Need help with bitcoin seed phrase recovery using 3rditeration BTCrecover

[kauai_lost_boi]

Back in late 2011 I got some bitcoin from New Liberty Standard. After researching the best way to put my coins in cold storage I resorted to using Electrum in Jan 2012 to verify my coins using an offline watching wallet. I then made a BIP 39 12 word mnemonic seed phrase. I also added more coins to the address but never verified and basically left it alone for years.

I created a txt file with the seed phrase but some of it got corrupted. Fortunately I stumbled across 3rditeration BTCrecover, who confirmed I am well within the recovery parameters as I am only short 3 words. I got a good guess as to what they might be but I'd rather deal in certainties.

I installed all the necessary software on a decent setup, got it up and running on a 44 core, 40 gb ram system and after testing, I have 24 workers available. My problem is I need to narrow down the search parameters otherwise I may be left with hundreds of billions of possibilities. I'm probably doing things wrong.

I was tinkering around with the seedlist/token file approach using symbols like "^ +" but it didn't like what I gave it to chew on. I'm also not sure of the derivation path let alone some of the commands I should pursue in my working argument for cmd line

Here is the situation with my 12 word bip39 seed phrase

The first 2 words are known and are a lock

I only know one of the the next 4 words and not sure which of the 3rd or 4th position the known word fits

the 7th and 8th are known but not sure of their exact position

same for the 9th and 10th

The last 2 words are known and are a lock


I know I'm close to having a breakthrough. If anyone out there is reading this and sees any possible methods of resolution, I'm here until this gets done.

[bitmover]

Here is the situation with my 12 word bip39 seed phrase

The first 2 words are known and are a lock

I only know one of the the next 4 words and not sure which of the 3rd or 4th position the known word fits

the 7th and 8th are known but not sure of their exact position

same for the 9th and 10th

The last 2 words are known and are a lock


I know I'm close to having a breakthrough. If anyone out there is reading this and sees any possible methods of resolution, I'm here until this gets done.

So you have something like this:

word1 word2 3 4 5 6 word7/8 word7/8 word9/10 word 9/10 word11 word12

Do you have any idea about word 3 4 5 and 6? Do they start with A, or BA, etc?

if you know, for example, the first letter of those words you can narrow your search considerably.
The llist of available words is here:
https://github.com/bitcoin/bips/blob/master/bip-0039/english.txt

brute forcing 4 words without more information would be hard. But you still need to try that changing words 7/8 9/10 ....

[kauai_lost_boi]

Thank you for the response! It's more like

1st word / 2nd word known and locked

3rd 4th 5th 6th words need processed but 1 word should be in at least the 3rd or 4th MAYBE but unlikely 5th positions. Got a good idea what 2 other words may be but want to be sure

7th & 8th are known & likely in correct positions but want to be sure

same for 8th & 9th words & positions

11th & 12th words are known and are a lock

Tried working out a seedlist token for btcrecover but still having trouble. I know I'm close, just need to get that last bit of advantage by correctly creating the cmd line argument to go along with my info.

[LoyceV]

I created a txt file with the seed phrase but some of it got corrupted.

I'm curious: how did this happen, and how do you still know for sure some words are correct?
I can understand if it was on a shredded piece of paper, but not when it's a text file.

[o_e_l_e_o]

I created a txt file with the seed phrase but some of it got corrupted.

For future reference - don't do this! Write it down on paper.

I was tinkering around with the seedlist/token file approach using symbols like "^ +" but it didn't like what I gave it to chew on.

Can you paste your tokenlist file and we can take a look at where you might be going wrong? Obviously change your actual words to "Word1", "Word2", etc. before you do!

Based on what you've told us so far, I would think your tokenlist file should look something like this:

Code:

^1^Word1
^2^Word2
^3^Word3 ^4^Word3 ^5^Word3
^7^Word7 ^8^Word7
^7^Word8 ^8^Word8
^9^Word9 ^10^Word9
^9^Word10 ^10^Word10
^11^Word11
^12^Word12
abandon ability able about above absent absorb abstract absurd abuse access accident account accuse achieve acid acoustic acquire across ...
abandon ability able about above absent absorb abstract absurd abuse access accident account accuse achieve acid acoustic acquire across ...
abandon ability able about above absent absorb abstract absurd abuse access accident account accuse achieve acid acoustic acquire across ...

As far as I am aware, there is no functionality on btcrecover to make a line say "Try any BIP39 word", so you just have to paste the entire BIP39 word list on to the last three lines for these three unknown words. Someone can correct me if I am wrong here.

This is not a completely rivial task. 3 unknown words is 2048³ combinations. With words 7/8 and words 9/10 possibly being switched, that's another factor of 4, and with word 3 being in one of three positions is a factor of 3. And this all assumes that the words you know are definitely correct and otherwise definitely in the correct position.

[NotATether]

As far as I am aware, there is no functionality on btcrecover to make a line say "Try any BIP39 word", so you just have to paste the entire BIP39 word list on to the last three lines for these three unknown words. Someone can correct me if I am wrong here.

You are correct, BTCrecover will only look for words on that line if it has positional anchors on it.

Though I was thinking, that somebody can make a feature of putting a wildcard ("*") at the end of the line and that will search for all of the words in the specified wordlist.

I think BTCrecover can detect if it's a BIP39 or Electrum seed that needs to be recovered, by looking at the wallet file or the question answers, but I'm not sure.

[kauai_lost_boi]

To LoyceV

It's a messed up story how all this happened, including recovering from a brain injury at the time. When I was doing all this I had the memory of a goldfish and the common sense of the Amoeba boys from the Powerpuff Girls. I saved the seed txt file onto a 3 1/2 inch floppy and disguised the text using wingdings. Years later, well you can imagine what it was like and probably play the Benny Hill theme song in the back ground for added effect.

To Oeleo

I've been talking with others including Crypto Guide/3rditeration and we had some major breakthroughs. We confirmed I used the Electrum seedlist which is nearly identical to the standard bip39 list, only a few words are different in the middle. So I will be using electrum1 or electrum2 for the wallet type in the future argument.

The format for the seedphrase token list I was trying was something like

+ ^17^Guinness Seventeenth_to_eighteenth_(but_never_last)

+ ^19^Motorhead Nineteeth_or_bust

It didn't like the + symbols and had to do other things to get the ^ to agree with the program. I was moving more towards your suggested format but wanted to understand how the syntax worked for the specific arguments. Now I'm beginning to understand. I didn't know I was supposed to throw in 3 lines of the seed word list at the bottom of the token and that would have been one major issue on the many others I came up against. Thank you for setting me straight and working out that template for my future token file.

If the 7th 8th 9th and 10th words being swapped are gonna make this a bigger mess, Ill confirm further what I got for at least 9 and 10. The one word I do know is likely the 3rd spot. I have an idea what the other words are. Altogether I'm really only unsure about 2 words and their positions. I'm 90% sure I got everything I need I just wanted to let the machine do most of the confirmation work since my brain is as reliable as Acme products in the use of a certain coyote.

[PawGo]

Here is the situation with my 12 word bip39 seed phrase

The first 2 words are known and are a lock

I only know one of the the next 4 words and not sure which of the 3rd or 4th position the known word fits

the 7th and 8th are known but not sure of their exact position

same for the 9th and 10th

The last 2 words are known and are a lock

Please take a look at my program: https://github.com/PawelGorny/lostword
if any of ready 'solvers' are suitable to your problem definition. If not, I may try to adapt it to make it ready to run.
Let me know if my reasoning is correct, the words configuration is like this:

Code:

known_word1
known_word2
unknown-any word
unknown-any word
unknown-any word or word X
unknown-any word or word X
unknown-any word or word Y
unknown-any word or word Y
unknown-any word or word Z
unknown-any word or word Z
known_word11
known_word12

It means that in the worst case you search for 5 completely unknown words * 6 different positions of suspected words, but maybe I do not really understand your situation. Examples above are more optimistic.

Anyway I think you may use solver 'POOL' from my app.

[o_e_l_e_o]

It didn't like the + symbols and had to do other things to get the ^ to agree with the program

So the "+" symbols should be unnecessary when making a tokenlist for a seed phrase as opposed to a tokenlist for a password or passphrase. The "+" symbols denote that btcrecover must include that line somewhere in the password, but since we are attempting to descramble a 12 word seed with a 12 line tokenlist, then it will automatically use one entry from each line anyway.

The "^x^" symbols denote that whatever follows should be attempted in that position. Since we know Word1 appears at position 1, then "^1^Word1" on the first line of my example does just that. Since we aren't sure if words 7 and 8 are swapped, then the line "^7^Word7 ^8^Word7" says to either try Word7 in position 7 or position 8. The last three lines without a "^x^" can be attempted anywhere in the seed phrase, but given that every other position is accounted for, the only places left for them to be tried are in positions 3, 4, 5, or 6, depending on the position of Word3.

