In my case it was (of course) the change address in a transaction I spent using a p2pool payment address and an emc payment address.
But the article doesn't make sense about tracking people.
Yes until I spend my own change, there is a (very small) doubt that the change address of my transaction was indeed in my wallet and if I spend any of the 1 Satoshis that transaction may confirm it - but that's no different to saying that spending my own change will confirm it's my change ... and ... do people NOT spend their own change?
Anything you do to protect your identity when spending the change you can simply add the 1 Satoshi to it also.
So, yeah I don't see the issue mentioned in the reddit article about
...
presumably in an attempt to more easily map the transaction graph (i.e., remove anonymity from the transactions)
...