#pragma GCC optimize ("Ofast")
#include<bits/stdc++.h>
using namespace std;
void *wmem;
char memarr[96000000];
template<class T> inline void walloc1d(T **arr, int x, void **mem = &wmem){
  static int skip[16] = {0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1};
  (*mem) = (void*)( ((char*)(*mem)) + skip[((unsigned long long)(*mem)) & 15] );
  (*arr)=(T*)(*mem);
  (*mem)=((*arr)+x);
}
template<class T1> void sortA_L(int N, T1 a[], void *mem = wmem){
  sort(a, a+N);
}
template<class T1, class T2> void sortA_L(int N, T1 a[], T2 b[], void *mem = wmem){
  int i;
  pair<T1, T2> *arr;
  walloc1d(&arr, N, &mem);
  for(i=(0);i<(N);i++){
    arr[i].first = a[i];
    arr[i].second = b[i];
  }
  sort(arr, arr+N);
  for(i=(0);i<(N);i++){
    a[i] = arr[i].first;
    b[i] = arr[i].second;
  }
}
template<class T1> void rsortA_L(int N, T1 a[], void *mem = wmem){
  sortA_L(N, a, mem);
  reverse(a, a+N);
}
template<class T1, class T2> void rsortA_L(int N, T1 a[], T2 b[], void *mem = wmem){
  sortA_L(N, a, b, mem);
  reverse(a, a+N);
  reverse(b, b+N);
}
inline void rd(int &x){
  int k;
  int m=0;
  x=0;
  for(;;){
    k = getchar_unlocked();
    if(k=='-'){
      m=1;
      break;
    }
    if('0'<=k&&k<='9'){
      x=k-'0';
      break;
    }
  }
  for(;;){
    k = getchar_unlocked();
    if(k<'0'||k>'9'){
      break;
    }
    x=x*10+k-'0';
  }
  if(m){
    x=-x;
  }
}
inline void rd(long long &x){
  int k;
  int m=0;
  x=0;
  for(;;){
    k = getchar_unlocked();
    if(k=='-'){
      m=1;
      break;
    }
    if('0'<=k&&k<='9'){
      x=k-'0';
      break;
    }
  }
  for(;;){
    k = getchar_unlocked();
    if(k<'0'||k>'9'){
      break;
    }
    x=x*10+k-'0';
  }
  if(m){
    x=-x;
  }
}
inline void wt_L(char a){
  putchar_unlocked(a);
}
inline void wt_L(long long x){
  int s=0;
  int m=0;
  char f[20];
  if(x<0){
    m=1;
    x=-x;
  }
  while(x){
    f[s++]=x%10;
    x/=10;
  }
  if(!s){
    f[s++]=0;
  }
  if(m){
    putchar_unlocked('-');
  }
  while(s--){
    putchar_unlocked(f[s]+'0');
  }
}
template<class S> inline void arrInsert(const int k, int &sz, S a[], const S aval){
  int i;
  sz++;
  for(i=sz-1;i>k;i--){
    a[i] = a[i-1];
  }
  a[k] = aval;
}
template<class S, class T> inline void arrInsert(const int k, int &sz, S a[], const S aval, T b[], const T bval){
  int i;
  sz++;
  for(i=sz-1;i>k;i--){
    a[i] = a[i-1];
  }
  for(i=sz-1;i>k;i--){
    b[i] = b[i-1];
  }
  a[k] = aval;
  b[k] = bval;
}
template<class S, class T, class U> inline void arrInsert(const int k, int &sz, S a[], const S aval, T b[], const T bval, U c[], const U cval){
  int i;
  sz++;
  for(i=sz-1;i>k;i--){
    a[i] = a[i-1];
  }
  for(i=sz-1;i>k;i--){
    b[i] = b[i-1];
  }
  for(i=sz-1;i>k;i--){
    c[i] = c[i-1];
  }
  a[k] = aval;
  b[k] = bval;
  c[k] = cval;
}
template<class S, class T, class U, class V> inline void arrInsert(const int k, int &sz, S a[], const S aval, T b[], const T bval, U c[], const U cval, V d[], const V dval){
