The rating is really only important for historical/statistical purposes. I am not a huge fan of the EF scale. Yes, NWS personnel are trained and can bring in other experts to help, but they are not civil/construction engineers. So depending on availability, an inexperienced survey crew might not have a very accurate rating. But mostly, the severity or strength of the tornado is tied to what it hits. And oftentimes it is nothing but open fields/dirt. Or it's a 220 mph monster but only hits poorly constructed buildings thus stays EF3 or EF4 when in reality, it may have been an EF5 had it hit the rare structure that qualifies as well-built.
I think many tornadoes are underrated in intensity because of it.
Which is why is important to remember that the EF-scale is not a wind-measurement scale.
The rating is on the damage, and not the tornado itself, directly speaking at least.
As far as using radar for ratings go, radar winds aloft are not that well correlated to what is happening at the surface. Severe Thunderstorm Warning verifications are all the further you have to look on that point.
Even the DOW measurement from Greenfield that set the internet on fire was still 144 feet in the air, and it was parked right outside of town! Then it's what fraction of 1% of all tornadoes have a DOW right on top of them?
Radar is sampling such an infinitesimally small percentage of tornadoes at anything close to approximating surface wind speed that it would add a whole new mess and layer of inconsistency to everything, so it's not a magic wand either.