Four prompts written as one family: same woman, same outfit, same railway yard, each changing only the pose and the crop. Four pictures, one per prompt, and this time every picture matches its paragraph.
All four landed, in order. One: a knee raised onto the platform edge, the same-side hand on the thigh, looking at the lens. Two: arms hanging, head turned to the side, and the “simple wrist accessories” only that paragraph mentions turned up as bracelets. Three: full length from behind, torso twisted back over the shoulder, with the open back, white socks and cream sneakers that only paragraph three asks for. Four: the same twist, cropped above the knee. Each detail showed up in the run whose paragraph named it.
This is finding one again, from another direction. Splitting a shoot into separate, near-identical prompts does the same thing as numbering a list: each change sits in its own block where nothing else competes with it. The nine-panel collage did it inside one prompt. This does it across four.
Holds across all four: the white high-neck long-sleeve crop top, sage-green shorts, light-brown low ponytail, the commuter train and tower blocks behind, deep focus, and 9:16.
Needs a reference image. Every paragraph says “match the reference image” for the background and composition. The same face across four separate prompts is the reference doing that work, not the text.
