Author

Topic: Easter egg (Steganographic Transactions) (Read 1483 times)

sr. member
Activity: 384
Merit: 258
April 06, 2015, 01:53:01 PM
#7
SPOILER. Don't read this post if you don't want to know the answer.











These 4 transactions look like classic payments:
- several inputs controlled by a payer are merged
- an amount is sent to an address controlled by a payee
- change is sent to an address controlled by the payer

Actually, it's possible (likely ?) that these transactions were classic payments, but they might be something else: "steganographic transactions"

Let's see the 3rd transaction:
- The 1st output (19z5fD6LhhiRupqezw7vi3fuumt5jCS9LU - 0.01 BTC) seems deterministically linked* to the 2nd input (1P3RfYxRTkTLdwXAVYzh41sfyMgzppELZA - 0.01 BTC)
- The 2nd output seems deterministically linked* to the others inputs

It means that this transaction might be:
- a classic payment transaction built by a single user
- a manually crafted transaction merging a txo controlled by a user A (the 2nd input) with txos controlled by a user B (the others inputs). No payment is done. The transaction just sends the coins to others addresses controlled by the users.
- ... (more weird scenarii)

This second interpretation has some "fun" properties:
- detection of this pattern is quite hard for human eyes
- it breaks the "merged inputs" heuristic used by some tools in order to clusterize addresses in wallets


A few remarks:
- The 4th transaction is similar to the 3rd transaction (no fee)
- The 1st and 2nd transactions have the same property but they pay a fee and the pattern is even more difficult to detect.
  Example: In the 1st transaction, the 3rd input (14okJQwaHJ3xHBtdU3LxqUEuXcsHhz9gtE - 0.01301568 BTC) seems deterministically linked to the 1st output (1Njw6FuxuVk293LwYREHxvUVUhx5MfzJLf - 0.01251568 BTC)


It remains a "mystery" for the 3rd and 4th transactions:
If they're classic payments, I don't know why the wallet has added an additional input/output. Hypotheses:
- a feature of the wallet, ensuring that there's always a minimum of 2 outputs ?
- a bug in the algorithm selecting the inputs ?
- manually crafted transactions ? Smiley


I wouldn't be surprised if someone already discussed this pattern. On my side, I've spotted the transactions this morning, while doing some tests, and found it was a funny coincidence (because, you know...easter eggs).


*: "deterministically linked" means the input and the output are linked whatever the correct interpretation of the transaction. I wrote "seems" because, with some more advanced scenarii, this statement might be proven wrong.
sr. member
Activity: 384
Merit: 258
April 06, 2015, 10:54:21 AM
#6
It's not about the size.

Another hint: steganography (...but don't lose your time with a message hidden in the transactions)
legendary
Activity: 2632
Merit: 1094
sr. member
Activity: 384
Merit: 258
April 06, 2015, 10:39:33 AM
#4
Another hint: it's not bad for privacy
sr. member
Activity: 384
Merit: 258
April 06, 2015, 08:39:05 AM
#3
I dont want to search so long but maybe you mean that you can, with the use of change addresses, find out, with a high certainty, more addresses someone owns?

Change addresses are a risk to anonymity in my eyes.

Only guessing...  Tongue
It's not related to change addresses.

Another hint: the title of the post may help  Wink
legendary
Activity: 2674
Merit: 1083
Legendary Escrow Service - Tip Jar in Profile
April 06, 2015, 08:05:26 AM
#2
I dont want to search so long but maybe you mean that you can, with the use of change addresses, find out, with a high certainty, more addresses someone owns?

Change addresses are a risk to anonymity in my eyes.

Only guessing...  Tongue
Jump to: