%27.6%
Bottom 50%: share of taps17.1%17.2%

The top tenth attracted almost three tenths of taps. The bottom half still attracted about one sixth. This mix of shared popularity and individual variation helps explain why a personal MyType result can feel more informative than a single popularity ranking.

05 / The details

Patterns in the reference deck.

Each percentage is the tap rate for AI faces with a model-assigned label. Labels overlap: these are associations within this deck, not evidence that a single feature caused a choice.

Apparent age

Apparent age: tap rate
Women’s facesMen’s faces

Under 25

18.4%
14.3%

25–29

12.2%
12.0%

30–39

7.4%
8.0%

40–49

4.0%
4.9%

50 and over

2.0%
2.1%
0%10%20%

The Face Report, November 10–16, 2026. Randomly dealt AI faces; labels may overlap.

View chart data
CategoryWomen’s facesMen’s faces
Under 2518.4%14.3%
25–2912.2%12%
30–397.4%8%
40–494%4.9%
50 and over2%2.1%
See full percentage table
LookWomen’s facesMen’s faces
Under 2518.4%14.3%
25–2912.2%12.0%
30–397.4%8.0%
40–494.0%4.9%
50 and over2.0%2.1%

Hair color

Hair color: tap rate
Women’s facesMen’s faces

Blonde

13.4%
11.6%

Red

13.0%
10.9%

Brown

12.5%
13.1%

Light brown

12.4%
12.0%

Dark brown

11.4%
12.3%

Black

8.8%
7.5%

Dyed

7.8%
5.8%

Gray

3.8%
4.9%
0%10%20%

The Face Report, November 10–16, 2026. Randomly dealt AI faces; labels may overlap.

View chart data
CategoryWomen’s facesMen’s faces
Blonde13.4%11.6%
Red13%10.9%
Brown12.5%13.1%
Light brown12.4%12%
Dark brown11.4%12.3%
Black8.8%7.5%
Dyed7.8%5.8%
Gray3.8%4.9%
See full percentage table
LookWomen’s facesMen’s faces
Blonde13.4%11.6%
Red13.0%10.9%
Brown12.5%13.1%
Light brown12.4%12.0%
Dark brown11.4%12.3%
Black8.8%7.5%
Dyed7.8%5.8%
Gray3.8%4.9%

Eye color

Eye color: tap rate
Women’s facesMen’s faces

Green

12.5%
11.6%

Brown

12.5%
13.3%

Blue

12.1%
11.2%

Hazel

11.0%
11.5%

Dark brown

9.2%
8.6%
0%10%20%

The Face Report, November 10–16, 2026. Randomly dealt AI faces; labels may overlap.

View chart data
CategoryWomen’s facesMen’s faces
Green12.5%11.6%
Brown12.5%13.3%
Blue12.1%11.2%
Hazel11%11.5%
Dark brown9.2%8.6%
See full percentage table
LookWomen’s facesMen’s faces
Green12.5%11.6%
Brown12.5%13.3%
Blue12.1%11.2%
Hazel11.0%11.5%
Dark brown9.2%8.6%

Details and expression

Details and expression: tap rate
Women’s facesMen’s faces

Glasses

8.9%
9.7%

No glasses

10.6%
10.2%

Freckles

11.9%
11.8%

No freckles

10.0%
9.8%

Dimples

10.7%
11.4%

No dimples

10.3%
10.0%

Neutral expression

10.3%
9.6%

Soft smile

10.4%
10.5%

Smile

10.4%
10.2%
0%10%20%

The Face Report, November 10–16, 2026. Randomly dealt AI faces; labels may overlap.

View chart data
CategoryWomen’s facesMen’s faces
Glasses8.9%9.7%
No glasses10.6%10.2%
Freckles11.9%11.8%
No freckles10%9.8%
Dimples10.7%11.4%
No dimples10.3%10%
Neutral expression10.3%9.6%
Soft smile10.4%10.5%
Smile10.4%10.2%
See full percentage table
LookWomen’s facesMen’s faces
Glasses8.9%9.7%
No glasses10.6%10.2%
Freckles11.9%11.8%
No freckles10.0%9.8%
Dimples10.7%11.4%
No dimples10.3%10.0%
Neutral expression10.3%9.6%
Soft smile10.4%10.5%
Smile10.4%10.2%

The age chart shows the steepest gradient: rates fell at every age band on both decks. In the hair and eye charts, several colors were much closer together. These are comparisons between portraits carrying those labels, rather than controlled experiments changing one feature at a time.

