How PSL Scores Are Calculated: The Metrics Behind the Number

The Anatomy of a PSL Score

A PSL score is not a single measurement. It is a composite evaluation built from the analysis of multiple facial regions, each assessed on its own merits and then synthesized into an overall rating that accounts for how those regions work together. If you are unfamiliar with the PSL scale itself, our complete PSL scale guide covers the fundamentals.

PSLScore evaluates eight distinct feature categories, each corresponding to a major component of facial aesthetics.

Eye area. The region most consistently identified as the single most important factor in facial attractiveness. This includes the eyes themselves, the surrounding orbital bone structure, brow position, and upper eyelid exposure. The eye area is where many of the most critical individual measurements — canthal tilt, palpebral fissure length, interpupillary distance — originate.

Jawline and lower third. The mandibular region, including jaw definition, gonial angle, chin projection, and the overall proportionality of the lower face. In male faces, a strong jaw is one of the most powerful halos available. In female faces, the ideal tends toward softer, more tapered contours.

Midface. The region between the eyes and the upper lip. Midface length relative to total face length, and relative to lower face length, significantly affects perceived youth and attractiveness. A compact midface is generally considered more aesthetically favorable.

Nose. Bridge width, tip projection, nostril width, and the nasofrontal angle all factor in. The nose sits at the geometric center of the face and has a disproportionate impact on perceived harmony — it does not need to be small, but it does need to fit the face it belongs to. For a deep dive into every nose measurement and what the ideals look like, see our nose analysis guide.

Facial symmetry. The degree to which the left and right sides of the face mirror each other. Research consistently links higher symmetry to higher perceived attractiveness across cultures.

Skin quality. Clarity, texture, evenness of tone, and visible signs of aging or damage. Unlike bone structure, skin quality is one of the most directly controllable factors in your score.

Facial harmony. How well all features work together as a unified whole. A face where every feature is individually average but perfectly harmonious can score higher than a face with one standout feature but poor overall coherence. Our facial harmony analysis guide explores why this relational quality matters more than any single measurement.

Sexual dimorphism. How clearly a face reads as masculine or feminine relative to its sex. Higher dimorphism — stronger masculine features in male faces, more feminine features in female faces — generally contributes positively to scoring.

Each category receives its own sub-score, providing a detailed map of your facial strengths and weaknesses. The overall PSL score synthesizes these into a single number, but the real diagnostic value lies in the breakdown. To understand what each score range means in practical terms, see the PSL rating chart.

Quantitative Measurements

Beneath the eight category scores lies the raw data: specific, measurable facial proportions and ratios that anchor the entire system in quantifiable metrics rather than subjective impression. PSLScore extracts more than fifteen of these measurements from each photo analyzed. Here are the most significant.

Canthal tilt

Canthal tilt measures the angle formed by a line drawn from the inner corner (medial canthus) to the outer corner (lateral canthus) of the eye, relative to the horizontal. A positive canthal tilt — where the outer corner sits higher than the inner corner — is one of the most consistently valued features in PSL assessment. It contributes to what the community describes as "hunter eyes," a term for an eye area with minimal upper eyelid exposure, positive tilt, and deep-set orbital structure.

The measurement is straightforward but its aesthetic impact is substantial. Even a few degrees of difference in canthal tilt can meaningfully change how the eye area is perceived. For an in-depth exploration, read our article on canthal tilt and why it matters.

Gonial angle

The gonial angle is the angle formed at the gonion — the point where the lower border of the mandible meets its ascending ramus, essentially the "corner" of the jaw. A sharper (more acute) gonial angle produces a more defined, angular jawline, while a more obtuse angle creates a softer, less defined jaw contour.

In male facial aesthetics, a gonial angle in the range of approximately 120-130 degrees is generally considered ideal, producing the kind of defined jaw that functions as a powerful halo. Angles significantly above 140 degrees tend to appear weak or undefined. For female faces, slightly higher angles (softer jaw contours) are typically favored, reflecting the dimorphic differences in ideal jaw structure between sexes.

