
Why Reputation Systems Fail Without Identity
Every reputation system — eBay feedback, Uber ratings, Amazon reviews — quietly depends on one assumption: that building a reputation is harder than throwing it away and starting over. Remove verified identity from the equation, and that assumption stops holding.
The Sybil problem
A Sybil attack is when one actor creates many fake identities to manipulate a system that assumes one identity per participant. On the ordinary web, this is expensive to pull off at scale — phone verification, credit card checks, and IP tracking all raise the cost of a fresh identity. In a pseudonymous, privacy-first system, creating a new identity can be nearly free, which means “just start a new account” is a viable response to any bad reputation, not a last resort.
What this breaks
- Vote manipulation: a seller can create buyer accounts to leave themselves glowing reviews, or leave negative reviews on competitors.
- Reputation laundering: a burned identity is simply abandoned in favor of a new one, with the negative history left behind rather than carried forward.
- First-mover fraud: build a small amount of genuine reputation cheaply, then cash it in with one large-scale scam before the identity is discarded — a pattern sometimes called “milking” reputation.
Partial fixes, and their costs
Vouching/invitation systems require an existing trusted member to vouch for a newcomer, making Sybil identities expensive to bootstrap — at the cost of making the system harder to join for legitimate newcomers too.
Stake-based reputation requires posting a deposit or bond before participating, so abandoning an identity has a real financial cost. This raises the barrier to Sybil attacks proportionally to the stake size, but also raises the barrier to entry for everyone.
Proof-of-work identity (making a new account computationally expensive to create) raises the cost of mass-producing fake identities without requiring any real-world identity information — a technique borrowed from anti-spam systems.
Escrow-linked reputation, where reviews can only be left by a wallet that actually completed a real, funds-moving transaction with the counterparty, at least ties reputation to genuine economic activity rather than free-to-fabricate social proof.
The honest conclusion
None of these fully replicate what real-world identity accountability provides — they raise the cost of gaming the system, they don’t eliminate it. Any pseudonymous marketplace or platform has to treat its reputation numbers as a noisy, gameable signal rather than ground truth, and design its dispute/escrow layer (see the previous post) to not depend entirely on reputation scores being trustworthy. The uncomfortable truth is that “no identity, but still a fully reliable trust score” is a contradiction most systems can only partially resolve, not fully solve.