Bitcoin puzzle transaction ~32 BTC prize to who solves it

[ilgizilgiz]

Hi, could you open the repository on GitHub? How does this code perform the search? Does it use some kind of algorithm, or is it completely random?

Hi! The repository is now public:

https://github.com/IlgizIlgiz/bitkee

The search works in several modes:

  1. Manual mode - you select a specific page (position in the key space) and scan 128 addresses at once
  2. Auto mode - generates random pages automatically and checks balances via Blockchain.info API
  3. Puzzle mode - brute-force search within known key ranges for Bitcoin Puzzle addresses

The search is essentially random/brute-force — there's no smart algorithm to predict private keys (that would be cryptographically impossible). It's more of an educational/lottery-style tool to visualize how vast the Bitcoin key space is.

Each "page" contains 128 sequential private keys. The total key space is ~2^256, so finding a used address by random search is astronomically unlikely.

[ee1234ee]

I actually wrote and tested all the code for 135. Works quite well.
It’s the same dispatcher as for 67 and 68, with run parameters varying by GPU model.
There’s a central server accumulating all the DP. It was actually the most complex bit to write, because handling the load from tens of thousands of GPUs sending points there was quite the challenge.

135 is roughly as complex as 68, so with 20k GPUs like I had back then it’s a matter of a couple months.
Probably 300k ish investment for the 13.5Btc reward.

Hello expert
Since you have already written the program to solve problem 135, why not create a pool? Like btcpuzzle, this is a pool with exclusive bonus. How nice would you put the pool on the Internet and let everyone join? 
If you don't want to do it, can you open source the program on GitHub? Let's build a pool together to solve problem 135.

[kTimesG]

I actually wrote and tested all the code for 135. Works quite well.
It’s the same dispatcher as for 67 and 68, with run parameters varying by GPU model.
There’s a central server accumulating all the DP. It was actually the most complex bit to write, because handling the load from tens of thousands of GPUs sending points there was quite the challenge.

Since you have already written the program to solve problem 135, why not create a pool? Let's build a pool together to solve problem 135.

Unlike address puzzles, where each and every worker simply needs a key range from a server and that's basically all, ECDLP puzzles need significantly higher data transfers and lots of central I/O.

It's also extremely difficult to have an "interruptible" worker this way, because a full state must be saved as often as possible, and restored on restart (without crashes and without data corruption due to the sudden interruption). With address puzzles: only the progress reporting is ever needed, zero state saves, zero state restores, simply scan whatever range the server wants next.

What I'm saying here, is that the operations are very sensitive and complex, so I doubt you'll see a (serious) pool for 135 or above soon, simply because clients can disappear forever in the middle of some long task, and managing that case (on the server) is kinda chasing ghosts in a dark room, for the rest of your life.

I'd say the closest to a "pool" would be to invest $ and let the stuff silently run on fully-trusted workers that have a clear and controlled lifecycle, not on random PCs. It's the same level of trust, since the ECDLP can only ever be solved server-side, when (for the first time) some whatever DP is found again, but from a different walk type.

[Bram24732]

I actually wrote and tested all the code for 135. Works quite well.
It’s the same dispatcher as for 67 and 68, with run parameters varying by GPU model.
There’s a central server accumulating all the DP. It was actually the most complex bit to write, because handling the load from tens of thousands of GPUs sending points there was quite the challenge.

Since you have already written the program to solve problem 135, why not create a pool? Let's build a pool together to solve problem 135.

Unlike address puzzles, where each and every worker simply needs a key range from a server and that's basically all, ECDLP puzzles need significantly higher data transfers and lots of central I/O.

It's also extremely difficult to have an "interruptible" worker this way, because a full state must be saved as often as possible, and restored on restart (without crashes and without data corruption due to the sudden interruption). With address puzzles: only the progress reporting is ever needed, zero state saves, zero state restores, simply scan whatever range the server wants next.

What I'm saying here, is that the operations are very sensitive and complex, so I doubt you'll see a (serious) pool for 135 or above soon, simply because clients can disappear forever in the middle of some long task, and managing that case (on the server) is kinda chasing ghosts in a dark room, for the rest of your life.

I'd say the closest to a "pool" would be to invest $ and let the stuff silently run on fully-trusted workers that have a clear and controlled lifecycle, not on random PCs. It's the same level of trust, since the ECDLP can only ever be solved server-side, when (for the first time) some whatever DP is found again, but from a different walk type.