Midface ratio

The midface ratio compares the length of the midface (typically measured from the center of the pupils to the upper lip) to the width of the face at the bizygomatic point (the widest point of the cheekbones). A lower midface ratio — indicating a shorter midface relative to facial width — is generally associated with more youthful and attractive facial proportions.

This measurement is particularly important because midface length is one of the features that changes most visibly with aging, as the midface tends to elongate over time due to gravitational effects on soft tissue and changes in bone density. A compact midface is therefore both an aesthetic positive and a signal of youth. For a complete breakdown of midface proportions and FWHR, see our midface analysis guide.

Facial width-to-height ratio (FWHR)

FWHR is calculated by dividing the bizygomatic width (the distance between the widest points of the cheekbones) by the upper face height (the distance from the upper lip to the brow). This ratio has been extensively studied in academic research, where it has been linked to perceptions of dominance, masculinity, and attractiveness.

Higher FWHR values — indicating a wider face relative to its height — tend to be associated with higher perceived masculinity and dominance in male faces. The "ideal" FWHR varies by sex, with male faces generally scoring better with slightly higher ratios and female faces with moderate ratios.

Facial thirds

Classical facial analysis divides the face into three vertical sections: the upper third (hairline to brow), the middle third (brow to base of nose), and the lower third (base of nose to chin). In idealized proportions, these three sections are approximately equal in height.

Significant deviation from equal thirds can indicate specific proportional issues. A long lower third might suggest excess chin projection or vertical maxillary excess. A short upper third might indicate a low hairline. Equal thirds, while not the only path to an attractive face, represent the balanced proportionality that scores well across multiple facial analysis frameworks.

Interpupillary distance

The distance between the centers of the pupils, typically evaluated relative to overall face width. Interpupillary distance affects how the eye area is framed within the face and contributes to perceived facial proportionality. Eyes that are too close together or too far apart relative to the face width can detract from overall harmony.

Nose width ratio

Nose width measured at the widest point of the nostrils, evaluated relative to intercanthal distance (the distance between the inner corners of the eyes). Classical proportional ideals suggest that nose width should approximately equal intercanthal distance, though the actual ideal varies across different facial structures and ethnic backgrounds. For a comparison of how proportional frameworks like the golden ratio stack up against PSL analysis, see our article on golden ratio vs PSL.

Chin projection and ratio

How far the chin extends forward relative to the vertical line dropped from the base of the nose. Adequate chin projection is important for facial profile balance and jawline definition. Chin-to-nose proportionality — the relationship between chin height, nose length, and overall lower third proportions — also contributes to scoring.

The Role of Facial Harmony

Individual measurements are necessary but not sufficient for calculating a PSL score. Two faces can have similar individual metrics and yet look quite different — and score quite differently — based on how those features work together.

Facial harmony refers to the overall coherence of facial features as a unified composition. It is the difference between a face that looks "right" at a glance and one where something feels slightly off even if you cannot immediately identify what. Our facial harmony analysis guide explores this concept in depth, including why harmony often matters more than any individual feature. In practice, harmony evaluation considers several factors.

Proportional consistency. Do the features scale appropriately relative to each other? A large nose can look harmonious on a face with broad bone structure and wide-set eyes, but the same nose on a narrow face with close-set eyes creates a proportional mismatch.

Feature compatibility. Some feature combinations are inherently more cohesive than others. A strong jaw paired with a strong brow ridge reads as powerfully masculine. The same jaw paired with a delicate, narrow brow can look incongruous. Harmony evaluation captures these interactions that individual measurements cannot.

Aesthetic flow. How smoothly the eye tracks across the face. Features that create visual interruptions — abrupt transitions in proportion, isolated features that do not match the surrounding structure — detract from harmony even if each feature is individually acceptable.

This is also where the concepts of halos and failos become mathematically relevant. A halo — a feature so strong it elevates the entire face — works precisely because it creates a focal point that pulls the overall perception of harmony upward. A failo works in the opposite direction, creating a point of visual disruption that anchors the overall perception downward. The PSL rating chart reflects these dynamics in the score ranges, where the gap between "above average" and "attractive" is often the presence or absence of harmonious feature interaction rather than any single superior measurement.

