[qnr]

A few weeks ago I made a script to look for transaction chains like this in the blockchain, and it turned out they have been going on every day since mid-2010. Currently they account for 20-25% of tx volume but the percentage was much higher in the past (up to 50-60% for some days in 2012). For most of 2014 it was about 15%.

I'm not sure what the sender of these transactions is doing (trying to inflate tx volume? some badly written software?strange mixing algorithm?) though even if you discard those transactions from analysis it doesn't change the overall picture much.

[sanswork]

It'd be great if you could post the data somewhere. I wanted to do a script to go through them but just haven't had the free time.

[qnr]

http://pastebin.com/whyuCDqi

Tab-separated columns are as follows:

- Date - Percentage of "strange" txs - Number of transactions with strange txs discarded - Length of the longest chain of strange txs

Note: coinbase transactions are ignored by the script as they are "involuntary" and do not represent economic activity.

[FatalLogic]

Thanks for posting this data. Here's a quick chart of it.

http://i.imgur.com/sgzVwYm.png

[mike_hearn]

They're just a wallet that's making lots of small payments. You're looking at a change address that gets used once.

I'm not sure why these sorts of transactions would be considered ignorable. If you have a large bank balance and most of your payments are for groceries, does that somehow make those payments irrelevant?

[qnr]

I'm talking about particularly large chains of transactions (like hundreds or even thousands of transactions within the same day, often with most of them included in just a few blocks).

If they actually need to pay to that many different addresses it would be far superior to make a single tx with multiple outputs.

My first theory was that someone is trying to inflate tx volume and the simplest idea they came up was scripting the Bitcoin client to make thousands of small payments in a row.

[martinko]

It can be just a popular service (hence the high number of transactions) with a standard wallet implementation, ie. one transaction per withdrawal. Yes this is inefficient, but I dont see a reason why to automatically assume that this is an effort to inflate transaction volumes.

[illuminate]

"I'm not sure why these sorts of transactions would be considered ignorable."

There's a great attitude for pumping!

[linhares]

Nice of you to completely change the subject; google "ignoratio elenchi" for your self-improvement.