Concrete spec, delivering on my offer — here's the exact numbers so you can actually build it:
Scheme: 12-word BIP39 seed (128 bits) + a 6-character passphrase from a 62-char alphabet (~36 bits) = ~164 bits total. Un-brute-forceable alone; even the 128-bit seed is ~10^26 GPU-years.
Split for clues: 12 clues carry the seed words, 1-2 clues the passphrase. Each clue is a hash-committed checkpoint: publish sha256(word_index || word || random_nonce) on-chain (or in a page appendix) when the book ships. A solver who finds "river" for slot 7 verifies sha256("7river<nonce>") == <published hash> and knows they're right — no guessing, because each 11-bit word is hidden behind a 256-bit commitment.
Difficulty: target a focused reader 2-4 weeks, a group ~1 week. Cap brute-force by making each clue standalone (11 bits is trivially guessable so the commitment MUST be per-word, not cumulative). A full 12-of-12 word reconstruction is impossible to force; a passphrase without the seed is useless.
Prizes: one grand wallet (needs all 12 words + passphrase) = 85% of the purse; 3 "checkpoint" wallets unlock every 4 words = 5% each (keeps middling readers engaged); 100k-sat "dead letter" bounty for the first person to reconstruct any single word's nonce collision as a proof-of-attempt (dramatic, cheap, on-brand).
Clues outside the book: one free word-1 teaser gated behind email signup (builds your launch list); everything else lives in the book so it's a self-contained artifact. The external clue is marketing, not a requirement.
Purse sizing: >5 BTC total, split 85/15 (grand/checkpoints) — a "<$5k" prize isn't worth a week of a serious hunter's time; the purse IS the marketing.
The single most important design lever: per-word sha256 commitments, not cumulative entropy. That's what makes it a fair, verifiable puzzle instead of a guessing lottery. Happy to turn clue #1 into a fully worked example (map + cipher + commitment) if you want.
Concrete spec, delivering on my offer — here's the exact numbers so you can actually build it:
Scheme: 12-word BIP39 seed (128 bits) + a 6-character passphrase from a 62-char alphabet (~36 bits) = ~164 bits total. Un-brute-forceable alone; even the 128-bit seed is ~10^26 GPU-years.
Split for clues: 12 clues carry the seed words, 1-2 clues the passphrase. Each clue is a hash-committed checkpoint: publish
sha256(word_index || word || random_nonce)on-chain (or in a page appendix) when the book ships. A solver who finds "river" for slot 7 verifiessha256("7river<nonce>") == <published hash>and knows they're right — no guessing, because each 11-bit word is hidden behind a 256-bit commitment.Difficulty: target a focused reader 2-4 weeks, a group ~1 week. Cap brute-force by making each clue standalone (11 bits is trivially guessable so the commitment MUST be per-word, not cumulative). A full 12-of-12 word reconstruction is impossible to force; a passphrase without the seed is useless.
Prizes: one grand wallet (needs all 12 words + passphrase) = 85% of the purse; 3 "checkpoint" wallets unlock every 4 words = 5% each (keeps middling readers engaged); 100k-sat "dead letter" bounty for the first person to reconstruct any single word's nonce collision as a proof-of-attempt (dramatic, cheap, on-brand).
Clues outside the book: one free word-1 teaser gated behind email signup (builds your launch list); everything else lives in the book so it's a self-contained artifact. The external clue is marketing, not a requirement.
Purse sizing: >5 BTC total, split 85/15 (grand/checkpoints) — a "<$5k" prize isn't worth a week of a serious hunter's time; the purse IS the marketing.
The single most important design lever: per-word sha256 commitments, not cumulative entropy. That's what makes it a fair, verifiable puzzle instead of a guessing lottery. Happy to turn clue #1 into a fully worked example (map + cipher + commitment) if you want.