This coin is brand new. Even if that post is 100% true we are in the infant stages of the coin. Progress will be what drives this coin. My target is roughly 200k.
Yeah, these are certainly infant stages of this coin. But devs said that anon was done. So why did developers launch anonymity feature if it clearly doesn't work (assuming this reddit post is 100% correct)? Did they at least test it? Did someone at least review their solution and concepts behind it?
The method outlined in that reddit post seems incorrect based upon my own test but it does sound like the wallet in question only had a single input to begin with.
In my case it spent the final amount in 3 completely separate random values in the first instance, these were spent from 3 completely separate inputs received by 3 completely separate wallet addresses on my client.
Is it possible to find where the coins came from by looking back in the blockchain and adding up smaller values trying to find the original wallet address(es)? I'd have to say, perhaps. But trying to figure this out will get much harder when there are a lot of transactions moving around.
One area of improvement the developers could look to undertake would be to create 3 separate deposit addresses for the account that the randomized values go to. That way there is no record of the 3 randomized values going to the exact same deposit address but they will still go to the exact same account.
I don't think that the problem lies in the number of inputs.
Imagine mixer as a big black box. Nobody knows what is happening there.
You send X coins to the mixer. The recipient must receive Y coins. The remaining R=X-Y coins you receive at "change" addresses (one or more). So R = R_1+R_2+...+R_n (in case of KeyCoin it seems that R = R_1, i.e. there is only one change address).
So on the blockchain you can always find transactions that have outputs X, Y, R_1,....,R_n and the following equation holds:
X = Y + R_1 + R_2 + .... + R_n.
(this equation assumes that the mixer isn't stealing part of your coins).
So if you can find such transactions in vicinity (block height is not far away) of the transaction that sent Y coins to recipient, then the addresses (or inputs) that were used to send X coins belong to the original sender.
Of course you may find different transactions that satisfy equation X = Y + R_1+R_2+....+R_n, but the probability may be low (depends on actual implementation and the number of transactions with different amounts in outputs).
This is general problem with such mixers. This can be ameliorated in such implementations where mixer "steals" some random amount of coins from you (and yes, such mixers exists).