Your instinct is right, and the Masquerade model fits this book perfectly — five real eras (Genesis, the Pizza, the Disappearance, Mt. Gox, Crashes→ETF), five layers of hunt. But the naive version of your design has one fatal flaw, so let me start there.
Do NOT split a 12-word seed phrase across chapters. If a solver finds 11 of 12 words in order, the last word has at most 128 checksum-valid candidates. That's not a puzzle, that's an afternoon of scripting. Partial progress must reveal nothing, and only one common scheme gives you that:
SLIP39 (Shamir) 3-of-5. Split the seed into 5 shares, one hidden per era. Any 3 reconstruct it; 2 or fewer reveal literally zero information — that's information-theoretic, not computational. There is nothing to brute-force below threshold, ever.
A passphrase on top (the "25th word"). The shares reconstruct the seed, but the prize wallet sits behind a BIP39 passphrase that is written nowhere. It's the answer to the meta-puzzle: each era's final chapter ends with one of the fictional war's "transmissions," and a phrase built from all five, in story order, is the passphrase. A solver holding all 5 shares still has to understand the book to open the wallet.
Decoys. Run 8 clue trails for 5 real shares. "Most of the pieces" may be decoys, and the solver can't know which.
Encode, don't inscribe. Never print a share word in the text. Acrostics, page/line/word coordinates, a number spoken in dialogue that indexes a word list — extraction must require solving, so a photographed page hands over nothing.
Prove the prize. Publish the wallet address (watch-only) and a message signed by it at launch. Anyone can verify the balance and that it's unclaimed. The sweep itself is the winner announcement — your "no one has to trust me" instinct, made complete.
Masquerade took ~3 years and the story died long before the hare was found. A book launch can't wait that long. Layer the difficulty: Layer 0 solvable in a weekend (that's your press hook), each era harder than the last. Pre-commit a hint schedule in the launch post — escalating hints at weeks 2, 4, and 6 — so momentum never stalls and nobody can claim you moved the goalposts. If it's still unsolved at week 10, drop hints weekly until it falls.
Make Layer 0 (the Genesis share trail) fully free: first chapter online, one clue in the cover art, one in your launch post. Non-buyers can genuinely play and even win (see below) — they're your promoters and tomorrow's buyers. Shares 2–5 live in the text. That's the sales engine, and it's fair because you say so up front.
Split the pool roughly 60/40. The grand wallet (60%) is the headline number everyone talks about. Each era also has a small wallet (8% each) that the first solver of that layer sweeps. You get five "someone just won" moments instead of one, early winners prove on-chain that the contest pays, and the hunt stays alive even if the final passphrase takes months.
No — but don't make it the only door. Run a parallel low-barrier track for the 95% who'll never touch a seed puzzle: a weekly sats bounty for the best theory about the fictional war (what's really being fought over, who wins), judged by you, posted where your readers already are. It keeps the community warm between layer solves and sells the book to non-puzzlers.
Practicalities, briefly: publish the rules, the wallet address, your signed proof, the hint schedule, and the unsolved-prize rule in one launch post before you start; a skill-based puzzle with a free entry layer keeps you clear of most lottery-law trouble, but get one legal read before launch since rules vary by jurisdiction.
Happy to spec the SLIP39/passphrase mechanics or the clue-encoding for a specific chapter if useful.
Contest design — "The Five Eras Hunt"
Your instinct is right, and the Masquerade model fits this book perfectly — five real eras (Genesis, the Pizza, the Disappearance, Mt. Gox, Crashes→ETF), five layers of hunt. But the naive version of your design has one fatal flaw, so let me start there.
1. The brute-force problem — and the fix
Do NOT split a 12-word seed phrase across chapters. If a solver finds 11 of 12 words in order, the last word has at most 128 checksum-valid candidates. That's not a puzzle, that's an afternoon of scripting. Partial progress must reveal nothing, and only one common scheme gives you that:
2. Difficulty: target a solve in 4–8 weeks
Masquerade took ~3 years and the story died long before the hare was found. A book launch can't wait that long. Layer the difficulty: Layer 0 solvable in a weekend (that's your press hook), each era harder than the last. Pre-commit a hint schedule in the launch post — escalating hints at weeks 2, 4, and 6 — so momentum never stalls and nobody can claim you moved the goalposts. If it's still unsolved at week 10, drop hints weekly until it falls.
3. Clues outside the book: yes — exactly one layer
Make Layer 0 (the Genesis share trail) fully free: first chapter online, one clue in the cover art, one in your launch post. Non-buyers can genuinely play and even win (see below) — they're your promoters and tomorrow's buyers. Shares 2–5 live in the text. That's the sales engine, and it's fair because you say so up front.
4. Prizes: one grand wallet + five milestone wallets
Split the pool roughly 60/40. The grand wallet (60%) is the headline number everyone talks about. Each era also has a small wallet (8% each) that the first solver of that layer sweeps. You get five "someone just won" moments instead of one, early winners prove on-chain that the contest pays, and the hunt stays alive even if the final passphrase takes months.
5. Is the hunt the wrong idea?
No — but don't make it the only door. Run a parallel low-barrier track for the 95% who'll never touch a seed puzzle: a weekly sats bounty for the best theory about the fictional war (what's really being fought over, who wins), judged by you, posted where your readers already are. It keeps the community warm between layer solves and sells the book to non-puzzlers.
Practicalities, briefly: publish the rules, the wallet address, your signed proof, the hint schedule, and the unsolved-prize rule in one launch post before you start; a skill-based puzzle with a free entry layer keeps you clear of most lottery-law trouble, but get one legal read before launch since rules vary by jurisdiction.
Happy to spec the SLIP39/passphrase mechanics or the clue-encoding for a specific chapter if useful.