My main concern with a pool would be that verifying DPs are real and not forged is a computationally intensive task.

[kTimesG]

My main concern with a pool would be that verifying DPs are real and not forged is a computationally intensive task.

Unless you traverse each walk that produced a DP, you can never know if it was forged or not, because the walks may be forged as well.

The DP may be 100% valid, but the walk might not have jumped at all according to your params, which means the DP is useless (the real DP set is 100% dependent on the jump parameters).

[Bram24732]

My main concern with a pool would be that verifying DPs are real and not forged is a computationally intensive task.

Unless you traverse each walk that produced a DP, you can never know if it was forged or not, because the walks may be forged as well.

The DP may be 100% valid, but the walk might not have jumped at all according to your params, which means the DP is useless (the real DP set is 100% dependent on the jump parameters).

My thoughts exactly. So no way to be protected against a bad actor filling your DB with junk

[Niekko]

Hi! The repository is now public:

https://github.com/IlgizIlgiz/bitkee

The search works in several modes:

  1. Manual mode - you select a specific page (position in the key space) and scan 128 addresses at once
  2. Auto mode - generates random pages automatically and checks balances via Blockchain.info API
  3. Puzzle mode - brute-force search within known key ranges for Bitcoin Puzzle addresses

The search is essentially random/brute-force — there's no smart algorithm to predict private keys (that would be cryptographically impossible). It's more of an educational/lottery-style tool to visualize how vast the Bitcoin key space is.

Each "page" contains 128 sequential private keys. The total key space is ~2^256, so finding a used address by random search is astronomically unlikely.

really nice, thank you

[brainless]

"who buy me a coffee"
This sentence is begging in admin view and always face deleted post
But asking a script, program, app, developments,
These are not begging in admin views.....
Thinking.........😂

[optioncmdPR]

The keys in column, 4 per row.
I think it looks rather tidy. The question is, does this
structure mean anything
regarding the possibility of
deriving the next two  to finish  row # 18?


1378
15314ce0
1d3202483a7b
1460293068f3c936
1764f3080d5749fd2c55
1ba5342de40f556e52dc2a04
1fa5ee5340326e6ac3875d916ce8
17e2551e3d94cd647d4fe747b862a62e
1a96ca8d834a65911d4aed211709de820a7c
1757756a9322382facd04b5f8303e9e9ae4933d6
153869acc5b2a221c58d8f6bd3b27c591e02b35a358f
122fca143c052ec18388d5446cd610b53cbaade6d7ce3b9b
174176b015f4d22bd43c2e935475070a1a009d4efae164cb9e3c
180788e47e326c236fb6d5ad1f436abe1f9b67e1149d18b63ac4ffdf
1eb25c90795d61c2c675b852189a217496cbb87cab44ffc07a1825367bbe
13c96a3742f64906363d541eb611abee7cce5efdaccf6808f7051f27b09112d4
1a838b13505b268672832ed74f2b5e35ee730fc235c1942c1aebebb3940cd0fc1491
101d83275fb2bc7e0c349b84b6431a6c4ef1

[damiankopacz87]

I don't say it is illegal or scam either but from the other side our probable creator is missing since 2019 ...

Creator is not missing, he is reading Your posts from another account and spinning this blockchain gambling.

Mr Creator, please, confirm, are You here?

Is 256th address original, or was there 0 at the beginning of its key?

BR
Damian

[analyticnomad]

The keys in column, 4 per row.
I think it looks rather tidy. The question is, does this
structure mean anything
regarding the possibility of
deriving the next two  to finish  row # 18?


