In blockchain technology, a fraud proof is a cryptographical evidence that a verifier submits to challenge a transaction's validity. Developers widely use fraud proofs to enable the on-chain scalability of blockchains while ensuring the accuracy and availability of on-chain data.

These proofs primarily act as a safeguard against incorrect state transitions. Since they are only activated when inconsistencies are identified, they conserve computational resources, making them ideal for environments where scalability is a priority. 

For example, Fraud proofs are crucial in Optimistic rollups to identify and handle invalid transactions. Whenever a transaction is approved, there's a specific dispute period where anyone can challenge the current state by presenting a valid fraud proof. If someone can prove a transaction is fraudulent during this window, it is nullified and the network readjusts to the previous state.

This system is in place to incentivize good behavior. If the network properly executes the rollups per the consensus rules, the concerned parties receive a financial reward. However, they face monetary penalties and risk forfeiting their fraud proof if they approve an erroneous transaction. This dual approach of rewards and penalties ensures that the rollups remain cost-effective and fast, optimizing the performance of decentralized applications on a blockchain.

Fraud proofs are not without their cons. They require constant communication between multiple parties. This back-and-forth can lead to system disruptions and open the gates for dishonest behavior or other illicit actions.

Another issue is their reliance on the assumption that all block data is available. If a miner only provides the block header without its accurate contents, it's impossible to determine its correctness. While fraud proofs offer solutions, they also present challenges that need addressing.

Also Learn: Optimistic vs. Zero-Knowledge Rollups: What’s the Difference?