The joke in a bobblehead-style sports meme is simple: one person in a real match photo keeps their full-size head while their body shrinks to a fraction, and everyone around them stays normal. The hard part is keeping it a photo. Scale the person by hand and the edges go soft, the shadow points the wrong way, and the feet float. This article builds the edit in CapCut from a single 16:9 photo, measures what the tool actually did to the proportions, shows the two prompts that turned the person into a toy, and reports what happened when the still was animated.
The finished still: same photo, same pose, same light, with the center player's body scaled down under his original head.
Adults photograph at roughly one head to four body lengths, measured from chin to sole. The meme lives around one to three and one to two and a half: the head reads as a real head, the body reads as a doll's, and the person still stands on the ground next to teammates who prove the scale. Push past that and the figure stops reading as a person at all, which is the failure this article will show later. So the target is a number, and the rest of the process is about asking for it and then checking whether it arrived.
The edit only works on a photo that contains three things. Someone at normal size in the same frame, because a shrunk body with nobody beside it is just a short person. A visible shadow on the ground, because the shadow is what the tool has to redraw and what a viewer checks first. And a subject who is fully in frame, feet included, since the contact with the ground is where the illusion is won or lost. The source used here is a generated stand-in for the reader's own photo: three players in plain teal kits, no numbers, no badges, celebrating on a pitch with a low sun from the left and long shadows on the grass. It was produced in the same composer as the edit, at 16:9, so that the edit and its source share one aspect ratio from the start.
The edit is a follow-up in the same chat, with the source photo attached as the reference. It names what must not change, gives the body one number, and spells out the ground contact and the shadow.
Edit this photo the same way but milder: shrink the center player's body, arms and legs to about half of their current size while keeping his head at its original size, so he stands at about his teammates' hip height. Keep everything else exactly as it is: the two teammates, the stadium, the low sun from the left, his teal jersey, white shorts, his running pose with arms spread and open-mouth shout. His feet stay on the grass at the same spot; redraw his shadow to match the smaller body. Realistic photo, same phone-camera look, no text.
Two things about this wording matter more than the number. It leaves out the words bobblehead, figure, and toy, and the reason is in the failure section below. And it says keeping his head at its original size rather than asking for a bigger head: the head is the one thing the tool should leave alone, so the prompt describes it as fixed and puts all the change on the body.
The prompt asked for half. Measured on the frame, from the top of the hair to the chin and from the chin to the sole, the source player is about one head to four body lengths and the result is about one to three. In other words, the request for half landed inside the range the meme uses, and the body did not become half of anything; the tool changed the proportion, not the arithmetic. A second follow-up on that result, asking for about 20 percent smaller again, moved it to roughly one to 2.8 and kept the running pose. The doll in the last panel is the failure explained below, measured for comparison at about one to two.
Guide lines at hair, chin, and sole. Read the ratio off the frame after every pass instead of assuming the prompt's number was applied.
Source and result with the center player enlarged. The face, the pose, the kit, and the direction of the shadow carried over; only the body scale changed.
The first attempt asked for one third and used the phrase like a bobblehead figure. It came back as a vinyl figurine on a round black base, with a molded cartoon head in place of the man's face. The word did the damage: the tool read it as an object to draw rather than a proportion to reach. The second attempt removed every toy word, asked for one third again, and named the real head, the real kit, and the boots on the grass. The head survived this time, but the body became a plush doll with stubby arms and a knitted jersey, and the spread-arm pose was lost. In these runs, asking for one third in one step pushed the body out of photography twice, while asking for half and then 20 percent more kept it real.
Both one-third requests, with the center player enlarged. Left, the object the word bobblehead produced. Right, a real head on a doll's body.
Before posting, check four places at full size. The edge where the small body meets the grass should show individual blades in front of the boots. The shadow should leave from the feet in the same direction as everyone else's. The head should be the person's own head, not a redrawn one. And the hands should have five fingers each, since the spread-arm pose puts both hands in plain view. If you animate the still, add a fifth check: scrub the clip and look at the face in every frame, because the face is what the motion attacks first.