/*
TODO:

*/
//#define NDEBUG

//#define ONLINE_JUDGE

#ifndef ONLINE_JUDGE
//#define OPTUNA
#endif

#ifdef ONLINE_JUDGE
#define NDEBUG
#pragma GCC target("avx2")
#pragma GCC optimize("O3")
#pragma GCC optimize("unroll-loops")
#endif

#include<bits/stdc++.h>
using namespace std;
using ll=long long int;
//using Int=__int128;
#define mask(x) ((1LL<<x)-1)
#define ALL(A) A.begin(),A.end()
template<typename T1,typename T2> bool chmin(T1 &a,T2 b){if(a<=b)return 0; a=b; return 1;}
template<typename T1,typename T2> bool chmax(T1 &a,T2 b){if(a>=b)return 0; a=b; return 1;}
template<typename T> int bitUP(T x,int a){return (x>>a)&1;}
enum Dir{
    Right,
    Down,
    Left,
    Up
};
//→　↓ ← ↑
int dh[4]={0,1,0,-1}, dw[4]={1,0,-1,0};
//上から時計回り
//int dx[8]={0,1,1,1,0,-1,-1,-1}, dy[8]={1,1,0,-1,-1,-1,0,1};
long double EPS = 1e-6;
const ll INF=(1LL<<62);
const int MAX=(1<<30);
using pii=pair<int,int>;
using pil=pair<int,ll>;
using pli=pair<ll,int>;
using pll=pair<ll,ll>;
using psi=pair<string,int>;
using pis=pair<int,string>;
using psl=pair<string,ll>;
using pls=pair<ll,string>;
using pss=pair<string,string>;
template<class T>
using minimum_queue=priority_queue<T,vector<T>,greater<T>>;

using Graph=vector<vector<int>>;

using i8=int8_t;
using i16=int16_t;
using i32=int32_t;
using i64=int64_t;

using u8=uint8_t;
using u16=uint16_t;
using u32=uint32_t;
using u64=uint64_t; 

void FastIO(){
    ios::sync_with_stdio(false);
    cin.tie(0);
    cout << fixed << setprecision(20);
}
//0-indexed vector cin
template<typename T>
inline istream &operator>>(istream &is,vector<T> &v) {
    for(size_t i=0;i<v.size();i++) is>>v[i];
	return is;
}
 
//0-indexed vector cin
template<typename T>
inline istream &operator>>(istream &is,vector<vector<T>> &v) {
    for(size_t i=0;i<v.size();i++){
        is>>v[i];
    }
    return is;
}


struct Xor32{
    using u32=uint32_t;
 
    u32 x=1234567;
    inline u32 rnd_make(){
        x^=(x<<13);
        x^=(x>>17);
        x^=(x<<5);
        return x;
    }
    inline u32 operator()(){
        return rnd_make();
    }
    //[a,b)
    inline int operator()(int a,int b){
        int dis=b-a;
        int add=rnd_make()%dis;
        return a+add;
    }
    //[0,b)
    inline int operator()(int b){
        return rnd_make()%b;
    }
    //http://web.archive.org/web/20200105011004/https://topcoder.g.hatena.ne.jp/tomerun/20171216/
    //[0,b)の中から異なる2つを選ぶ [0]の値<[1]の値
    inline array<int,2> two(int b){
        assert(b>=2);
        int v1=rnd_make()%b;
        int v2=rnd_make()%(b-1);
        if (v2>=v1) return {v1,v2+1};
        else return {v2,v1};
    }
    inline float random01(){
        return float(rnd_make())/mask(32);
    }
    inline double random_double(double a,double b){
        double sa=b-a;
        a+=random01()*sa;
        return a;
    }
    //確率pでtrueを返す
    inline bool gen_bool(float p){
        return p>random01();
    }
};

struct Xor64{
    u64 x=1234567;
    inline u64 rnd_make(){
    	x ^= x << 13;
    	x ^= x >> 7;
    	x ^= x << 17;
    	return x;
    }
    inline u64 operator()(){
        return rnd_make();
    }
};

struct Timer{
    chrono::high_resolution_clock::time_point st;
    float local;
    Timer(){
#ifndef ONLINE_JUDGE
        local=1.0;
#endif
        start();
    }
    void start(){
        st=chrono::high_resolution_clock::now();
    }
    int span()const{
        auto now=chrono::high_resolution_clock::now();
        return chrono::duration_cast<chrono::milliseconds>(now-st).count();
    }
};
struct TestTimer{
    chrono::high_resolution_clock::time_point st;
    unordered_map<string,ll> sum_time;
    unordered_map<string,chrono::high_resolution_clock::time_point> start_time;

