#include <bits/stdc++.h>
using namespace std;
const int maxn = 1<<18;
vector<int> seg(maxn<<1, 0);
int q(int a, int l, int r, int lq, int rq) {
if (l >= lq && r <= rq) return seg[a];
int mid = (l+r)/2;
int x = 0;
if (mid > lq) x = q(a<<1, l, mid, lq, rq);
if (mid < rq) x = max(x, q(a<<1|1, mid, r, lq, rq));
return x;
}
void upd(int a, int v) {
for (a+=maxn; a > 0; a>>=1) seg[a] = max(seg[a], v);
}
int main() {
int N, X; cin >> N >> X;
vector<int> T(N);
for (int i = 0; i < N; i++) cin >> T[i];
vector<int> comp = T;
sort(comp.begin(), comp.end());
comp.erase(unique(comp.begin(), comp.end()), comp.end());
unordered_map<int, int> end;
for (int i = 0; i < N; i++) end.insert({comp[i], i});
// Right way LIS
vector<int> pref(N);
for (int i = 0; i < N; i++) {
pref[i] = q(1, 0, maxn, 0, end[T[i]])+1;
upd(end[T[i]], pref[i]);
}
// inverse LDS
vector<int> S = T;
reverse(S.begin(), S.end()); reverse(pref.begin(), pref.end());
vector<int> dp(N+1, -1);
int l = 1; dp[1] = 0;
for (int i = 1; i < N; i++) {
int fake = S[i]-X;
if (S[dp[l]] > fake) pref[i] += l;
else if (S[dp[1]] > fake) {
int ll = 1, rr = N;
while (ll < rr-1) {
int mid = (ll+rr)/2;
if (S[dp[mid]] <= fake) rr = mid;
else ll = mid;
}
pref[i] += ll;
}
if (S[dp[l]] > S[i]) dp[++l] = i;
else if (S[dp[1]] <= S[i]) dp[1] = i;
else {
int ll = 1, rr = N;
while (ll < rr-1) {
int mid = (ll+rr)/2;
if (S[dp[mid]] <= S[i]) rr = mid;
else ll = mid;
}
dp[rr] = i;
}
}
int m = 0;
for (int i = 0; i < N; i++) m = max(m, pref[i]);
cout << m << '\n';
}