How PSLScore's AI Analysis Works

Traditional PSL rating required posting photos in forum threads and waiting for experienced raters to evaluate you manually. This process was slow, inconsistent, and publicly vulnerable. PSLScore modernizes the entire approach using AI-based facial analysis.

Landmark detection

The first step in PSLScore's analysis pipeline is automated facial landmark detection. Using computer vision, the system identifies key anatomical points on your face — the corners and centers of the eyes, the tip and bridge of the nose, the borders of the lips, the contours of the jawline, the brow position, and dozens of other reference points. These landmarks serve as the anchor points from which all measurements are derived.

Modern facial landmark detection is highly precise, typically identifying points to within a few pixels. This precision is what enables the consistent, reproducible measurements that differentiate AI analysis from human rating.

Measurement extraction

With landmarks identified, the system calculates the full suite of ratios and measurements described above: canthal tilt angles, gonial angle estimates, midface ratio, FWHR, facial thirds proportionality, interpupillary distance, nose width ratio, chin projection, and more. Each measurement is computed algorithmically from the landmark positions, eliminating the variability that comes with human estimation.

Category scoring

The raw measurements are then fed into scoring models that evaluate each of the eight feature categories. These models incorporate research on facial aesthetics, proportional ideals, and the statistical distributions of each measurement in the general population. A measurement is not simply rated as "good" or "bad" — it is evaluated relative to where it falls on the distribution and how it interacts with other measurements in the same category.

Composite scoring and harmony evaluation

Individual category scores are synthesized into the overall PSL score through a process that accounts for inter-feature harmony. This is where the system evaluates how features work together, applying the principles of proportional consistency and feature compatibility described above. The result is a single number that reflects both the quality of individual features and how cohesively they combine.

Consistency advantage

The most significant advantage of AI analysis over human rating is consistency. The same photo submitted to PSLScore will always produce the same scores and measurements. There is no variation from mood, time of day, comparison effects, or personal preference. This makes AI analysis ideal for tracking changes over time — if you implement a softmaxxing routine and want to measure its impact three months later, you need a baseline that will not shift on its own. Want to see these metrics applied to your own face? Try the PSL calculator to get your measurements and scores.

Limitations and Caveats

No facial analysis system — human or algorithmic — is without limitations, and intellectual honesty requires acknowledging them clearly.

Photo dependence

AI analysis works from photographs, and photographs are not a perfect representation of a face. Lighting, angle, lens focal length, and distance from the camera all affect how facial proportions appear in an image. A photo taken with a wide-angle lens at close range will distort facial proportions (making the nose appear larger and the ears smaller), while a photo taken from further away with a telephoto lens will compress depth. For the most accurate results, use consistent photo conditions: even lighting, a neutral expression, the camera at face level, and a moderate distance (roughly arm's length or slightly further). Our guide on taking the best photo for an accurate PSL rating covers every variable that matters.

Two-dimensional analysis of three-dimensional structure

A photograph is a 2D projection of a 3D structure. Certain features — like chin projection, the depth of the orbital bones, and the forward growth of the maxilla — are more difficult to assess from a frontal photo than from a profile view. PSLScore extracts as much information as possible from each image, but some three-dimensional characteristics are inherently better evaluated in person or from multiple angles.

Cultural and demographic considerations

Facial aesthetics research has historically drawn disproportionately from certain demographic populations. While some proportional preferences (symmetry, averageness, clear skin) appear consistently across cultures, others are more culturally specific. The "ideal" nose shape, jawline contour, and facial fullness all vary across cultural contexts. PSLScore's models work to incorporate diverse reference data, but the system's assessments should be understood within this context rather than taken as universal truth.

