Consider a regular Codeforces round consisting of three problems that uses dynamic scoring. You are given an almost final scoreboard. For each participant (including yourself), the time of the accepted submission for each of the problems is given. Also, for each solution you already know whether you are able to hack it or not. The only changes in the scoreboard that will happen before the end of the round are your challenges. What is the best place you may take at the end? More formally, n people are participating (including yourself). For any problem, if it was solved by exactly k people at the end of the round, the maximum score for this problem is defined as: If n < 2 k ≤ 2 n , then the maximum possible score is 500 ; If n < 4 k ≤ 2 n , then the maximum possible score is 1000 ; If n < 8 k ≤ 2 n , then the maximum possible score is 1500 ; If n < 16 k ≤ 2 n , then the maximum possible score is 2000 ; If n < 32 k ≤ 2 n , then the maximum possible score is 2500 ; If 32 k ≤ n , then the maximum possible score is 3000 . Let the maximum possible score for some problem be equal to s . Then a contestant who didn't manage to get it accepted (or his solution was hacked) earns 0 points for this problem. If he got the the solution accepted t minutes after the beginning of the round (and his solution wasn't hacked), he earns points for this problem. The overall score of a participant is equal to the sum of points he earns for each problem plus 100 points for each successful hack (only you make hacks). The resulting place you get is equal to one plus the number of participants who's overall score is strictly greater than yours. The first line of the input contains a single integer n ( 1 ≤ n ≤ 5000 ) — the number of participants. You are the participant number 1 . Each of the following n lines contains three integers a i , b i and c i . Here a i = 0 means that the participant number i didn't manage to accept first problem. If 1 ≤ a i ≤ 120 , then the participant number