On women’s faces, light makeup had a 12.0% tap rate, defined makeup 11.4%, and no makeup 7.2%. As with the other labels, lighting, age, hair and other visible details may contribute to that association.

For men’s faces in the reference study, short beards were tapped at 11.6%, compared with 10.1% for no facial hair and 5.9% for full beards. Apparent ages are model labels for AI-generated adult faces, not verified ages.

06 / Styling & labels

Hair, facial hair and type tags add more detail.

The strongest-looking row is not necessarily the best styling advice. Hair length, age, expression and other attributes appear together in the same portraits. In this deck, long hair had a higher tap rate on women’s faces, while medium-length hair led on men’s faces. Wavy hair performed above each deck’s overall rate.

Hair length
Women’s facesMen’s faces

Very short

4.1%
8.1%

Short

8.6%
10.2%

Medium

9.9%
12.2%

Long

11.6%
5.2%
0%7.5%15%

Source: The Face Report, November 10–16, 2026. Rates refer to randomly shown AI faces in its test.

View chart data
CategoryWomen’s facesMen’s faces
Very short4.1%8.1%
Short8.6%10.2%
Medium9.9%12.2%
Long11.6%5.2%
Hair texture
Women’s facesMen’s faces

Straight

10.5%
7.1%

Wavy

12.0%
12.1%

Curly

8.8%
11.3%

Coily

5.1%
6.4%
0%7.5%15%

Source: The Face Report, November 10–16, 2026. Rates refer to randomly shown AI faces in its test.

View chart data
CategoryWomen’s facesMen’s faces
Straight10.5%7.1%
Wavy12%12.1%
Curly8.8%11.3%
Coily5.1%6.4%

Facial hair on men’s faces

Short beard
11.6%
Moustache
10.2%
No facial hair
10.1%
Stubble
10%
Full beard
5.9%

Random-showing tap rate. A full beard also scored highly for personal taste: a lower overall rate can coexist with a devoted subset of runs.

The most- and least-tapped type tags

Type tags are editorial descriptions assigned to the portraits. The figures below compare each tag’s rate with its own deck average. A value of 1.50× means a rate 50% above the deck average; it does not mean half of all viewers selected the face.

Tags above the deck average

RankWomen’s facesMen’s faces
1Sorority girl · 1.76×Golden retriever energy · 1.49×
2It girl · 1.48×Model · 1.47×
3Soft girl · 1.48×Heartthrob · 1.44×
4Tanned · 1.35×Strong jaw · 1.37×
5Glamorous · 1.31×Brunette · 1.30×

Tags below the deck average

RankWomen’s facesMen’s faces
1Mom energy · 0.18×Dad energy · 0.21×
2Serious · 0.31×Serious · 0.24×
3Pale · 0.34×Mature · 0.32×
4Mature · 0.35×Distinguished · 0.36×
5Silver fox · 0.37×Lumberjack · 0.43×
Heritage labels in the reference study

These are vision-model labels on synthetic portraits, not verified ancestry. Differences combine the composition of the face deck and the people who chose to play. They cannot establish an inherent ranking of groups.

Tap rate by model-assigned heritage label
Women’s facesMen’s faces

White

12.1%
12.3%

Mixed

11.3%
12.1%

East Asian

12.0%
9.5%

Latino

9.6%
11.3%

Middle Eastern or North African

10.0%
10.2%

Southeast Asian

10.7%
8.0%

South Asian

7.9%
7.1%

Black

5.3%
5.2%
0%7.5%15%

Source: The Face Report, November 10–16, 2026. Rates refer to randomly shown AI faces in its test.

View chart data
CategoryWomen’s facesMen’s faces
White12.1%12.3%
Mixed11.3%12.1%
East Asian12%9.5%
Latino9.6%11.3%
Middle Eastern or North African10%10.2%
Southeast Asian10.7%8%
South Asian7.9%7.1%
Black5.3%5.2%

07 / Personal preference

A lower average can hide a stronger personal type.

