Peer-to-peer file sharing networks have gained a lot of popularity and are burgeoning day by day. The widespread use inevitably produces malicious peers who try to disrupt the network by flooding it with inauthentic files. This necessitates the need for a robust reputation system to identify bad peers and files. The former has been extensively studied and many peer based reputations already exist. In this paper we present a novel framework for a Hybrid P2P Reputation System (HPRS) using peer and file rating. This combined rating enables good peers to judge individual files and thus eliminate any inauthentic downloads. It also helps a peer extract the maximum from the system by enabling him to get those occasional good files from bad peers and reject accidental bad files from good peers. Simulation results illustrate that HPRS significantly improves performance in a network.