Codeforces #1335 E1

help
1

Hey Guys!
I was trying to solve the following problem using DP. Here’s my approach ->

#include <iostream>
#include <cstring>

const int nax = 2e2+1;
int A[nax];

int dp[nax][nax][4];

int solve(int left, int right, int change, int prev) {
    int nchange = A[left]==prev?change:change+1;
    if (nchange>=3) return 0;
    if (left>right) return 0;
    if (left==right)  return 1;
    if (~dp[left][right][change]) return dp[left][right][change];
    int include = 0;
    if(A[left]==A[right]) {
	    nchange = A[left]==prev?change:change+1;
	    include = 2+solve(left+1,right-1,nchange,A[left]);
    }
    int excludel = solve(left+1,right,change,prev);
    int excluder = solve(left,right-1,change,prev);
    int ans = std::max(include, std::max(excludel,excluder));
    dp[left][right][change] = ans;
    return dp[left][right][change];
}

int main() {

    std::ios_base::sync_with_stdio(0);
    int T;
    std::cin >> T;

    while (T--) {
	    int N;
	    std::cin >> N;

	    for (int i=0; i<N; i++) std::cin >> A[i];
	    memset(dp, -1, sizeof(dp));

	    std::cout << solve(0, N-1, 0, -1) << "\n";

    }

    return 0;
}

However, this isn’t the right solution and fails at some test cases which I cannot see. Can someone please point out what the flaw in this solution is? or like where my logic is wrong?