I didn't know I was supposed to throw in 3 lines of the seed word list at the bottom of the token and that would have been one major issue on the many others I came up against.

So when creating a tokenlist for a 12 word seed, you should have 12 lines in it. btcrecover will try exactly one entry from each line. The bottom lines - every BIP39 word without any additional arguments - basically says "try any word anywhere it will fit".

If the 7th 8th 9th and 10th words being swapped are gonna make this a bigger mess, Ill confirm further what I got for at least 9 and 10. The one word I do know is likely the 3rd spot. I have an idea what the other words are. Altogether I'm really only unsure about 2 words and their positions.

If that's the case, then you might be better off just making a tokenlist file with only two unknowns and running it. If it doesn't return the answer, you could manually swap words 7/8, or 9/10, or move word 3, or whatever, in your tokenlist and run it again. There are also other programs, like the one mentioned above (I have never used it though so cannot vouch for it) or The FinderOuter which are potentially easier to use for this simplified case.

[kauai_lost_boi]

Thanx to everyone for their gracious insight and help in me doing this the right way the next go around. I'm gathering all the information from multiple sources and gonna throw everything I got at the box tomorrow. I will post the working arguments and the seed template before moving forward and try to do everything in real time. Fingers crossed peoples. If I get results tomorrow I'll become a church goer again 

[o_e_l_e_o]

I will post the working arguments and the seed template before moving forward and try to do everything in real time.

Once we've taken a look and settled on a final tokenlist, you can use the option --listpass to test your tokenlist produces the kind of combinations you want it to, before spending days running it to find out it isn't doing what you think it is. There is more info here: https://btcrecover.readthedocs.io/en/latest/TUTORIAL/#testing-your-config

I also wanted to ask: I presume you know the address the coins are on from your watch only wallet? Makes the whole process that bit easier than having to use an address database.

[kauai_lost_boi]

I will post the working arguments and the seed template before moving forward and try to do everything in real time.

Once we've taken a look and settled on a final tokenlist, you can use the option --listpass to test your tokenlist produces the kind of combinations you want it to, before spending days running it to find out it isn't doing what you think it is. There is more info here: https://btcrecover.readthedocs.io/en/latest/TUTORIAL/#testing-your-config

I also wanted to ask: I presume you know the address the coins are on from your watch only wallet? Makes the whole process that bit easier than having to use an address database.

Good day to everyone and hope you're enjoying your Monday or Sunday

So I'm working out all the routines and scenarios but I keep coming back to the bip39 list on electrum.  

The words I used in my seed can only be found in the bip39 list unless I'm not looking in the right places. I'm also not completely confidant in what I think I used to generate the seeds.  I do however have that memory of comparing the 2 different word lists and making the choice to go with what I have now. But I'm sure I used the electrum client to check the watching wallet in verifying funds.

Some of the words I was playing with that I remember were

turtle wedding husband

2 words can only be found in the bip39 list and one can only be found in the electrum. I may have settled in on the choice of a bip39 list and just used electrum as a secondary utility.

Unfortunately I lost any old addresses and must rely on the generated address database, so what would be my best arguments for cmd line?

So far I'm working with

Code:

python seedrecover.py --no-dupchecks --mnemonic-length 12 --language EN --dsw *--wallet-type BIP39 or electrum1 or electrum2*  --addr-limit 1 --bip32-path *if it is electrum then leave it blank or best derivation path for what I got*   
 *I have tokens that may reduce this value* --seedlist .\seedtoken.txt       Removing  the --typos 3 --big-typos 3 from the argument because it's a redundant command when working with a seed token file as confirmed by Crypto Guide

the format for the tokenfile I'm working on goes like

Code:

^1^Word1
^2^Word2
^3^Word3 ^4^Word3 ^5^Word3
^7^Word7 ^8^Word7
^7^Word8 ^8^Word8
^9^Word9 ^10^Word9
^9^Word10 ^10^Word10
^11^Word11
^12^Word12

I can be more confidant (presumptuous) and assume my memory isn't that bad and make a token file with only words 5 and 6 as big typos

Code:

^1^Word1
^2^Word2
^3^Word3 
^4^Word4 
^7^Word7 
^8^Word8
^9^Word9 
^9^Word10 
^11^Word11
^12^Word12 

I will certainly be checking out the listpass command function for sure. I'm not opposed to running multiple arguments over time or letting one run for weeks if necessary. Let me know what I may be missing or any other approaches I might not be seeing

[kauai_lost_boi]

Im running the arguments and it halts as soon as it gets to the 3rd word hurdle. I ran the listpass tests to see if my seed files are correctly configured and the same thing happens at the 3rd word hurdle. So now I gotta configure the seedlist token file, figure out my exact wallet path whether it be bip32 or bip 44, and the precise derivation path whether it be m/44'/0'/0'/0/0 or something else

[o_e_l_e_o]

Unfortunately I lost any old addresses and must rely on the generated address database, so what would be my best arguments for cmd line?

