Two Identical Parallelograms. They Do Not Look Identical.
Which line/shape is bigger?
You are looking at the Shepard Tables illusion, popularised by the cognitive scientist Roger Shepard in his 1990 book Mind Sights. Two parallelograms · drawn to look like the tops of two tables · sit side by side. One reads as a long, narrow table stretching away from you; the other as a shorter, wider table seen from the side. They look radically different. They are geometrically congruent. Trace one onto tracing paper, slide it over, and it lines up perfectly with the other.
What you are about to learn. What the illusion actually is, why it is perhaps the single cleanest demonstration of perspective-based size constancy, the role of the table-legs cue in strengthening the effect, and why spinning a figure by 90 degrees on a page can transform its dimensions in your head.
What the Illusion Looks Like
Draw a parallelogram, perhaps 140 pixels wide by 60 tall, skewed so it appears to recede into the distance. Now draw a second parallelogram of the same width and height, but rotated 90 degrees so that what was the long dimension is now pointing toward you rather than away.
The two parallelograms are congruent (same shape, same size, just rotated). But your brain does not read them as congruent. It reads them as two tables with two distinctly different proportions · one tall and narrow, one short and wide.
The minimal recipe. Two identical parallelograms at orthogonal orientations. The illusion works without drawing the table legs at all · but legs strengthen the effect by forcing the 3D interpretation. A plain outline is enough to produce a 25 to 30 percent perceived dimension difference.
Why It Works
Your brain treats the parallelograms as tabletops viewed from above and at a modest angle. Each parallelogram carries a coherent depth cue: the longer axis of the parallelogram must be the one pointing into the depth of the scene.
Size-constancy machinery kicks in: an edge that points into depth is foreshortened on the retina (objectively shorter) but should represent a genuinely longer physical dimension. So your brain upscales the apparent depth dimension to compensate.
The two parallelograms, rotated 90 degrees from each other, have their depth-axes pointing in orthogonal directions. Your brain upscales different physical edges in each one. Result: the same parallelogram looks long-and-narrow in one orientation and short-and-wide in the other.
This is depth constancy applied to a flat figure. It is the same mechanism as Ponzo, only here the depth cue is internal to each shape rather than external (rails). You are watching your visual system undo perspective foreshortening twice, on two shapes, in two different directions · and the result is two shapes that feel different even though they are not.
The Cover-and-Trace Test
The single best proof that the two tabletops are identical is to cover one with tracing paper, outline it, slide the tracing onto the other, and watch it click into place.
The direct proof. Print the figure. Cut out one tabletop. Rotate it 90 degrees. Overlay it on the other. The outlines match exactly. Your conscious perception, which had the two tabletops read as wildly different, is now staring at the arithmetical proof of their equality. It is one of the most striking sensory experiences in all of illusion-land · because even after the cut-and-overlay proof, looking back at the original figure still shows two different tables.
Why Legs Help
In the classical version with tabletops drawn with legs attached, the illusion is stronger than in the bare-outline version by roughly 10 percent. Why? The legs dispatch an unambiguous depth cue: they must be vertical, descending from the table edge to the floor. Given legs, your brain is fully committed to a 3D interpretation · and any 3D interpretation triggers size-constancy scaling.
Without legs, the parallelograms can in principle be read as flat 2D shapes (they are, after all, on a flat page). A fraction of viewers manage to hold that reading, and for them the illusion is much weaker.
Common misconception: “I can turn off the 3D reading and the illusion goes away.” You cannot, reliably. Even observers explicitly instructed to see the parallelograms as 2D shapes still experience the effect · the 3D reading is automatic and pre-attentive. Training yourself to see flat shapes as flat takes considerable practice and still does not fully eliminate the size-constancy scaling.
The Magnitude
In controlled studies, the Shepard Tables effect sits at 20 to 30 percent · one of the largest size illusions ever recorded. Compare that to Müller-Lyer (15 to 20 percent) and Ponzo (10 to 20 percent). The Shepard routinely beats them because it combines two depth cues (the parallelogram skew and the legs) with a whole-shape geometric difference that forces the brain into its most aggressive scaling mode.
Why Shepard beats Müller-Lyer. Müller-Lyer delivers a depth cue at the line endpoints (the fins). Shepard delivers a depth cue across the whole shape (the parallelogram’s internal geometry). More of the visual input supports the scaling inference, so the scaling is stronger. This is the same “illusions stack when cues agree” principle we met in the Sander illusion.
A Harder Variant
Below is a Shepard Tables figure at difficulty 3 · the parallelograms are sharper, the geometric illusion cleaner. The two tabletops are, as always, congruent.
Which line/shape is bigger?
Where Shepard-Style Illusions Hide
- Furniture catalogues. A table photographed from the short side looks more elegant than the same table photographed from the long side · size-constancy tilts the perceived depth, and elegant catalogues tend to use the short-side view.
- Room planning. When you stand in a rectangular room and look along its long axis, the room feels stretched. When you stand in the same room looking along its short axis, it feels wider. The physical room has not changed; your viewing axis has, and the Shepard-style scaling follows.
- Architectural photography. Photographers working for real estate routinely choose camera angles that make rooms feel longer in the direction the client is trying to sell. The Shepard mechanism is in their compositional toolkit.
- Video-game level design. Corridors drawn in perspective feel longer or shorter depending on the designer’s camera angle. The same corridor, photographed head-on versus obliquely, produces very different spatial impressions.
Test Yourself on 50 More Illusions
The Shepard Tables illusion is one of more than 50 classical illusions on PlayMemorize. Each round draws a deterministic SVG scene and asks one grounded question: which is larger, which is brighter, which is actually parallel. The reveal overlay shows the true geometry plus a one-line “why it works” caption.
- Keep playing Shepard Tables → · the standalone game, pinned to this one figure with fresh seeds each round
- Play Illusions → · spot the tricks across size, colour, orientation, and impossible figures
- Play Spatial → · train mental rotation and area estimation
- Play Matrix → · abstract pattern reasoning under time pressure
Turn your head 90 degrees. Tilt the phone or monitor sideways so the original vertical orientation is now horizontal. The illusion does not vanish · the parallelograms still look different · but which one looks long-and-narrow versus short-and-wide swaps. This is because the depth reading is pinned to the viewer’s vertical axis, not to the figure’s original orientation. You are watching your own perceptual machinery in real time.
The takeaway. The Shepard Tables is a demonstration that size is not a property of objects. Size is a computation your brain runs on the fly, using depth cues that may or may not correspond to physical reality. Two shapes that are congruent become two shapes of “different sizes” the moment a perspective interpretation kicks in. Once you see this clearly, you see it everywhere · in furniture, in buildings, in photographs, in your own estimation of distances and areas. The gap between your measurement and your perception is where illusion · and insight · lives.
Illusions
Your eyes lie - the math knows the truth. Spot equal lengths, identical greys, and truly parallel lines across 57 classic optical illusions
เล่นตอนนี้ - ฟรีไม่ต้องสมัครสมาชิก ใช้ได้ทุกอุปกรณ์
