There’s a thing called “Bitcoin Puzzles” where someone (anonymous) has placed relatively large amounts of bitcoin under private keys and publicly advertised that those keys are mostly-all-bits-zero. So the game is to discover the mostly-zero private key by brute force, and then claim the bitcoin.
But of course claiming the bitcoin requires submitting a transaction to the pool, and that transaction is going to include the corresponding public key; and (for math reasons I have not investigated personally) if you know the public key then the difficulty of brute-forcing the private key goes from 2^{k} down to something like 2^{k/2} — i.e., instantaneous, for the number of nonzero bits we’re talking about in these puzzles.
So, when Bitcoin Puzzle #66 was recently solved, it was quickly rumored that what had really happened was:
Theory #1, the clever bot
- Some poor guy spends years brute-forcing a 66-bit private key.
- He submits a transaction, including the public key, asking for 6.6BTC to be transferred from the prize address to himself.
- Some random bot has been watching the pool for years for exactly this moment. The bot takes a few tens of seconds to brute-force the now-effectively-33-bit private key, and submits its own transaction consuming the same input but with a higher transaction fee
- The bot’s transaction thus replaces the poor guy’s transaction, and is mined, giving the prize money to the bot. Poor guy spent years of brute-force time for nothing.
But of course there are at least two competing theories:
Theory #2, the boring one
- Some guy spends years brute-forcing a 66-bit private key. He submits the transaction; nobody front-runs him; he just made $400,000. The end.
Theory #3, the puzzlemaster cheated
- The anonymous puzzlemaster, who knows all the private keys by definition, could have submitted his own transaction to take out the money at any time (say, to buy another house). That would look the same as if anyone else brute-forced the puzzle legitimately; so as far as evidence goes, Theory #3 is indistinguishable from Theory #2.
All of the Bitcoin Puzzle addresses are (apparently) listed here:
https://privatekeyfinder.io/bitcoin-puzzle
This is the address for Bitcoin Puzzle #66:
https://www.blockchain.com/explorer/addresses/btc/13zb1hQbWVsc2S7ZTZnP2G4undNNpdh5so
And this is the winning transaction (IIUC):
https://www.blockchain.com/explorer/transactions/btc/57a88f47e4c047740b782a5562fca143ce85de0373cbff3a7d406e9ae7fc2f5f
My question is, Is there any “physical evidence” remaining that might corroborate which theory is correct? Specifically, if there were something like blockchain.com/explorer for unconfirmed transactions, then we could look at it and say “Oh, that’s interesting; an unconfirmed transaction against the prize address at 18:57:00 with a 50000-sat fee, followed by another unconfirmed transaction at 18:59:20 with a 75000-sat fee, and then a confirmed transaction at 18:59:39 with a 76400-sat fee, and then several more invalid transactions in the five minutes after that.” — That would be strong evidence for Theory #1 (IIUC); but as far as I know, there’s no public (or reliable private) record that would be able to provide that kind of information — right?
…Actually, I just found this stacker.news story, which points to an unconfirmed transaction 8c8ec6b and links to a page on mempool.space/tx as corroboration of Theory #1. That page is… well, it’s gone now. And it was never archived by the Wayback Machine nor by archive.is. So I guess I might be asking whether there’s any other historical record of the information ephemerally available on mempool.space; or perhaps some subset of such information “of historical interest.”