    TestTimer(){}
    void start(const string &s){
#ifndef ONLINE_JUDGE
        start_time[s]=chrono::high_resolution_clock::now();
#endif
    }
    void end(const string &s){
#ifndef ONLINE_JUDGE
        auto now=chrono::high_resolution_clock::now();
        sum_time[s]+=chrono::duration_cast<chrono::nanoseconds>(now-start_time[s]).count();
#endif
    }
    void output()const{
#ifndef ONLINE_JUDGE
        for(auto m:sum_time){
            cerr<<m.first<<": "<<m.second/1e6<<"ms"<<endl;
        }
#endif
    }
};
struct TestCounter{
    unordered_map<string,ll> cnt;

    TestCounter(){}

    void count(const string &s){
#ifndef ONLINE_JUDGE
        cnt[s]++;
#endif
    }
    void output()const{
#ifndef ONLINE_JUDGE
        for(auto m:cnt){
            cerr<<m.first<<": "<<m.second<<endl;
        }
#endif
    }    
};
Timer TIME;
Xor32 Rand32;
Xor64 Rand64;
TestTimer testTimer;
TestCounter testCounter;

//https://atcoder.jp/contests/asprocon9/submissions/34659956
template<class T,int CAPACITY>
class DynamicArray{
public:
    array<T,CAPACITY> array_={};
    int size_=0;

    DynamicArray(){}

    DynamicArray(int n){
        resize(n);
    }

    void push_back(const T &e){
        array_[size_++]=e;
    }
    //ソートされた状態を保つように挿入
    void insert(const T &e){
        int ng=-1,ok=size_;
        while(ok-ng!=1){
            int mid=(ok+ng)/2;
            if(array_[mid]>e) ok=mid;
            else ng=mid;
        }
        for(int i=size_;i>ok;i--){
            array_[i]=array_[i-1];
        }
        array_[ok]=e;
        size_++;
    }
    //eをこえる一番左の添字
    int find_binary_search(const T &e)const{
        int ng=-1,ok=size_;
        while(ok-ng!=1){
            int mid=(ok+ng)/2;
            if(array_[mid]>=e) ok=mid;
            else ng=mid;
        }
        return ok;
    }
    void pop_back(){
        size_--;
    }

    inline T& operator[](int index){
        return array_[index];
    }
	inline const T& operator[](int index) const {
		return array_[index];
	}
    inline int size()const{
        return size_;
    }
    inline T& back(){
        return array_[size_-1];
    }
    const inline T& back()const{
        return array_[size_-1];
    }
	inline auto begin() -> decltype(array_.begin()) {
		return array_.begin();
	}

	inline auto end() -> decltype(array_.begin()) {

		return array_.begin() + size_;
	}

	inline auto begin() const -> decltype(array_.begin()) {
		return array_.begin();
	}

	inline auto end() const -> decltype(array_.begin()) {

		return array_.begin() + size_;
	}
    inline void resize(int new_size){
        size_=new_size;
    }
    void operator=(const DynamicArray &e){
        for(int i=0;i<e.size_;i++){
            array_[i]=e[i];
        }
        size_=e.size_;
    }
    bool operator==(const DynamicArray &v){
        if(size_!=v.size_) return false;
        for(int i=0;i<size_;i++){
            if(array_[i]!=v[i]){
                return false;
            }
        }
        return true;
    }
    bool operator==(const vector<T> &v){
        if(size_!=v.size()) return false;
        for(int i=0;i<size_;i++){
            if(array_[i]!=v[i]){
                return false;
            }
        }
        return true;
    }
    void clear(){
        size_=0;
    }
    //O(1)
    //末尾と入れ替える。順序が保持されないが高速
    void swap_remove(int idx){
        array_[idx]=array_[size_-1];
        size_--;
    }
    //最初から見ていき、一致したものを削除(remove)
    bool erase(T value){
        for(int i=0;i<size_;i++){
            if(array_[i]==value){
                remove(i);
                return true;
            }
        }
        return false;
    }
    void reverse(){
        for(int i=0;i<size_/2;i++){
            swap(array_[i],array_[size_-1-i]);
        }
    }
    //O(size)
    //順序を気にしない場合、swap_removeの方がいい
    void remove(int idx){
        for(int i=idx;i<size_-1;i++){
            array_[i]=array_[i+1];
        }
        size_--;
    }
    void fill(T x){
        for(int i=0;i<size_;i++){
            array_[i]=x;
        }
    }
};

