Mismatch negativity (MMN) and P3b are well known for their clinical utility. There exists no gold standard, however, to acquire these markers. This may explain why the within-individual sensitivity of MMN/P3b is often quite poor and why seemingly identical markers can behave differently across studies. Here we compare two traditional paradigms for MMN or P3b assessment with the recently more popular local-global paradigm which promises to assess MMN and P3b orthogonally within one oddball sequence. All three paradigms were administered to healthy participants (N=15) with concurrent EEG. A clear MMN and local effect were found for 15/15 participants. The P3b and global effect were found for 14/15 and 13/15 participants, respectively. There were no systematic differences between the global effect and P3b. Indeed, P3b amplitude was highly correlated across paradigms. The local effect differed clearly from the MMN, however. It occurred earlier compared to MMN and was followed by a much more prominent P3a effect. The two sets of peak latencies and amplitudes were also not correlated across paradigms. We conclude that the local-global paradigm is effective in evoking the traditional P3b component, but it does not capture the MMN. Caution should therefore be exercised when comparing the local effect and MMN across studies. Nevertheless, the within-individual sensitivity of both MMN and the local effect was satisfactory. The within-individual sensitivity of P3b was lower than expected in a healthy control group, which may explain the often-low sensitivity of P3b in patients with disorders of consciousness.