A visible scoring route is only the first candidate. A useful choice also fits your execution, respects the contact sequence and leaves a position you have considered. There is no universal ranking in which every ticky beats every bricole, or the shortest path always wins.
Ball-first and cushion-first use different event orders
Neither branch wins automatically; choose the clearest contact window and post-contact paths on the actual table.
First exclude an invalid sequence
Identify your cue ball and the order of the two object balls. Before the second object-ball contact, the cue ball must have made at least three cushion contacts. A cushion-first route can count that first cushion: one before the first ball plus two between the balls can satisfy the three-cushion requirement. It does not automatically need three more after the first ball.
Then check access, a frozen ball or cushion, and the ability to strike without a push, double hit or wrong-ball foul. At the UMB opening, the direct red-first requirement removes options that might be available in an ordinary position.
Compare two usable routes
For each candidate write the first ball, contact thickness, cushion order, intended arrival and likely extra-collision area. Note where a small execution error matters: a thin first-ball contact, a narrow passage or a pace-sensitive third-cushion arrival. A longer path may solve one problem while adding others.
Follow the first object ball as well as the cue ball. A clear line on a static diagram can still produce a collision after that ball moves. If two choices look similar, prefer the one you can reproduce under the present table conditions, not the one with the more impressive name.
A worked comparison without false probabilities
Suppose, in a practice exercise, route A scores 6 of 10 attempts and B scores 5 of 10. A also suffers three unwanted collisions; B has none. These are invented example counts, not measurements from 3Ball or proof that A has a true 60% match probability. Such a small sample does not settle which route is better.
Retest the collision-prone part of A, keep both starting layouts identical and record why each miss happened. If you change the pace or cut, start a new comparison rather than mixing results. Add the final ball locations to the sheet; a score count alone hides important differences.
Look from the correct ball after a miss
If you are playing white, the incoming opponent plays yellow; your white becomes one of their object balls. Assess the leave from yellow, not from the ball you just struck. Large separation or a ball near a cushion is not automatically a difficult leave: inspect the actual available route.
Score, remaining target and event format affect the decision, but do not turn a guessed percentage into an expected-value formula. If you have no measured probabilities, describe the trade-off plainly: repeatability, collision risk, required speed and the resulting position.
Use the replay as evidence of a model run
In 3Ball keep the position fixed, try each candidate and inspect its replay. Compare first contact, cushion count and the final layout. A solver finding a route does not establish your physical success rate or prove that every alternative was searched. Confirm a promising candidate on the real table and retain separate notes for simulator and physical attempts.