Double-Spending

On the main Bitcoin network, there has never been a successful "confirmed" double-spend (meaning a transaction with 6+ confirmations being reversed). The cost of the hardware and electricity required to overpower the massive global hash rate makes such an attack prohibitively expensive, likely costing billions of dollars for a very uncertain outcome. However, "unconfirmed" (0-conf) double-spends happen frequently, where a user sends a transaction and then quickly "cancels" it by sending the same funds with a higher fee to another address. This is why you should never consider a transaction final until it is mined into a block.

A credit card chargeback is a "centralized" reversal. The bank uses its authority to take money back from a merchant and return it to the customer. This is essentially a "feature" of the centralized banking system to protect against fraud, but it is often abused. A double-spend on a blockchain is a "technical" reversal that exploits the protocol itself to erase a transaction from the history. While they look similar to the merchant, the difference is that a blockchain transaction is designed to be permanent and irreversible once confirmed, whereas a credit card payment is always "probabilistic" for several months.

The Lightning Network handles double-spending through a series of "Smart Contracts" and "Time-Locked" penalties. Instead of broadcasting every transaction to the main blockchain, users open "Payment Channels." If one party tries to cheat by broadcasting an old, higher-balance state of the channel (a form of double-spending), the other party has a window of time to provide a "Revocation Key." This key allows the honest party to "punish" the cheater by taking 100% of the funds in the channel. This economic deterrent allows for instant payments without the need to wait for block confirmations.

Not necessarily, but the security model is different. In Proof-of-Work, the security is "Physical" (the cost of electricity and mining hardware). In Proof-of-Stake, the security is "Economic" (the value of the tokens staked). If a validator in a PoS system tries to sign two conflicting blocks (a double-spend), they are detected by the protocol and "slashed," meaning a portion or all of their staked wealth is destroyed. While PoW is considered more "battle-tested," a properly designed PoS system can be just as expensive to attack, provided the market cap of the staked asset is high enough.

The number 6 was suggested by Satoshi Nakamoto in the original whitepaper as a point where the probability of an attacker being able to catch up and rewrite the chain becomes mathematically negligible, even if they control 10% of the network hash rate. At 6 confirmations, the chance of a successful reorganization is nearly zero. For very large transactions (millions of dollars), some institutions wait for 10 or 20 confirmations. For smaller transactions, 1 or 2 is often sufficient. It is a sliding scale of "Risk vs. Time" that every user must decide for themselves.