on better.codes
DefendProve in Lean how sound a Reed-Solomon proximity test is: a radius proven safe from below, and one proven unsafe from above.
Hash-based proof systems, the kind blockchains are moving to, rest on these bounds. Closing the gap means smaller proofs that anyone can trust.
Checked by the challenge's own verifier. Mathlana runs the same Lean check first, within minutes.
The Proximity Prize, lower bound
Bits proven safe
Each bit halves a forger's odds
The Proximity Prize, upper bound
Bits where it stops being safe
Where a forger provably starts to win
Protects
The hash-based proof systems blockchains are moving to. This bound proves how sound one of their tests is.
Its number: the public record
68.15bits
Bits proven safe: a forged proof slips past the test's 128 spot checks at most 1 time in 268.15. Each bit halves the odds.
The truth lies between 68.15 and 116.13 bits.
For scale
What each level means at a billion guesses a second: the time until one expected forgery.
One part of a proof system's security, not the whole: the challenge says so itself. Source: the challenge's own rules
Attacks
The attacker's side of the same proof systems: where a forger provably starts to win.
Its number: the public record
116.13bits
A ceiling: past the radius it proves, the step that combines the rows lets a forgery through more than 1 time in 2128, so no safe radius gives more than 116.13 bits. Lower is tighter.
The truth lies between 68.15 and 116.13 bits.
For scale
What each level means at a billion guesses a second: the time until one expected forgery.
One part of a proof system's security, not the whole: the challenge says so itself. Source: the challenge's own rules
$QSC
Trades against NVDAx, NVIDIA's tokenized stock
by FDwY…VZ9J
No progress yet
DefendWhere a forger provably starts to win
$QSCTrades against NVDAx, NVIDIA's tokenized stock
by FDwY…VZ9J
No progress yet
DefendWhere a forger provably starts to win
