Cup Arc

Landing throws in Cup Arc

Aim, power, loft — and a descending rim crossing

The table runs 690 units away from you and 570 across, and the ball is thrown from a fixed origin near your edge. Aim swings up to 24 degrees either side of straight ahead. Horizontal speed and vertical launch speed are each adjustable within their own fixed ranges, and gravity acts on the ball for the whole flight.

A cup only counts if the ball crosses its rim while descending. A flat, fast throw that reaches the cup on the way up passes over it; the arc has to peak before the target. That is why loft is a separate control rather than something derived from power — the two together decide where the peak of the flight sits.

Layouts change between rounds. Cup counts and arrangements vary, including diamond and line patterns, so the distances you are throwing to are not the same from one round to the next.

Fix two variables and adjust one

Three interacting inputs is one too many to tune by feel all at once. The reliable method is to settle on a loft you like, leave it alone, and treat horizontal speed as your only distance control. That reduces every throw to a single-variable problem you can correct systematically: short means more power, long means less.

Aim is best handled separately and first. Line up the cup, throw, and see whether you missed left or right before touching power at all — otherwise you cannot tell whether a miss came from direction or distance, and you end up adjusting both and learning nothing.

Higher loft is more forgiving on distance because the ball comes down more steeply, which widens the window in which it is descending across the rim. Flatter throws travel further for the same power but demand much more precise distance control. When a layout puts cups close together, loft up and accept the shorter reach.

The three opponents

The computer opponents differ in four ways at once: how precisely they aim, how consistently they hit their intended power and loft, how often they simply make a mistake, and how many candidate throws they search before choosing one.

The gap is wide. The easy opponent aims with roughly nine times the angular error of the hard one and misfires about one throw in five, while the hard opponent searches seven candidate throws and misfires around one in forty. That is a difference in kind rather than a difficulty slider.

Common mistakes

The recurring error is changing more than one thing between throws.

• Adjusting power and loft together after a miss. You lose the information the miss gave you, because you can no longer tell which change was responsible.
• Throwing flat at distant cups. A low arc can reach the cup while still climbing, and a ball crossing the rim on the way up does not drop in no matter how accurate the line.
• Correcting aim and distance in the same throw. Establish the line first; distance is much easier to tune once you know you are pointed at the right cup.
• Assuming last round’s throw still works. Layouts are regenerated between rounds, so a setting that was perfect for one arrangement is only a starting point for the next.

What is different about this version

The opponent solves for a genuinely valid throw before any error is applied. It searches launch options for one that crosses the target rim while descending, then deliberately perturbs the aim, power, and loft by its own profile’s deviations. It is not rolling for a hit and animating backward from the result, which is why its misses look like plausible throws rather than arbitrary ones.

Because the rim test requires a descending crossing rather than mere proximity, the physics decide every outcome. There is no snap-to-cup assist and no hidden success chance: a throw that would physically drop in does, and one that would skim past does not.

Written by the Silverweb Games team from this game’s own implementation. Last reviewed July 31, 2026.