Skip to content
Lizely

Spot the Difference

Compare two original harbor scenes and locate five precise visual changes.

Privacy: your files never leave your device. All processing happens locally in your browser.

How to use

  1. 1.Compare the Original harbor scene with the Changed scene, paying attention to the kite, lamp, boat, fish, and crate details.
  2. 2.Select a difference on the right scene, or use the labeled A through E inspection buttons for a fully keyboard-playable route.
  3. 3.Find all five differences with no misses for exactly 1,000 points; one distinct miss is recoverable and a second ends the file.

About Spot the Difference

Spot the Difference is a compact visual comparison game built from an original Lizely harbor illustration. The page draws both scenes locally with SVG and CSS. It does not upload a photo, fetch stock artwork, depend on a third-party image library, or randomize target locations. The original and changed panels share a fixed 320 by 220 coordinate system, so every visible change and accepted hit area can be audited precisely.

Five details differ. The kite tail changes near zone A, the lamp window changes near zone B, the boat flag points differently near zone C, the fish bubbles change near zone D, and the crate mark changes near zone E. These are not vague whole-panel guesses. Each target has an explicit center and circular radius. The logic accepts a point when its squared distance from a target center is less than or equal to the radius squared. A point exactly on the boundary counts; a point just beyond it does not.

To play with a pointer or touch screen, compare the left Original panel with the right Changed panel and select a changed detail on the right. Found differences receive a visible marker and cannot be counted twice. The layout preserves both panels as responsive vectors, avoiding blurry bitmap scaling. The Changed scene translates the pointer position back into the fixed coordinate system before checking it, so the same geometry applies on desktop and mobile widths.

A complete keyboard route is available beneath the scenes. The A through E buttons correspond to the five disclosed inspection zones. Tab to a zone and press Enter to inspect it. A found zone locks, and activating A, B, C, D, then E completes the same production logic as selecting the five centers with a pointer. This is a playable alternative, not a decorative accessibility note.

Each new find adds 200 points. Five finds with zero misses produce exactly 1,000 points. The first distinct miss is recoverable and applies a disclosed 100-point deduction, so you can continue searching rather than losing the scene immediately. Misses are normalized to rounded scene coordinates; selecting the same normalized empty point again is an atomic no-op and cannot manufacture a second mistake. A second distinct miss deadlocks the file. Completion and deadlock both freeze further interactions until Restart restores the original fixed state.

Independent tests do more than call the production target table and agree with it. They own a separate literal fixture for all five centers and radii, check the center, exact boundary, and 0.01-unit outside point for each target, assert five unique IDs and five unique zones, and replay both pointer and keyboard winning paths. Other tests confirm invalid coordinates do nothing, repeated finds do nothing, a first miss remains recoverable, the second distinct miss ends the run, and a perfect route scores exactly 1,000.

The puzzle rewards careful comparison of this one authored scene. A useful search method is to divide each panel into bands, compare silhouette before color, and check small directional marks such as flags, tails, bubbles, and symbols. Avoid clicking broad empty areas merely to probe for invisible hit boxes; every difference is visibly drawn in the changed scene.

The result is entertainment, not a standardized vision, attention, memory, cognitive, educational, employment, or health assessment. Screen size, zoom, color rendering, input accuracy, and familiarity with the fixed scene can affect a run. The disclosed geometry and repeatable Restart are provided for transparency and reliable play, not to assign a general ability score.

Methodology & sources

Author two fixed 320 by 220 SVG harbor scenes and five literal difference targets with unique IDs, zones A through E, centers, and circular radii. Accept a point when squared distance is less than or equal to radius squared, reject non-finite and out-of-scene coordinates atomically, and deduplicate found targets. Round miss coordinates to a stable pixel signature, permit one distinct recoverable miss with a 100-point deduction, and deadlock on the second. Independently assert every center, exact radius boundary, and 0.01-unit outside point, verify unique IDs and zones, complete both pointer-center and keyboard-zone routes for exactly 1,000 perfect points, and freeze both terminal states.

Frequently asked questions

Does this game use uploaded or stock images?
No. Both 320 by 220 harbor scenes are original fixed vectors drawn locally with SVG and CSS, with no external artwork or network image request.
How precise does a pointer click need to be?
Each difference has a disclosed circular hit radius. The exact boundary counts as a hit, while a point outside the radius is a miss.
Can I play without a mouse?
Yes. Tab to the A through E inspection controls and press Enter. They activate the same tested targets and produce the same perfect result.

Related tools

Mini Games guides

View all