A bat and a ball cost $1.10 together, and the bat costs $1.00 more than the ball. Most people's first answer for the ball is 10 cents. That is computation running ahead of reasoning: 1.10−1.00 is quick, correct arithmetic, but it answers a question nobody asked. Reasoning sets the problem up first: if the ball costs b, the bat costs b+1.00, so 2b+1.00=1.10 and b=0.05. The ball costs 5 cents, and the arithmetic was no harder.
The two are different kinds of work. Reasoning answers 'what is this, and why does this method fit?' Computation answers 'what number does the method produce?' Reasoning is where structure pays off: Gauss is said to have added the numbers 1 to 100 as a boy by pairing 1+100, 2+99, and so on, getting 50×101=5050 with one multiplication. Computation is where care pays off: carrying digits, keeping signs, not dropping a term.
Not every problem needs much reasoning. 43+32 has an obvious method, a common denominator, and the only risk is arithmetic: 129+128=1217. Hunting for a clever approach there wastes time. The skill is telling the situations apart: when the method is in doubt, reason first; when it is settled, compute carefully.
Think of it like this. Reasoning is planning a trip; computation is driving it. Careful driving cannot rescue a route to the wrong city, and the best plan still fails if you miss a turn. A math problem lets you check the plan cheaply: a 10-cent ball would make the bat $1.10 and the total $1.20, which exposes the wrong route before any driving.
Core idea. Reasoning chooses which structure and method fit; computation mechanically executes the chosen steps.