Need some help understanding how Algorand's secure
I'm no blockchain expert by any means, just some hobbyist. I've come across Algorand just a few days ago and was immediately intrigued. Traditional PoW based blockchains always seemed so pointless with its inability to scale, and there's this magical thing that can have all of scalability, security, and decentralization. Wow!
So my understanding is that to add a new block to the chain, a single proposer is selected, who then proposes a new block, then 1000 others join in to vote whether the block seems good or not. Based on the probability mathematics and disincentive to act maliciously when you hold a majority stake, I don't have a problem believing that the majority of the participants would behave in good faith. My problem is with the information these participants are basing their decisions upon.
As a participant node you're buffering up transactions that should go on subsequent blocks, and trying to achieve consensus on those transactions, right? How do you know that the stream of pending transactions you receive are legitimate? Are we just betting that it would be very difficult to flood a majority of participant nodes with fraudulent transactions? Actually, aren't participant nodes only receiving data from relay nodes, of which much fewer exist, and the bad actor only needs to target the relay nodes?
But wait what are fraudulent transactions anyway? You can't steal money from someone's wallet unless you have their private key to sign the transaction with. Perhaps the threat we should be worried about is the problem of double spending. But is that even possible if forking is impossible? Ok, forget about trying to score some ALGOs at all. Place some bearish bets on ALGO through external means (e.g., short it on a crypto exchange) launch a network attack flooding relay nodes with all kinds of conflicting transactions, preventing consensus and grinding the block progression down to a halt. ALGO becomes unusable, confidence shattered, value drops,...