//ソート
template<typename T>
inline void vsort(vector<T> &v){
    sort(v.begin(),v.end());
}
//逆順ソート
template<typename T>
inline void rvsort(vector<T> &v){
	sort(v.rbegin(),v.rend());
}
//1ビットの数を返す
inline int popcount(int x){
	return __builtin_popcount(x);
}
//1ビットの数を返す
inline int popcount(ll x){
	return __builtin_popcountll(x);
}
template<typename T>
inline void Compress(vector<T> &C){
    sort(C.begin(),C.end());
    C.erase(unique(C.begin(),C.end()),C.end());
}
//要素数n 初期値x
template<typename T>
inline vector<T> vmake(size_t n,T x){
	return vector<T>(n,x);
}

//a,b,c,x data[a][b][c] 初期値x
template<typename... Args>
auto vmake(size_t n,Args... args){
	auto v=vmake(args...);
	return vector<decltype(v)>(n,move(v));
}

//vは昇順
bool is_in(const vector<int> &v,int x){
    int n=v.size();
    if(n==0) return false;
    int ng=-1,ok=n-1;
    while(ok-ng!=1){
        int mid=(ok+ng)/2;
        if(v[mid]<x) ng=mid;
        else ok=mid;
    }
    if(v[ok]==x) return true;
    return false;
}

template<typename T >
struct edge {
	int to;
	T cost;
    int id;
	edge()=default;
	edge(int to, T cost,int id) : to(to), cost(cost), id(id) {}
};
template<class T >
struct Edge {
    int from, to,id;
    T cost;
    Edge(int from,int to,T cost,int id):from(from),to(to),cost(cost),id(id){}
    Edge()=default;

    bool operator<(const Edge<T> &e){
        return cost<e.cost;
    }
    bool operator<=(const Edge<T> &e){
        return cost<=e.cost;
    }
};
template<typename T>
using WeightGraph=vector<vector<edge<T>>>;
//vector cout
template<typename T>
inline ostream &operator<<(ostream &os,const vector<T> &v) {
    bool sp=true;
    if(string(typeid(T).name())=="c") sp=false;
    for(size_t i=0;i<v.size();i++){
        if(i and sp) os<<" ";
        os<<v[i];
    }
    return os;
}
//vector<vector> cout
template<typename T>
inline ostream &operator<<(ostream &os,const vector<vector<T>> &v) {
    for(size_t i=0;i<v.size();i++){
        os<<v[i];
        if(i+1!=v.size()) os<<"\n";
    }
    return os;
}
//pair cout
template<typename T, typename U>
inline ostream &operator<<(ostream &os,const pair<T,U> &p) {
	os<<p.first<<" "<<p.second;
	return os;
}
//map cout
template<typename F, typename S>
inline ostream &operator<<(ostream &os,const map<F,S> &M) {
	bool first=false;
    for(auto x:M){
        if(first) os<<endl;
        first=true;
        os<<x;
    }
	return os;
}
//set cout
template<typename T>
inline ostream &operator<<(ostream &os,const set<T> &S) {
	bool first=false;
    for(auto x:S){
        if(first) os<<endl;
        first=true;
        os<<x;
    }
	return os;
}
 
//pair cin
template<typename T, typename U>
inline istream &operator>>(istream &is,pair<T,U> &p) {
	is>>p.first>>p.second;
	return is;
}
template<typename T>
void append(vector<T> &v,const vector<T> &vp){
    for(auto p:vp){
        v.push_back(p);
    }
}
//Fisher–Yatesアルゴリズム
template<typename T>
void shuffle(vector<T> &v){
    int sz=v.size();
    for(int i=sz-1;i>0;i--){
        swap(v[Rand32()%(i+1)],v[i]);
    }
}

template<class T,class U>
T linear_function(U x,U start_x,U end_x,T start_value,T end_value){
    if(x>=end_x) return end_value;
    if(x<=start_x) return start_value;
    return start_value+(end_value-start_value)*(x-start_x)/(end_x-start_x);
}

