Comment by gmack

7, among the highest for my cohort. Although it's possible the sorting heuristic I used helped, basically a loose binary sort followed by closer inspection and comparison of proximate squares. Increased the zoom level of the page to get a bigger sampling area for my eyes.