To use an address database your simply use the command --addressdb .PATH.db. See more here: https://btcrecover.readthedocs.io/en/latest/Creating_and_Using_AddressDB/

So far I'm working with

I've not tested it, but on a quick look it seems fine to me. The only thing I would pick you up on is the address limit of 1. Given that you have lost your old addresses, and this was so many years ago, can you be absolutely sure that you used the first address in the wallet? Would it make sense to bump it up to 5 just to be on the safe side?

There is a typo in the second tokenlist you posted in code tags, as you have both Word9 and Word10 appearing in the the 9th position.

[kauai_lost_boi]

Unfortunately I lost any old addresses and must rely on the generated address database, so what would be my best arguments for cmd line?

To use an address database your simply use the command --addressdb .PATH.db. See more here: https://btcrecover.readthedocs.io/en/latest/Creating_and_Using_AddressDB/

So far I'm working with

I've not tested it, but on a quick look it seems fine to me. The only thing I would pick you up on is the address limit of 1. Given that you have lost your old addresses, and this was so many years ago, can you be absolutely sure that you used the first address in the wallet? Would it make sense to bump it up to 5 just to be on the safe side?

There is a typo in the second tokenlist you posted in code tags, as you have both Word9 and Word10 appearing in the the 9th position.

I only had an electrum xpub to verify the initial funds back when I very first received it. Then there might have been one more transaction of bitcoin being added to an address but can't confirm. I can increase the addy limit if needs be. Right now the argument and the seedlist token file ain't having a friendly Sunday playday

Also, do I need to include the entire bip39 wordlist at the bottom of the token for it to run probabilities or will it do that from its own built in list?

[o_e_l_e_o]

I only had an electrum xpub to verify the initial funds back when I very first received it.

Do you still have the xpub? That's even better than having the address.

Also, do I need to include the entire bip39 wordlist at the bottom of the token for it to run probabilities or will it do that from its own built in list?

I was just looking in to this as being a possibility for why you are running in to problems. I think you probably do, but I'm not able to test that for myself at the moment. It could be the reason it is failing when it comes to Word3 is because after it decides if it is going to try it in position 3, 4, or 5, it has nothing telling it what to try for the other positions. All the seed recovery examples on https://btcrecover.readthedocs.io/ specify at least 12 lines for a 12 word seed phrase. It's worth a try.

[kauai_lost_boi]

I wish I had the old info like the xpub or even one of the addys. I would have had this done years ago. But that was when BTC was worth $6000. Maybe it was worth the wait 

 

I got btcrecover to run and give me an ETA which is far better than the last week of attempts. But then it hit a snag when it came to the 3rd 4th 5th 6th word lines. Mainly the one word I do know that should be in the 3rd 4th or 5th position.
^1^locked
^2^locked
^3^maybe ^4^maybe ^5^maybe (maybe being the same known word. {I'm probably doing it wrong [again]})

It's been a productive morning. I got more lines in command than before!

[kauai_lost_boi]

Happy Monday to everyone. So I've been using advice from many sources, links and webpages to try to refine the btcrecover seedlist token file to work correctly, unfortunately it's stuck at the 3rd and likely most volatile lines where it needs to fit at least one word out of 4 spots. Been tinkering around with any avenue of approach but I'm not giving the program the right things to even get a nibble. I'll be checking to see if anyone has a method or insight but for now I'm kind of in a corner playing tiddley winks 

[o_e_l_e_o]

Have you tried including 3 lines of the entire BIP39 word list as suggested sbove?

Another thing you could try would be to change the positional anchor for that word to just ^3^ and run your file. If you don't get a result, change it to ^4^ and run it again. And so on.

Or you could try removing the positional anchor for that word altogether. Given that every other position other than 3, 4, 5, and 6 is accounted for, that word would only ever be tried in those 4 positions anyway.

[kauai_lost_boi]

I've tried so many different ways with the token I've lost count or the ways that didn't work before may work now with a different argument. What I will do tomorrow is post pics of the examples and the cmd window for each instance. Hopefully we can narrow down the approach to get a working configuration.