//http://gasin.hatenadiary.jp/entry/2019/09/03/162613
struct SimulatedAnnealing{
    long double temp_start; //差の最大値(あくまでも参考)
    long double temp_end; //差の最小値(あくまでも参考)
    float time_start;
    float time_end;
    bool is_hill;
    bool minimum;
    int interval; //intervalごとに温度の再計算

    long double temp_now;
    int cnt_calc_temp;
    /*
    0:線形
    1:pow pow
    2:指数
    */
    int temp_type;
    

    //SimulatedAnnealing(){}
    SimulatedAnnealing(long double temp_start,long double temp_end,float time_start,float time_end,bool is_hill,bool minimum,int temp_type=0,int interval=1):
        temp_start(temp_start),temp_end(temp_end),time_start(time_start),time_end(time_end),
        is_hill(is_hill),minimum(minimum),temp_type(temp_type),interval(interval),temp_now(temp_start),cnt_calc_temp(0){
    }
    float calc_temp(){
        if(cnt_calc_temp%interval==0){
            float progress=float(TIME.span()-time_start)/(time_end-time_start);
            if(progress>1.0) progress=1.0;
            if(temp_type==0){//線形
                temp_now=temp_start*(1.0-progress)+temp_end*progress;
            }else if(temp_type==1){ //https://atcoder.jp/contests/ahc014/submissions/35326979
                temp_now = pow(temp_start,1.0-progress)*pow(temp_end,progress);
            }else{ //https://ozy4dm.hateblo.jp/entry/2022/12/22/162046#68-%E3%83%97%E3%83%AB%E3%83%BC%E3%83%8B%E3%83%B3%E3%82%B0%E6%97%A9%E6%9C%9F%E7%B5%82%E4%BA%86%E5%8D%98%E7%B4%94%E5%8C%96%E3%81%95%E3%82%8C%E3%81%9F%E8%A8%88%E7%AE%97%E3%82%92%E4%BD%BF%E7%94%A8%E3%81%99%E3%82%8B
                temp_now = temp_start*pow(temp_end/temp_start,progress);
            }
        }
        cnt_calc_temp++;
        return temp_now;
    }
    //diff: スコアの変化量
    //確率を計算
    float calc_prob(float diff){
        if(minimum) diff*=-1;
        if(diff>0) return 1;
        long double temp=calc_temp();
        return exp(diff/temp);
    }
    inline bool operator()(long double diff){
        testCounter.count("try_cnt");
        if(minimum) diff*=-1;
        if(diff>=0){
            if(diff==0) testCounter.count("zero_change");
            else testCounter.count("plus_change");
            return true;
        }
        if(is_hill) return false;

        long double prob = exp(diff/calc_temp());

        if(Rand32.gen_bool(prob)){
            testCounter.count("minus_change");
            return true;
        }
        else return false;
    }
    //最大化の場合,返り値<変化量なら遷移してもよい
    long double calc_tolerance(long double prob){
        long double tolerance=log(prob)*calc_temp();
        if(minimum) tolerance*=-1;
        return tolerance;
    }
    //log(prob)*temp prob:[0,1]の乱数
    long double calc_tolerance(){
        long double prob=Rand32.random01();
        return calc_tolerance(prob);
    }
};

///++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
///++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
///++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
//https://atcoder.jp/contests/asprocon9/submissions/34659956
#ifndef OPTUNA 
#define REGIST_PARAM(name, type, defaultValue) constexpr type name = defaultValue
#else
#define REGIST_PARAM(name, type, defaultValue) type name = defaultValue
#endif

namespace OP{
    REGIST_PARAM(yama,bool,false);
    REGIST_PARAM(startTemp,double,500000);
    REGIST_PARAM(endTemp,float,0);
    REGIST_PARAM(TIME_END,int,1900);
};

const int num_cards=45;
struct Card{
    ll A;
    ll B;
};

vector<Card> inputs;
ll middle=500000000000000000;

