The GRACE satellite mission provides gravity field estimates of the Earth with unprecedented accuracy. Nevertheless,the realistic detection of oceanic mass redistribution remains challenging due to comparatively small signalamplitude, aliasing by tides and other short-term variability, and smoothing of small spatial scales. To verify thecapability of GRACE to measure oceanic mass variability, a validation with in-situ timeseries of Ocean BottomPressure (OBP) timeseries is essential.Here, different GRACE gravity fields provided by the GRACE Science Data System (CSR, GFZ, JPL), GRGS,ITG and others are compared with more than 140 timeseries of OBP sensors deployed throughout all oceans.The performance of the different GRACE products to capture oceanic mass variability is assessed by a weighedcorrelation analysis, taking into account the length and data quality of the in-situ time series. Both Gaussianfiltering and an ocean-model derived spatial pattern filtering method are used for the GRACE data, whereas forthe in-situ timeseries, different de-tiding and de-trending methods are applied to reduce aliasing and sensor drift.The analysis aims (a) to quantify the skill of different GRACE products and to quantify the advances made byrecent GRACE gravity field releases with improved data processing, and (b) to identify regions where GRACEperforms exceptionally well (e.g. high latitudes), and in which parts of the oceans GRACE fails to detect real OBPvariability. Spatial patterns related to the performance of GRACE may help to predict the quality of spacebornegravity measurements also for those oceanic regions where no in-situ data are available. This is critical for thefuture use of GRACE to remotely determine water mass redistribution in all oceans.