1378
15314ce0
1d3202483a7b
1460293068f3c936
1764f3080d5749fd2c55
1ba5342de40f556e52dc2a04
1fa5ee5340326e6ac3875d916ce8
17e2551e3d94cd647d4fe747b862a62e
1a96ca8d834a65911d4aed211709de820a7c
1757756a9322382facd04b5f8303e9e9ae4933d6
153869acc5b2a221c58d8f6bd3b27c591e02b35a358f
122fca143c052ec18388d5446cd610b53cbaade6d7ce3b9b
174176b015f4d22bd43c2e935475070a1a009d4efae164cb9e3c
180788e47e326c236fb6d5ad1f436abe1f9b67e1149d18b63ac4ffdf
1eb25c90795d61c2c675b852189a217496cbb87cab44ffc07a1825367bbe
13c96a3742f64906363d541eb611abee7cce5efdaccf6808f7051f27b09112d4
1a838b13505b268672832ed74f2b5e35ee730fc235c1942c1aebebb3940cd0fc1491
101d83275fb2bc7e0c349b84b6431a6c4ef1

No it does not mean anything. It takes awhile to admit to yourself that there is no secret, no structural sequence, and no hidden meaning in any of this. There is only straight up sequential scanning at a large scale. The sooner you choke on this hard to swallow truth, the better off you will be.

[optioncmdPR]

I feel that they are ,yes, "deterministic" , but due to a custom formula , and not related to any bip derivation schemes. Furthermore, I propose this custom formula has an output consistent with bitspaces and that the masking is used on integers 5 steps ahead to represent a bitspace 5 steps below it. One can easily claim  any set of addresses is a wallet. I also firmly believe that the creator carefully ommited defining the deterministic method, so that he cant be accused of being misleading should it turn out to be true.  And then there is a thrill in experimenting with maths . Taken as a hobby only , and not an obsession, I think its okay to dare conceive such a fantasy as winning a puzzle (or 4) , with mere logic. So shoot me dead, im having fun.

[ee1234ee]

The creator creating this set of puzzles is both a good and a bad thing
If someone becomes addicted to it and abandons their main business, it is very scary. In the end, they may have nothing, even lose their job and house, because you may not be able to solve any of the puzzles in the end.
Everyone should still treat it with a calm mind, use extra energy and money to play and give it a try, and not take a gamble, as it can easily lead to failure.

[hoanghuy2912]

Even now, many people are still obsessed with scanning keys, even though they know the success rate is almost zero. Along with that, the puzzle creator disclosed only some public keys for a few puzzles within large ranges, rather than for all puzzles in similar large ranges, accompanied by unconvincing reasons. This has already become an issue that few people are aware of.

[Bram24732]

Even now, many people are still obsessed with scanning keys, even though they know the success rate is almost zero. Along with that, the puzzle creator disclosed only some public keys for a few puzzles within large ranges, rather than for all puzzles in similar large ranges, accompanied by unconvincing reasons. This has already become an issue that few people are aware of.

“I don’t understand things so I call them unconvincing”

[hoanghuy2912]

Even now, many people are still obsessed with scanning keys, even though they know the success rate is almost zero. Along with that, the puzzle creator disclosed only some public keys for a few puzzles within large ranges, rather than for all puzzles in similar large ranges, accompanied by unconvincing reasons. This has already become an issue that few people are aware of.

“I don’t understand things so I call them unconvincing”

That's right, I don't know anything.

[damiankopacz87]

Even now, many people are still obsessed with scanning keys, even though they know the success rate is almost zero.

0,001 is almost zero and 0,00001 is almost zero. To be precise, if You have RTX 5090 and good software and You run it all day long (24h) Your chance to solve 71st puzzle is 0,0000002. In other words, You need 5.500.000 days / 15.000 Years to be 100% sure that You solve it (or 7.500 Years to have 50% chance )

135 is even worse.

BR
Damian

[brainless]

Even now, many people are still obsessed with scanning keys, even though they know the success rate is almost zero.

0,001 is almost zero and 0,00001 is almost zero. To be precise, if You have RTX 5090 and good software and You run it all day long (24h) Your chance to solve 71st puzzle is 0,0000002. In other words, You need 5.500.000 days / 15.000 Years to be 100% sure that You solve it (or 7.500 Years to have 50% chance )

135 is even worse.

BR
Damian

Let me repeat
For 70 bit puzzle should be 7500 years, 69 should be 4000 years, but here these puzzle found in months, not even 1 year
Think and jump to try your luck, remember if you have 1gb gpu or have ryx 5090 even hundreds gpu all have same chances to find, as reward is 1 key

[damiankopacz87]

70 was found in minutes becouse of Pubkey availability. 69 was found fast becouse Privkey was at very begining of range.

BR
Damian

[Fadymak]

Hi
You can't solve puzzles from 71 by GPU
it can only be solve by CPU
PIZZLE 71 WELL BE SOLVED SOON