A Block Withholding attack, also known as selfish mining, occurs when a miner on a pool intentionally does not submit any blocks they find to the pool. The intent of the attack is to cause the profitability of the mining pool to drop. This type of attack, if sustained for a long enough period, can bankrupt a pay-per-share pool. Mitigation of a Block Withholding attack is complex because of the random nature of mining, but some methods have been developed such as various cryptographic commitment schemes combined with hash functions. These schemes typically prevent the pool administrator from cheating on the entire pool and make it impossible for miners in the pool to distinguish between a partial proof of work and complete proof of work.
One form of block withholding attack is operated when a miner finds a block in a victim pool and decides to avoid submitting immediately to the pool operator, and focuses all available mining power to the victim pool to increase the number of relative shares in the pool. After waiting, the blocker then releases the formerly found block. This is defended by using oblivious tasks so the miner is unable to determine between a full solution and a share.
