pbremen01: It's amazing how full of crap you are and any time you are challenged to prove something you always have an excuse.
Don't worry I'm in the process of setting up a test for you now with several completely unconnected wallets. If you can't deanonymise it yourself due to [insert excuse here] then feel free to pass it on to one of your friends who claim that they can do so.
Hell I'll even provide all of the wallet.dat files upon the conclusion of the test so that it can be independently verified if the test has been a success or failure. (After they've been emptied of funds of course).
If you send the same amount at the same time then of course this program could print the wrong addresses. What is here to prove?
But by manually sending the same amount of coins you're actually manually replicating other technologies for anonymization that are not implemented in KeyCoin. In fact, you're actually claiming that there exists better methods for anonymization than these that are implemented in KeyCoin (for example, CoinJoin with "denomination").
And your method isn't practical at all. If you own 10 coins, then with your method you can't send 10 coins anonymously (because you need to have at least (for example) 50 coins to send 5 separate transactions of 10 coins).
Of course my method isn't practical for the average user but it will prove that you can't really track from the destination to the source in this kind of scenario which would likely be replicated when there are many transactions moving at any one time.
The only thing that makes it possible currently is a blockchain with very few transactions. The program merely makes an educated guess on the source transactions and therefore the wallet address(es) that sent it in the first place. It doesn't disprove the anonymity solution works, it is merely exploiting the lack of transactions to jump to that conclusion easier.