A vehicle always drifts to the side with the least caster.
Don’t know what’s done with our modern cars but in older cars I thought caster was usually set positive 1/4 degree more on the passenger side than the drivers side to cater for road crown. The larger the caster difference, the higher the force that pull the car to the side with less caster.
Making the passenger side 0.25 degree more will compensate for a crowned road but it’s a balancing act as one tries to equal the left drift induced by the road crown itself with a right drift induced by the lesser driver side caster. You could get a car to steer straight on some crowned roads but then you’d have unbalanced forces on flat freeways that result in a pull to the right
So cars used to be aligned for the roads one travelled on, using string, older alignment machines and operator knowledge and such old school methods. And this was at a time when power steering was a luxury so caster was very positive to induce steering self centring forces.
But newer cars have wider and lower aspect ratios tyres and power steering which must impact the caster induced forces somewhat…
It’s all interesting subject matter but in this case, as mentioned, it’s likely an issue with the electric power assistance and the lack of calibrating the electronic system to recognise the actual physical steering centre position after the alignment. Doubt the EPS system is capable of determining whether the calibrated centre and the actual physical centre position are in agreement. Such is probably why the EPS system pulls the steering to the calibrated centre which obviously causes the vehicle to steer (because of the calibration error)… Either that or one of the new front tyres are faulty…