Yes that is correct.
And it is also correct to assume that predictions don't represent the actual observed weather parameter.
But this seems a systematic underestimation of 30-40%. That can occur one day, maybe two but not in such long row.
So you could start by observing the prediction (-2 days) versus the observation (of high average windspeed) on the actual day. And maybe compare this to an official measurement close by, possibly a handheld equipment to calibrate yourself. Wind does constitute a significant part of pwsFWI when temperature and dryness are such that wood (litter and undergrowth in general) becomes easily ignitable.