Problem setters choose the limits so the intended solution passes and the naive one times out. Read the constraints before the story, multiply by the budget, and you know the target complexity before writing a line. Often you also learn which technique to reach for.
A judge gives you roughly 1–2 seconds, and a typical machine does about 10⁸ simple operations per second. Divide that budget by your algorithm's cost and you get the largest n each complexity can survive. Drag across the bars to move n.
Paste the constraints block from any problem, or pick an example, and read it the way a strong candidate does: sizes, value ranges, then the small words. Edit the numbers and watch the reading change.
Some lines don't change the budget. They tell you which tool to use. Tap a card for an example.
Grids, queries, graphs and test cases give you two numbers. The cost depends on how they combine: m·n cells, n·q work, V + E edges. Move both sliders and see which approaches survive.
Pick the target: read real-looking constraints and choose the slowest approach that still passes. Pass or TLE?: 60 seconds, one n and one complexity at a time. Decide fast. Keys: 1–4 for options, ← pass and → TLE, Enter for next.