It is common for tropical depressions / hurricanes to be under-reported in a GRIB file. The typical grid used in a grib is 2 degrees, which can be as large as 120 NM. A tight depression can "hide" between the datapoints. I usually see some evidence of the depression in the surface
wind barbs, and sometimes in the pressure contours (in a grib), but the max windspeed reported in the grib is almost always going to be less than the actual windspeed.
You can increase the grib resolution to 1, or even 1/2 degree, but this is at the cost of a 4x or 16x increase in filesize. This is one reason to also get the wfax chart, or text
forecast, as a supplement to the gribs. I would also trust the storm track predictions of the wfax and text forcasts, more than the grib predictions.
I find it useful to regularly download gribs and wfax to my home computer, when I am not sailing. This way I gain
experience interpreting a wide range of data. Of course you have to actually study it, which I don't always do.