Introducing a scoring model
Mon, 18 Nov 2019 12:57:19 GMT
We at Blender believe that the increasing expansion of so-called blockchain analysis services (BAS) poses the direst threat to the anonymity and privacy of Bitcoin users.
That's why we are starting a set of improvements today by introducing a scoring model, which is intended to fight blockchain analysis services. The scoring model brings a new algorithm of labelling transaction inputs.
The scoring model adds a mixing code based score to each input and recalculates scores each time an outgoing transaction is made, based on the input and output amounts of the transaction proportionally. This approach is similar to the one used for bitcoin labelling in above-mentioned BAS, so it happens that Blender uses BAS approaches to withstand analysis and provide more privacy to our customers.
Starting this update, not only the presence of a mixing code on a transaction input plays role on a decision whether it will be sent to a customer or not, but also it's score.
What does it mean to a customer? There are two main aspects:
- much more transaction security: the number of mixing cycles for each user now has unpredictable nature;
- there is no way BAS can now determine if your bitcoins are back to you: even if it happens, it's part in a transaction will be small enough to be a statistically indistinguishable from a random ocassion.
Of course, all this will work for you only if you always use your mixing code.