3 ms·

With 51% hashpower, you can make illegal (by consensus) transactions on the longest chain. This will fork off the validating nodes, but most light clients won’t

With 51% hashpower, you can make illegal (by consensus) transactions on the longest chain. This will fork off the validating nodes, but most light clients won’t see an issue and will follow the malicious chain.

nullc 5y ago
In practice it doesn't actually work that way. Lite clients use a semi-trusted validator, and the semi-trusted validator won't be tricked. I say practice rather than theory because it's not just theory: There are many different forks of Bitcoin, including one created by Wright (BSV, "bitcoin satoshi vision" lol) which exist and are used and have lite clients. For many users, the brief exposure to the wrong chain if their validator starts giving them the wrong one isn't an actual problem-- they will notice and switch eventually and only get robbed if someone pays them on the malicious chain before they notice. Though this is the reason why the Bitcoin whitepaper recommends that users that accept frequent payments should prefer to run their own node.
- jimmydorry 5y ago
  Thanks for your clarification nullc. I don’t quite follow: * How will the light clients notice if they are on the malicious chain if it has the most work (for simplicity they see 1 malicious node and 1 non-malicious)? The light wallet surely goes with the malicious validator with the longest chain, when it sees it. It’s obvious (to me) that the non-malicious validators won’t accept invalid blocks, but not so obvious to me why light wallets would follow the smaller chain. This is at the crux of what appears to be my misunderstanding of light clients. If you have a link, I would like to read more and it would probably save you time from repeating something you’ve likely said before. As a context reminder: GP claimed that it wasn’t possible to spend someone else’s money, even with 100% of the mining rate.

nullc 5y ago

In practice it doesn't actually work that way. Lite clients use a semi-trusted validator, and the semi-trusted validator won't be tricked. I say practice rather than theory because it's not just theory: There are many different forks of Bitcoin, including one created by Wright (BSV, "bitcoin satoshi vision" lol) which exist and are used and have lite clients. For many users, the brief exposure to the wrong chain if their validator starts giving them the wrong one isn't an actual problem-- they will notice and switch eventually and only get robbed if someone pays them on the malicious chain before they notice. Though this is the reason why the Bitcoin whitepaper recommends that users that accept frequent payments should prefer to run their own node.

jimmydorry 5y ago
Thanks for your clarification nullc. I don’t quite follow: * How will the light clients notice if they are on the malicious chain if it has the most work (for simplicity they see 1 malicious node and 1 non-malicious)? The light wallet surely goes with the malicious validator with the longest chain, when it sees it. It’s obvious (to me) that the non-malicious validators won’t accept invalid blocks, but not so obvious to me why light wallets would follow the smaller chain. This is at the crux of what appears to be my misunderstanding of light clients. If you have a link, I would like to read more and it would probably save you time from repeating something you’ve likely said before. As a context reminder: GP claimed that it wasn’t possible to spend someone else’s money, even with 100% of the mining rate.