Searching for Weak Bitcoin Keys and Low-Entropy Addresses

[Plogan]

Hello everyone!

I recently came across a fascinating (but mostly forgotten) topic on: https://bitcointalk.org/index.php?topic=2488493.75;wap, discussing weak Bitcoin addresses — wallets generated using predictable or low-entropy methods. Although this issue was first noticed 5–10 years ago, I’ve found hints suggesting that some of these addresses are still being used for fund transfers even in the past couple of years.

I am not a professional cryptographer or blockchain expert — just an enthusiast deeply interested in this subject. I would like to gather like-minded people to research, share ideas, and possibly discover new methods for detecting weak or predictably generated addresses.

Why This Matters

Historically, insiders or attackers have sometimes used predictable key generation methods (or perhaps partially exploited certain patterns or bugs) to stealthily steal funds. Even today, in 2025, I suspect that weak key vulnerabilities may still exist in Bitcoin and other cryptocurrencies.

Despite a few isolated discussions, there has been very little deep exploration into this topic. Most of the conversations ended after a few examples were found, but it’s very possible that the problem has simply evolved — with new predictable patterns that have yet to be studied or understood.

Known Methods of Predictable Key Generation:

Several fascinating examples already exist where public or semi-public data was used to derive private keys:

Using a Bitcoin block hash as a private key
Example: Genesis block hash → derived address.

SHA-256 of an existing Bitcoin address as a private key
Example: SHA256("1A1zP1eP5QGefi2DMPTfTL5SLmv7DivfNa") → derived address.

Using a block's Merkle root as a private key
Example: Merkle root of block #229670 → derived address.

Using a transaction hash (txid) as a private key
Example: Specific txid → derived address.

Reusing Bitcoin block hashes for Ethereum wallets
Example: Bitcoin block #100 hash used as an Bitcoin private key.


My Hypotheses About Possible Use Cases:

1. Experiments by early crypto exchanges
(e.g., generating keys that could allow a password reset for users who lost access)

2. Insiders creating weak wallets intentionally for hidden BTC thefts

3. Exploitation of possible unknown patterns or bugs
(maybe even something deep in blockchain mechanics)


What I Propose

I would like to form a small research community focused on:

Analyzing historical and current blockchain data,

Developing hypotheses on new predictable generation patterns,

Studying entropy vulnerabilities,

Testing scanning methods (possibly I can using full nodes) to identify anomalies.


I have created a Telegram group to kickstart discussions:
https://t.me/WeakKeysLab

Final Thoughts

This is a massive challenge and might sometimes feel like chasing shadows — but even the chance of finding a single active low-entropy address is incredibly exciting. Furthermore, improving our understanding of such vulnerabilities can only strengthen the overall security of the crypto ecosystem.

If you have experience in:

Blockchain analysis,

Cryptography,

Entropy research,

Programming (especially for parsing large blockchain datasets),

or if you're simply curious and want to learn together, you are more than welcome to join!

[ABCbits]

Even today, in 2025, I suspect that weak key vulnerabilities may still exist in Bitcoin and other cryptocurrencies.

FWIW, the vulnerabilities actually lies on software used to generate Bitcoin/altcoin address where Bitcoin/non-centralized altcoin can't stop people from using address with bad security.

Most of the conversations ended after a few examples were found, but it’s very possible that the problem has simply evolved — with new predictable patterns that have yet to be studied or understood.

Isn't it mainly because of these reasons?
1. People may use entropy/data related with their life (e.g. book they like to read or their favorite musician).
2. Many weak RNG usually use few same data source (e.g. current date/time or process ID).

[apogio]

Isn't it mainly because of these reasons?
1. People may use entropy/data related with their life (e.g. book they like to read or their favorite musician).

Most of the time, the problems occur when people try to generate randomness with something that they can remember. It doesn't work this way. There is no true randomness, but generating entropy from things that you can memorise leads to the exact opposite thing than randomness. I guess we could say it leads to predictability 

[hd49728]

If you use a trusted wallet softwares for generating your wallet, and addresses offline, you don't have to concern about this issue too much.

J. Lopp has an article How Many Bitcoin Seed Phrases Are Only One Repeated Word?.

If you read his article, see his warnings, you won't use those cool but dangerous wallet mnemonic seeds with 12 or 24 repeated words.

[satscraper]

~

Regarding low-entropy addresses , yeah, some wallets may still generate predictable private keys due the various flaws in architecture they used. The latest example of such flaw is "seeding mechanism used in Libbitcoin Explorer 3.0.0 through 3.6.0", i.e. Mersenne Twister PRNG was seeded with just 32 bits of system time.

 Effected wallets: