A well-known feature of observed and simulated climate is enhanced land surface warming compared to the ocean. This difference in warming, frequently expressed as a ratio, is often contrasted between the observed record and climate model output as both an evaluation metric for climate models and as a global index used for climate change detection and attribution. Latest simulated estimates of the ratio use full global coverage and marine surface air temperature, making genuine comparisons with observations difficult, since global observed datasets typically use sea surface temperatures and have limited spatial coverage. We show that re-calculating simulated ratios when using sea surface temperatures and limited spatial coverage (to resemble the observations) raises the ratio by ~0.25. We also update the observed ratio using latest observations and we find a close convergence of observed and simulated ratios towards ~1.6 for the 2000-2016 period accompanied by a decline in temporal variability. If we revise estimates of the likely range of ratios from climate models to account for the above factors, then our new observed ratio estimate is slightly less than the median of the GCM ensemble range (1.54 to 1.81).