void input(){
    int dummy;
    cin>>dummy;
    inputs.resize(num_cards);
    for(int i=0;i<num_cards;i++){
        cin>>inputs[i].A>>inputs[i].B;
    }
}

struct Solver{
    vector<int> perm;
    vector<int> ratio;
    ll a=0,b=0;

    ll calc_score(const vector<int> &perm,bool output=false){
        a=0,b=0;
        for(int i=0;i<45;i++){
            int j=i+1;
            if(i==44) j=44;
            a+=(inputs[perm[i]].A>>ratio[i]);
            b+=(inputs[perm[i]].B>>ratio[i]);
        }
        if(output) cout<<a<<" "<<b<<endl;
        //cout<<abs(middle-a)<<" "<<abs(middle-b)<<endl;
        return max(abs(middle-a),abs(middle-b));
    }
    ll calc_score_diff(const vector<int> &perm,array<int,2> &uv,bool update=false){
        ll new_a=a,new_b=b;
        for(int i=0;i<2;i++){
            new_a-=(inputs[perm[uv[i]]].A>>ratio[uv[i]]);
            new_b-=(inputs[perm[uv[i]]].B>>ratio[uv[i]]);

            int j=1-i;
            new_a+=(inputs[perm[uv[j]]].A>>ratio[uv[i]]);
            new_b+=(inputs[perm[uv[j]]].B>>ratio[uv[i]]);
        }
        if(update){
            a=new_a;
            b=new_b;
        }
        return max(abs(middle-new_a),abs(middle-new_b));
    }

    void solve(){
        ratio.push_back(11);
        for(int i=14;i>=4;i--){
            ratio.push_back(i);
            ratio.push_back(i);
            ratio.push_back(i-1);
            ratio.push_back(i-2);
        }
        for(int i=0;i<num_cards;i++){
            perm.push_back(i);
        }
        SimulatedAnnealing SA(1e14,1e5,TIME.span(),900,false,true,0,10000);
        ll best_score=calc_score(perm);
        auto best_perm=perm;
        int loop=0;
        while(true){
            loop++;
            if(loop%100000==0){
                if(TIME.span()>900) break;
            }
            auto uv=Rand32.two(num_cards);
            if(uv[0]==0 or uv[1]==0) continue;
            ll now_score=calc_score_diff(perm,uv);
            
            //swap(perm[uv[0]],perm[uv[1]]);
            //cerr<<now_score<<endl;
            //cerr<<calc_score(perm)<<endl;
            //assert(calc_score(perm)==now_score);
            //swap(perm[uv[0]],perm[uv[1]]);

            if(SA(now_score-best_score)){
                //cout<<best_score<<endl;
                calc_score_diff(perm,uv,true);
                swap(perm[uv[0]],perm[uv[1]]);
                //cerr<<d<<endl;
                //cerr<<calc_score(perm)<<endl;
                //assert(calc_score(perm)==d);
                if(best_score>now_score){
                    best_score=now_score;
                    best_perm=perm;
                }
            }
        }
        perm=best_perm;
    }

    void output(){
        cout<<44<<endl;
        for(int i=0;i<11;i++){
            for(int j=0;j<4;j++){
                int id=1+i*4+j;
                if(j==3) cout<<perm[id]+1<<" "<<1<<endl;
                else cout<<perm[id]+1<<" "<<perm[id+1]+1<<endl;
            }
        }
        // for(int i=44;i>=1;i--){
        //     cout<<perm[i]+1<<" "<<perm[i-1]+1<<endl;
        //     assert(perm[i]!=perm[i-1]);
        // }
        ll score=calc_score(perm,true);
        cerr<<ll(2e6-1e5*log10l(score+1))<<endl;
    }
};


void solve(){
    input();
    Solver solver;
    solver.solve();
    solver.output();

    
#ifndef ONLINE_JUDGE
    testTimer.output();
    testCounter.output();
    cerr<<TIME.span()<<"ms"<<endl;
    //cerr<<"score: "<<simulate(best_grid,true)<<endl;
#endif
}
 
int main(const int argc,const char** argv){
#ifndef OPTUNA
    if(argc!=1){

    }
#endif
    FastIO();
    int T=1;
    //cin>>T;
    while(T--) solve();
}