I've been perplexed as to why there is so much activity for the recently launched SatoshiDice.com betting site. Certainly, there are many, many options for wagering using bitcoins, as Mem's list shows: http://bitcointalk.org/index.php?topic=75883.0
. So why is this two-week-old site pulling in a thousand wagers or more each day now?
The reason I am so curious is that the number of transactions on the blockchain each day have skyrocketed recently:
And it appears much of the growth can be attributed to wagers placed with SatoshiDice.com:
The Recent bets table shown on the site shows the 200 most recent bets made but now that consistently shows four to six hours worth before hitting 200 rows. So that means between 800 to 1,200 wagers are placed on the site in a day (this may be off as I haven't measured each of the addresses to determine a full day's betting volume, but 800 looks to be safe to use as the lower bound.)
Each player making a wager sends a payment to an address for one of the bet types shown on the site. The service then runs a calculation and returns either a transaction to the winner that includes the appropriate payout or it returns a transaction to the loser that includes a tiny fraction of bitcoins as a consolation prize. The only way the player knows if the bet won or not is based on the return transaction.
So there are two transactions on the blockchain for each wager. If there were 800 wagers for the day then there were 1,600 transactions on the blockchain. With the total per-day activity to the blockchain exceeding 10,000 transactions recently, SatoshiDice wagering represents at least 15% of all blockchain traffic in a day.
In trying to come up with reasons why this site might be growing so fast, one of the suggestions offered to me was that it worked as a great mixing service. That explanation didn't register with me at first, as the spend transaction is returned instantly, so there is a direct connection between the payment sent to SatishiDice and the transaction that is returned.
A mixing service not only mixes coins but also has a time factor. Coins can be returned a little at a time to different addresses which makes tracing more difficult.
With SatoshiDice.com returning on average more than 99% of the wager amounts then a mixing strategy might be to send coins through multiple times. Some earlier passes will be winners and payouts with "clean" funds are sent on the return transactions. The coins used for wagers that lose get chewed up and only a tiny morsel is returned. After a few passes the makeup of the wallet at the end would be coins that look significantly different from the coins that were held before the wagering.
Could it be then that these blockchain-based wagering services are being used widely for mixing coins?