The introduction of new bitcoins into the network is one of the main objectives of mining, but this reward is simply an incentive to do the other more important part of mining, adding transactions to a block. When a block is resolved, no bit can be changed without invalidating it. This means that all transactions in the block have been processed and are final, in that they are in the blockchain and will not change. For example, if I have a bitcoin, I can send it to Alice or Bob, but not both. So, given the two possible transactions, there must be a way to decide which one is valid, the two answers being equally good. So, it’s not enough to check them, but to pick winners and losers in a way that allows others to agree. This is the Nakamoto consensus at work, requiring billions of mathematical steps to solve a block. This also means that it is very difficult to accomplish. The honest miners who mines a new block are rewarded in this form of bitcoins. But attackers who try to go back in history and replace old blocks with new blocks are at a disadvantage, because that requires a huge amount of hashpower. While the attacker returns and tries to re-mine old blocks, the rest of the network continues its extraction operations on the tip of the blockchain, which makes it even more difficult to catch up. Therefore, to create blocks faster than the rest of the network, the attacker must control more than 50% of the network’s hash power.