Not sure what aspect of that is actually an error. All we know is that we judge ourselves different than others, but there is no information whether we judge others accurately or ourselves accurately.

In the past when people used to be stricter, there was a saying nemo iudex in causa sua, nobody be allowed to judge in his own case, because we would be far, far too lenient with ourselves, finding excuses. Today, we live in a world where being soft and forgiving is more fashionable, so today it would be more popular to think we judge ourselves correctly and others far too harshly. But besides these changing fashions of sentiment, do we actually know?

Assuming you, like most modern people, accepted the soft trend, and you think the lenient judgements we tend to give to ourselves are correct and should also be extended to others, have you ever tried to consider the other leg of the dilemma, i.e. what if we judged others as harshly as usual but ourselves too, maybe that would be the most accurate?

The important thing to understand is that prediction and moral judgement are different things. Circumstances predict better than personalities people's behavior. But this simply means being as good as the average is still very bad, allowing our behavior to be made bad by bad circumstances, like the vast majority does, is still something deserving censure. So this line of thought only works if you accept the average, typical person is a bad person - that predicting non-avereage behavior does nothing for judgement.