focus load seems to focus heavily on #kms rather than time & power

The calculation of excercise load seems to focus heavily on #kms ridden, rather than time spend and average power. 
My workout on sunday 24 Nov was outside, riding flat, with an average power of 164Watt over 2.5h. 
My workout on Wednesday 27 Nov was on Zwift, climbing a virtual hill, with an average power of 145Watt for 2h to consciously work in low aeroob load. 

The load calculation for Sunday was 305 whereas Wednesday was 80 (x3.85). This seems heavily biased by the kms  (70.9km on sunday, 22.8 virtual kms on Wednesday ==> x3.1). This seems out of balance with time spend (x1.25) or average power (x1.13)

In zwift I can ride a big gear and have a small power in ERG mode, leading to an out of ballance high average speed, or I can ride a downhill course. so virtual kms seems to me like the worst parameter to take into account to calculate load.

what am I missing?