You figure out your body shape by taking three measurements — bust (or chest), waist, and hips — and comparing the ratios. If your bust and hips are within an inch of each other and your waist is at least 25% smaller than both, you're an hourglass. If your hips are noticeably wider than your bust, you're a pear (triangle). Wider on top? Inverted triangle. Roughly equal all the way down? Rectangle. Carrying the volume in the midsection with narrower hips? Apple (oval).
That's the snippet answer. Below is the longer version — including why "body shape," "body type," and "physique" are three different things people keep conflating, what the actual research says about prevalence, and how AI does the same ratio math without a tape measure.
This is the cleanest way to keep them straight:
The Google search "what are the 4 body shapes" is one of those well-loved typos. There are five standard silhouette categories in the fashion literature (Healthline's overview lists more than four), and the FFIT classification system used in apparel research actually distinguishes seven to nine (Lee et al., 2007, using the SizeUSA database). Four is a folk-memory simplification of five.
These are the categories almost every fashion source uses, defined by the relationships between three measurements. We'll cover prevalence using the SizeUSA dataset — a database of body scans of more than 6,300 American women, processed using the Female Figure Identification Technique developed by Simmons et al. and applied by Lee et al. (2007).
Bust and hips are within roughly an inch of each other; waist is at least 25% smaller than both. Waist-to-hip ratio typically 0.70–0.75. The famous "36-24-36" numbers describe a textbook hourglass — so do "34-26-36" (waist-to-hip ratio of 0.72), which is why that specific search keeps coming up.
In the Lee et al. SizeUSA analysis, the hourglass category as strictly defined appeared in about 11.8% of the sample, with two near-hourglass variants ("top hourglass" and "bottom hourglass") adding another 12%.
Hips are noticeably wider than the bust — usually by more than 5%. The waist is defined but the silhouette widens toward the bottom. SizeUSA's "triangle" classification appeared in about 4.8% strictly defined, but the related "spoon" shape (similar but with a slightly larger gap between waist and high hip) covered another ~21%.
The waist is the widest measurement, with the bust and hips narrower or roughly equal to it. Most weight sits in the midsection. The FFIT system calls this the "oval" shape. The classic apple silhouette is more common with age — somatotype studies show endomorphic and mesomorphic profiles both increase in older adults, and weight redistribution toward the abdomen is a well-documented postmenopausal pattern.
Note: some popular guides use "inverted triangle" to mean the same thing as apple (broad on top, narrow on bottom). That's a different shape in the fashion-research literature. We're using "apple" for waist-dominant and "inverted triangle" for shoulder-dominant. If a quiz disagrees, check what ratios it's actually measuring.
Bust, waist, and hips are roughly within 5% of each other. The waist isn't significantly smaller than the bust or hips, so the silhouette reads as straight up and down. This is the most common shape in the SizeUSA dataset by a wide margin — about 46–49% of women. If you've ever wondered why the "hourglass = standard" assumption felt off, this is why.
Bust or shoulders are wider than the hips by more than ~5%. Common in athletes who build upper-body muscle, but also a natural skeletal pattern for many people. Rare in the strict FFIT classification of the SizeUSA female sample (about 0.5%), though more common in male body-shape datasets where shoulder breadth tends to exceed hip breadth.
A note before the measuring section: most people are not pure examples of one shape. "Rectangle leaning hourglass" or "pear-rectangle" is normal. The categories are buckets imposed on a continuous distribution.
You need a soft fabric tape measure, a mirror, and ideally another set of hands. Wear thin clothes or none — bulky fabric throws off every measurement.
Then match against the ratios:
| Shape | Rule |
|---|---|
| Hourglass | Bust ≈ hips (within ~1 inch). Waist at least 25% smaller than both. |
| Pear (triangle) | Hips > bust by more than 5%. Defined waist. |
| Apple (oval) | Waist measurement ≥ bust and hips. Less defined waist. |
| Rectangle | Bust, waist, hips all within ~5% of each other. |
| Inverted triangle | Bust > hips by more than 5%. |
Worked example for "34-26-36": bust 34 ≈ hips 36 (within 5%), waist 26 is 23.5% smaller than bust and 27.8% smaller than hips, waist-to-hip ratio 0.72. That meets the hourglass criteria. The waist-to-bust reduction sits just under the 25% threshold some systems use, so depending on which calculator you run it through, "34-26-36" can also classify as "near-hourglass" or "bottom hourglass." Both reads are defensible.
Body type is a separate framework — somatotype theory. William Sheldon proposed in 1940 that humans cluster into three constitutional types: endomorph (rounder, higher body fat tendency), mesomorph (muscular, broader shoulders), and ectomorph (lean, narrow frame). He paired each with a personality profile, which is the part that has been thoroughly discredited and is connected to early-20th-century eugenics.
