Zero-knowledge trading performance
How a trading result can be proven correct while the trades behind it stay private.
The idea
A zero-knowledge proof lets someone show that a computation was done correctly on inputs the verifier never sees. For performance, the computation is the metric calculation and the private inputs are the trading history.
How Linvesther does it
The metric calculation is an ordinary program that runs inside the RISC Zero zkVM. Running it produces a receipt that binds three things: the identifier of the exact program that ran, a public journal (for example the period and the result) and a proof that the program executed correctly. The private inputs appear only as commitments, so the receipt reveals the result and nothing about the trades.
A verifier needs the receipt and an independent source for the accepted program identifier. With those, they can check the proof offline, without trusting Linvesther's servers.
Where the data comes from
The journal also records the fingerprint of the collector key that signed the source data. A proof shows that a calculation is right, not that the data was honest, so a verifier compares that fingerprint with a list of collectors they trust. A signer that is not on the list is reported as self-attested.
Current scope
Proofs exist today for each exchange account, covering return and maximum drawdown. Proofs for the combined record and for claims are planned. Generating a proof is memory-hungry, which is why it runs on demand rather than continuously.