Operating data from the Syncrude fluid cokers have shown that reducing the fluid coker temperature increases yields of valuable liquids and reduces sulphur emissions. It can, however, lead to the catastrophic loss of bed fluidity called "bogging". This paper develops a non-invasive method to monitor the fluidity of the fluidized bed from pressure signals. A pilot plant coker was used to test this method with bogging index measurements performed on coke samples. Spectral analysis, the signal entropy, its correlation dimension, cycle analysis and the S statistic failed to provide an accurate estimate of the bed fluidity. A new signal analysis method, the W statistic, which compares small and large fluctuations, provided a quick and reliable evaluation of the bed fluidity from dynamic pressure signals. (C) 2003 Published by Elsevier B.V.