What scores cannot capture

A PSL score measures facial aesthetics. It does not measure overall attractiveness, which involves grooming, body language, voice, style, personality, and countless other factors. It does not measure your worth as a person. And it does not account for the way a face looks in motion — the way someone smiles, how their expression shifts during conversation, or the animation that makes a living face fundamentally different from a photograph. These limitations are inherent to any static facial analysis system and are important to keep in mind when interpreting results.

Frequently Asked Questions

How accurate are AI-based PSL scores?

AI-based PSL scores offer a specific kind of accuracy: measurement consistency. The landmarks are detected the same way every time, the ratios are calculated with the same formulas, and the scoring models apply the same criteria uniformly. This eliminates the inter-rater variability that plagues human assessment, where the same face might receive a 4.5 from one rater and a 5.5 from another based on their mood, personal preferences, or recent exposure to other faces. What AI cannot claim is "objective truth" about attractiveness — no system can, because attractiveness is inherently multidimensional and partially subjective. AI-based PSL scores are best understood as highly consistent, well-calibrated assessments of facial proportions and harmony according to a defined framework. They are more reliable than casual human ratings, more reproducible than expert ratings, and particularly valuable for tracking changes over time.

Do different PSL tools give different scores?

Yes, and this is expected. Different tools may use different landmark detection algorithms, different measurement methodologies, different scoring models, and different weighting schemes for how individual features contribute to the overall score. One tool might weight the eye area more heavily than another. One might use slightly different reference proportions as ideals. The result is that a score of 5.0 on one platform may not correspond to exactly 5.0 on another. This is why comparing scores across platforms is not particularly meaningful. What is meaningful is using a single, consistent tool to track your own changes over time. The absolute number matters less than the trend — and a consistent tool will accurately reflect whether your facial aesthetics are improving, declining, or staying the same.

Can makeup or grooming change your measured ratios?

Makeup can alter the perceived proportions of facial features but generally does not change the underlying structural measurements that PSL analysis relies on. Contouring can create the illusion of a sharper jawline or higher cheekbones, and lip liner can change apparent lip proportions, but the landmark positions used for measurement — the actual corners of the eyes, the bony margins of the jaw, the bridge of the nose — remain in the same physical location. That said, some grooming choices can influence certain measurements. Eyebrow grooming affects brow position and shape, which can alter the apparent height of the upper third and the framing of the eye area. Facial hair can obscure jawline definition, making mandibular angles harder to assess. And hairstyle can influence perceived face shape and proportional framing. For the most consistent baseline measurements, photographs with minimal makeup and a pulled-back hairstyle that reveals the full facial structure will produce the most accurate results.

Why does PSLScore analyze 8 features instead of just a few?

Facial attractiveness is not determined by any single feature — it emerges from the interplay of multiple facial regions working together. Analyzing only the eyes and jaw, for example, would miss the contribution of midface proportions, nose harmony, skin quality, and overall symmetry. Someone with a strong eye area and jaw but a long midface and poor skin would be inaccurately assessed by a system that only measured two features. The eight-category approach provides a comprehensive picture that captures both individual feature quality and the harmony between features. It also produces the actionable detail that makes PSL analysis useful for self-improvement: knowing that your eye area scores well but your skin quality is below average tells you exactly where to focus your efforts. A single overall number without the feature breakdown would tell you where you stand but not how to move forward.

How often should I recheck my PSL score?

The optimal frequency depends on what you are doing between checks. If you are actively implementing changes — a new skincare routine, a body recomposition phase, a change in grooming approach — checking every three to six months gives enough time for changes to produce measurable results. Skin cell turnover takes roughly 28 days, meaningful body composition changes require at least two to three months of consistent effort, and the cumulative effects of a softmaxxing routine compound over several months. Checking more frequently than every three months risks measuring noise (day-to-day variation in lighting, hydration, sleep) rather than signal (actual structural or quality changes). If you are not actively making changes and simply want to monitor over time, an annual check is sufficient to capture age-related shifts and the effects of maintained routines. When you do recheck, use consistent photo conditions — same lighting, same angle, same distance, same expression — to ensure that differences in your score reflect actual facial changes rather than photographic variation.