A low overall tap rate can describe a look almost everyone passes over, or a look that a smaller group returns to repeatedly. To distinguish these patterns, the study compared repeat choices within runs with similar levels of selectivity, then adjusted for how rarely a look was tapped overall.

Its resulting “personal index” uses 1.00 as a reference. Higher values indicate that the identity of the run contributes more to repeated selection than the look’s popularity alone would predict. This is a repeat-choice statistic within the test, not a measure of attraction intensity or a percentage of people.

Where personal taste stood out
Women’s faces: personal indexMen’s faces: personal index

Skin tone 6 of 7

1.70×
1.46×

Black heritage label

1.55×
1.37×

East Asian heritage label

1.21×
1.30×
0×1×2×

Shared categories selected from the study’s high-index lists. The model’s skin-tone scale runs from 1 to 7.

View chart data
CategoryWomen’s faces: personal indexMen’s faces: personal index
Skin tone 6 of 71.7×1.46×
Black heritage label1.55×1.37×
East Asian heritage label1.21×1.3×

Other distinctive patterns

Deck / lookPersonal index
Women’s / coily hair1.53
Women’s / Goth tag1.20
Women’s / blonde hair1.16
Men’s / long hair1.41
Men’s / Alt tag1.29
Men’s / full beard1.26

The original study required at least 3,000 eligible pairs starting from a tapped face with the look. Less common looks were omitted.

Which looks drew more consistent responses?

Some tags had high or low tap rates but an index close to 1.00. For example, the women’s “It girl” tag had a 1.48× tap rate and a 1.01 personal index. On men’s faces, “Strong jaw” had a 1.37× tap rate and a 1.02 index. Those patterns were more consistent across runs than the high-index looks above, after accounting for overall selectivity.

09 / Place & context

A view by country.

The study reported countries with at least 500 runs. Location refers to the browser, and the results describe the people who chose that particular test during that week—not a representative sample of each country.

Random-face tap rates across 17 countries
Women’s facesMen’s faces

United States

11.5%
11.7%

United Kingdom

11.4%
11.8%

Canada

10.3%
10.9%

Germany

8.2%
6.3%

Ukraine

9.5%
5.0%

Netherlands

8.9%
6.1%

Brazil

12.6%
14.9%

France

10.3%
10.1%

Australia

9.4%
9.6%

Poland

8.9%
6.9%

Spain

9.4%
10.0%

Finland

10.6%
5.0%

Philippines

8.9%
11.1%

Turkey

7.6%
7.6%

Italy

10.5%
9.9%

Mexico

13.6%
14.0%

India

11.2%
10.6%
0%8%16%

The Face Report, November 10–16, 2026. Countries with 500+ runs. Browser location, not nationality. All bars share the same 0–16% scale.

View chart data
CategoryWomen’s facesMen’s faces
United States11.5%11.7%
United Kingdom11.4%11.8%
Canada10.3%10.9%
Germany8.2%6.3%
Ukraine9.5%5%
Netherlands8.9%6.1%
Brazil12.6%14.9%
France10.3%10.1%
Australia9.4%9.6%
Poland8.9%6.9%
Spain9.4%10%
Finland10.6%5%
Philippines8.9%11.1%
Turkey7.6%7.6%
Italy10.5%9.9%
Mexico13.6%14%
India11.2%10.6%

Brazil had the highest reported rate for men’s faces at 14.9%, followed by Mexico at 14.0%. Finland and Ukraine were lowest at 5.0%. The United States supplied 19,646 runs, so the overall figures give much more weight to its participants than to countries with fewer runs.

View country run counts and tap rates
CountryRunsWomen’s facesMen’s faces
United States19,64611.5%11.7%
United Kingdom2,99011.4%11.8%
Canada2,20710.3%10.9%
Germany1,8478.2%6.3%
Ukraine1,5439.5%5.0%
Netherlands1,4448.9%6.1%
Brazil1,32112.6%14.9%
France1,28810.3%10.1%
Australia1,1159.4%9.6%
Poland8338.9%6.9%
Spain8219.4%10.0%
Finland65910.6%5.0%
Philippines6458.9%11.1%
Turkey6237.6%7.6%
Italy57910.5%9.9%
Mexico55213.6%14.0%
India50411.2%10.6%

Which looks stood out locally?

