There's a few things happening here, but it boils down to EV range being just barely adequate so people are dependent on the range estimate and a 10% error is s

There's a few things happening here, but it boils down to EV range being just barely adequate so people are dependent on the range estimate and a 10% error is sometime a serious problem. This is especially true if you need to drive in conditions which disfavour an EV. To start with, if you are driving a significant fraction of the ideal range, then unexpectedly short range is a problem. Most ICEs have so much range this is never a concern except on long road trips, where refueling is easy and quick. Then EVs are relatively sensitive to temperature changes. ICEs mostly don't care because they don't have a battery to heat and cabin heat is free. EVs pay for cabin heat with range. EVs have a negatively proportional efficiency curve with respect to speed and load. That is, the faster you go the less efficient an EV is and the higher the load you are drawing from the battery the less efficient they are. ICEs have a positive proportional efficiency curve. That is, as you increase load to 80% throttle at the given RPM they get more thermally efficient. This makes the net efficiency loss due to high highway speeds less bad as a percentage of 'normal' consumption for ICEs than EVs. This latter also comes into play if you live somewhere very windy. If you sometimes drive into 80 KM/h head winds, that's relatively a better case for ICEs than EVs. Similarly, if you live somewhere hilly the consumption for driving uphill rises slower in an ICE than an EV. EVs regen on the way down which helps, but only if you have sufficient empty battery capacity on the way down the hill. If you consider all these and other more minor properties, like the fact that EVs are heavier and often need heavier tires as a result, then EV range is less predictable because it's so efficient already. If an EV is already 95% (of theoretical) efficient, then losing a straight 5% efficiency (95% -> 90%) because it's windy is a big deal. In a 60% (of theoretical) efficient ICE, losing a straight 5% efficiency (60% -> 55%) isn't so bad.