Drillstrings can experience a number of detrimental dynamic phenomena. Their causations are primarily related to: the strings geometry, borehole diameter, physical properties of steel, and the instantaneous drilling parameters. The detection of such phenomena at the surface is very limited. The enormous lengths drillstrings usually achieve cause their spring like mechanical characteristic. As a result downhole dynamic effects are either considerably attenuated while transmitted to the top end or even do not affect it at all. Latest drilling technology enables downhole detection of dynamic dysfunctions. Sensors placed close to the bit provide the necessary data. The available rate for data transmission from downhole to surface is by far too low to transmit in real-time the whole quantity of sampled data. Therefore, relevant data interpretation must already be performed downhole. Stable diagnostics algorithms are executed for this purpose. A currently used algorithm to diagnose stick-slip has shown some deficits at its diagnostics ability under certain conditions. This work examines the whole course of automated downhole stick-slip detection - from data sources to diagnostic words. The diagnostics algorithm itself, as an elementary part of the entire stick-slip monitoring sequence, is extensively analysed and all discovered faults are discussed. The development of a new approach for stick-slip identification and severity classification finally concludes the work.