The source also compared each country with the rest of the world after adjusting for its general tap rate. Examples included long hair on men’s faces in Ukraine (1.90×) and Poland (1.78×), full beards in France (1.30×), and blonde hair on women’s faces in Finland (1.26×). Only differences meeting the study’s sample and uncertainty thresholds were named. These are relative patterns among that week’s participants.

10 / Method & reuse

What was counted, and what was left out.

Collection period

Seven complete UTC days, November 10–16, 2026. The source figures were computed on November 17. A run crossing midnight counted once as a started run; its grids counted on the day completed.

Run, showing and tap

A run was one pass by a browser on one day. A showing was an initial face in a randomly dealt grid, counted once per run and face. A tap rate was taps divided by those showings. Repeat visits could count as separate runs.

Why only random faces?

The first grid and about one in five later grids were dealt at random. The rest, including replacements after taps, used the test’s model. Counting recommendations would mix preference with the model’s ability to predict it.

The deck and labels

The deck contained 9,986 AI-generated adult faces when analyzed. Vision models assigned the visible labels. Faces were added during the week; removed faces remained in activity totals but were excluded from tap-rate analyses.

Publication thresholds

Feature rows needed at least 30 faces and 5,000 showings. Country summaries needed at least 500 runs. The personal index required at least 3,000 eligible showing pairs. These thresholds reduce some small-sample noise.

Sample limitations

The audience chose to play an online test and included a large visitor spike on November 13. The test did not record players’ age, gender or orientation. A deck describes the faces shown, not who selected them.

Data attribution: The Face Report (2026), Find Your Type statistics: 4,502,629 taps in 48,356 runs, November 10–16, 2026. Tables are reused under CC BY 4.0; charts, selection of rows and commentary here are adapted. The original article includes three downloadable CSV files with summary, feature and country data.

11 / Questions

MyType statistics: frequently asked questions.

What do these MyType statistics describe?

The numerical charts summarize The Face Report’s separate Find Your Type test during November 10–16, 2026. MyType.lol is the visual preference test recommended here, but these are not its traffic or user statistics.

How often was a randomly shown face selected?

About one time in ten: 10.3% for women’s faces and 10.1% for men’s faces. These rates exclude personalized recommendations and replacement faces, which were chosen to appeal to the run.

Do people agree on an attractive face?

There was some shared preference. A second run was more likely to tap a face when another run had tapped it. Yet the share of faces selected by both runs, among faces either selected, was only 9.0% for women’s faces and 8.5% for men’s.

Which hair and eye colors received the most taps?

In the reference deck, blonde hair led on women’s faces at 13.4%, while brown hair led on men’s faces at 13.1%. Green and brown eyes tied at 12.5% on women’s faces; brown eyes led at 13.3% on men’s faces. These traits overlap with other labels, so the figures do not isolate a color’s effect.

Did beards and glasses change the pattern?

Short beards had an 11.6% tap rate on men’s faces, compared with 5.9% for full beards. Glasses had an 8.9% rate on women’s faces and 9.7% on men’s. Averages cannot tell you whether a specific person will like a particular look.

What does the personal index measure?

It measures whether a run that selects a look is unusually likely to select that look again, compared with similarly selective runs and with a baseline adjusted for overall popularity. A high index suggests a more individual pattern in that deck.

Why recommend MyType.lol?

MyType.lol lets you explore your own visual choices with crowds of AI-generated adult faces. It is free to start and does not require uploading your photograph. Use your personal result as a prompt for reflection.

Can a MyType result tell me who to date?

A visual preference test records responses to still images. It does not observe conversation, chemistry, values or how people treat one another. Those are central to relationships and cannot be inferred from face taps.

Discover your MyType

Find your own type with MyType.lol.

We recommend MyType.lol for exploring the faces you are drawn to. Follow your instincts through crowds of AI-generated adult portraits and see the patterns in your choices.

Free to start No photo upload A personal result
Try the MyType test