### _doomboo_'s blog

By _doomboo_, history, 9 months ago,

I was solving this problem with a combinatoric based approach. So I compute binomial coefficient using Fermat Little Th. and modular inverse. I have correct values for N up to 34, then my code starts giving wrong values. I belive the approach is correct but there is some issue I cant figure out related to modulo, modular inverse or powmod function. Can anybody help me understanding why my code fails? CODE

• +3

 » 9 months ago, # |   0 Auto comment: topic has been updated by _doomboo_ (previous revision, new revision, compare).
 » 9 months ago, # |   0 Your C(n, k) function should be: Spoilerreturn fact[n] * inv(fact[k] * fact[n - k] % MOD) % MOD; 
•  » » 9 months ago, # ^ |   0 it doesnt work anyway. I tried also approach with pascal triangle to compute nCk but I get the exact same issue: everything is fine for N up to 34, then it starts failing... it's so weird
 » 9 months ago, # |   0 Nobody able to help? :(
 » 9 months ago, # |   +5 I might be wrong, but the 70'th line (ll half = nCk(n, n/2, mod)/2;) might mess your calculations up, maybe that nCk returns an odd number but you divide it by 2. I suggest using *inverse(2, mod). I haven't verified the code logic so I cannot say if the approach is correct or not. Hope this helps.