The research project is motivated by the Singular-Target-Focus theory (Klar and Giladi 1997). In particular, it was verified whether this theory could account for comparative biases observed in unrealistic optimism. It was predicted that the size of comparative biases would depend rather on probability of events, than on their evaluation. The obtained results confirmed the hypothesis. For very frequent events (i.e. p equals ca 0.8), subjects evaluated higher the probabilities for individual object (Self or other, identified by a name, e.g. Pawel) than for a comparative group. In contrast, for rare events (i.e.p equals ca 0.1), subjects evaluated such probabilities lower. The comparative biases were reduced when participants were presented with events of moderate frequencies (i.e. p=ca 0.5). Moreover, such relations hold both for positive and negative events.