  int i;
  sz++;
  for(i=sz-1;i>k;i--){
    a[i] = a[i-1];
  }
  for(i=sz-1;i>k;i--){
    b[i] = b[i-1];
  }
  for(i=sz-1;i>k;i--){
    c[i] = c[i-1];
  }
  for(i=sz-1;i>k;i--){
    d[i] = d[i-1];
  }
  a[k] = aval;
  b[k] = bval;
  c[k] = cval;
  d[k] = dval;
}
template<class S, class T> inline S divup_L(S a, T b){
  return (a+b-1)/b;
}
template<class S, class T> inline S chmax(S &a, T b){
  if(a<b){
    a=b;
  }
  return a;
}
int N;
long long A[52];
long long B[52];
long long C[52];
int s1;
long long m1[1000000];
long long g1[1000000];
int s2;
long long m2[1000000];
long long g2[1000000];
void dfs(int s, int e, long long m = 0, long long g = 0){
  if(s == e){
    arrInsert(s2,s2,g2,g,m2,m);
    return;
  }
  dfs(s+1, e, m+A[s], g);
  if(s+2 <= e){
    dfs(s+2, e, m-C[s+1], g+B[s+1]);
  }
}
void ord(void){
  int i;
  int k = 0;
  rsortA_L(s2, g2, m2);
  for(i=(0);i<(s2);i++){
    if(k && m2[i] <= m2[k-1]){
      continue;
    }
    arrInsert(k,k,g2,g2[i],m2,m2[i]);
  }
  s2 = k;
}
long long solve(int n1, int n2){
  int i;
  int k;
  long long res = 0;
  s1 = s2 = 0;
  dfs(0,n1);
  ord();
  s1 = s2;
  for(i=(0);i<(s1);i++){
    m1[i] = m2[i];
  }
  for(i=(0);i<(s1);i++){
    g1[i] = g2[i];
  }
  s2 = 0;
  dfs(n1,n1+n2);
  ord();
  k = s2-1;
  for(i=(0);i<(s1);i++){
    while(k-1>=0 && m1[i]+m2[k-1] >= 0){
      k--;
    }
    if(m1[i]+m2[k] < 0){
      continue;
    }
    chmax(res, g1[i] + g2[k]);
  }
  return res;
}
int main(){
  wmem = memarr;
  long long res = 0;
  rd(N);
  {
    int KrdatlYV;
    for(KrdatlYV=(0);KrdatlYV<(N);KrdatlYV++){
      rd(A[KrdatlYV]);
      rd(B[KrdatlYV]);
      rd(C[KrdatlYV]);
    }
  }
  chmax(res, solve(N/2,divup_L(N,2)));
  chmax(res, solve(N/2+1, (divup_L(N,2))-1));
  wt_L(res);
  wt_L('\n');
  return 0;
}
// cLay varsion 20200112-1

// --- original code ---
// int N; ll A[52], B[52], C[52];
// 
// int s1; ll m1[1d6], g1[1d6];
// int s2; ll m2[1d6], g2[1d6];
// 
// void dfs(int s, int e, ll m = 0, ll g = 0){
//   if(s == e){
//     arrInsert(s2,s2,g2,g,m2,m);
//     return;
//   }
//   dfs(s+1, e, m+A[s], g);
//   if(s+2 <= e) dfs(s+2, e, m-C[s+1], g+B[s+1]);
// }
// 
// void ord(void){
//   int i, k = 0;
//   rsortA(s2, g2, m2);
//   rep(i,s2){
//     if(k && m2[i] <= m2[k-1]) continue;
//     arrInsert(k,k,g2,g2[i],m2,m2[i]);
//   }
//   s2 = k;
// }
// 
// ll solve(int n1, int n2){
//   int i, k;
//   ll res = 0;
//   s1 = s2 = 0;
// 
//   dfs(0,n1); ord();
//   s1 = s2;
//   rep(i,s1) m1[i] = m2[i];
//   rep(i,s1) g1[i] = g2[i];
//   s2 = 0;
//   dfs(n1,n1+n2); ord();
// 
//   k = s2-1;
//   rep(i,s1){
//     while(k-1>=0 && m1[i]+m2[k-1] >= 0) k--;
//     if(m1[i]+m2[k] < 0) continue;
//     res >?= g1[i] + g2[k];
//   }
// 
//   return res;
// }
// 
// {
//   ll res = 0;
//   rd(N,(A,B,C)(N));
//   res >?= solve(N/2, N/+2);
//   res >?= solve(N/2+1, (N/+2)-1);
//   wt(res);
// }