What survived is the morphological component, reworked in the 1960s and 70s by Heath and Carter into the Heath-Carter anthropometric somatotype method. The modern version uses ten measurements — three skinfolds (triceps, subscapular, supraspinale), two bone breadths (humerus, femur), two limb girths (arm, calf), height, and weight — to produce three scores from roughly 1 to 7. Your somatotype is the triple, like "3-5-2" (low endomorphy, high mesomorphy, low ectomorphy = textbook muscular build).
This is the framework still used in sports science and athlete profiling. Elite endurance athletes cluster toward higher ectomorphy; throwers and weightlifters toward higher mesomorphy. The categories are continuous, not discrete — a 2025 study of 341 young adults concluded that somatotype "should not be interpreted as discrete and homogeneous groups, but as positions along a continuum." In that study, men predominantly showed an Endomorphic-Mesomorph profile and women Mesomorph-Endomorph — pure ectomorphs are statistically rare in adult populations, despite how often "I'm an ectomorph" gets used as fitness shorthand.
AI body shape detection runs a three-step pipeline:
The technical pipeline is solid. The failure modes are about the photo:
Want the AI read on your specific photo? Our Body Shape Analyzer runs the full pose-and-ratio analysis on a single image and returns a category with confidence. Free, no signup, instant. If you're asking a different question — fitness aesthetic rather than silhouette category — the Rate My Physique tool grades muscular development and proportions instead. Pick the one that matches what you're actually trying to learn.
Body shape is a geometry observation. It correlates with some things and not with others. Here's what the evidence supports:
Real, with evidence: Body shape affects which silhouettes of clothing fit well. Apparel research literally uses these classifications to design sizing systems — Lee et al.'s SizeUSA work was motivated by the fact that standard sizing (which assumes an hourglass-shaped target) only fits the actual hourglass minority well, which is why so many garments fit weirdly on so many people.
Modest, contested: Waist-to-hip ratio shows up in attractiveness research as a cross-cultural signal. The Singh studies from the 1990s found a preference for WHR around 0.7 across multiple cultures. More recent work challenges the universality — preferred ratios shift by culture and BMI, and composite "curviness" outperforms WHR alone as a predictor.
Not real: Body shape doesn't determine personality, intelligence, or "type." Sheldon's original personality claims came out of the same intellectual lineage as physiognomy and eugenics and have not survived modern research. It also doesn't determine clothing size — size is absolute measurements, shape is ratios. A size 4 hourglass and a size 16 hourglass share the same shape.
This is the body-shape-adjacent number with actual clinical relevance. Waist-to-hip ratio (WHR) is calculated as waist circumference divided by hip circumference. It's a proxy for visceral fat — fat stored around organs, which is much more metabolically harmful than subcutaneous fat.
The WHO expert consultation on waist circumference and waist-hip ratio set the thresholds for elevated metabolic risk at:
A 2024 meta-analysis in Frontiers in Cardiovascular Medicine found elevated WHR was associated with a roughly doubled odds of myocardial infarction (pooled OR 1.98), with each 0.01 increase in WHR associated with about a 2% increase in cardiovascular risk. WHR has been shown to outperform BMI as a linear predictor of mortality in middle-aged adults.
This is the part of body shape that's actually about health. The hourglass vs rectangle vs pear question is a styling question. WHR is the metric your doctor cares about.
A few claims that keep cycling and are mostly wrong:
"Hourglass is the rarest body shape." In the SizeUSA data, strict hourglass appears in about 12% of the female sample. Inverted triangle is rarer at under 1%. Rectangle is by far the most common at around half.
"You can change your body shape with exercise." Partially. Your skeletal frame — shoulder width, ribcage shape, pelvic width — is fixed in adulthood. What you can change is fat distribution (somewhat) and muscle development (significantly), which shifts apparent proportions. A rectangle can build glutes and approach an hourglass-adjacent silhouette. The category your tape measure returns can move; the bones underneath don't.
"There are only 4 body shapes." Standard fashion taxonomies use 5. The FFIT system used in apparel research uses 7–9. The "four shapes" framing doesn't appear in the underlying literature.
"Ectomorphs can't build muscle." Fitness version of body-shape determinism. Somatotype is a continuous descriptor of current body composition, not a fixed cap on what you can change.
Want the AI read on your own silhouette? Start with the Body Shape Analyzer for silhouette categorization. For the fitness/aesthetic side, use Rate My Physique. For the rest of the analytical stack, Face Shape Analyzer and Facial Harmony handle the face. All free, no signup.