consider this input: 3 1 1 -1 -1 100 1000

should output: 1000

your solution: -1

As I told you before, greedy approach will not work here.

You can try Dynamic programming to solve this type of problem where greedy doesn't work. But for this there is a brute force solution.

for each number in the array, let's say that this is a_j and we multiply it with q. So we are fixing the middle element every time. Now we have to find a_i and a_k [i ≤ j and j ≤ k] such that a_i * p + a_k * r is maximized. So we can easily get an O(n²) solution. How to make it O(n)? Well the array is static. So we can pre-calculate the minimum and maximum in each suffix and prefix of the array and use them instead of running a loop each time to find a_i and a_k. This is the same solution described in the editorial btw.