That's not how it works.

Invalid blocks are a function of how far away (from an average node distance standpoint) the pool's bitcoind is from the rest of the bitcoin network. A high number of invalid block submissions implies that the bitcoind instance is relatively closer to an edge of the network than the center.

The pool is working on a block with a prevhash pointing to the previous found block. But if the pool's bitcoind doesn't have the latest information from the blockchain, then it will be working against the wrong prevhash. The further away the pool is from the center of the network, the longer it will take (on average) for the pool to get blockchain updates.