The Lineartronic gearbox, although CVT and therefore without any obvious changes in gear because, well, there aren’t any, surprised me because I thought I could detect just such a gear change, which was weird. But it was smooth and quiet.

After a 45-mile drive, mainly cross country but with quite a few braking and accelerating spells in villages and overtaking buses and lorries, the Levorg’s trip recorder showed a fuel consumption of 29.1 mpg – not too impressive, given the company’s claims that include engineering for economy. The official combined consumption figure is also a touch depressing at 39.8 mpg, while the emissions are 164 gm/km, meaning an annual road tax outlay of £185. The insurance group is 24E.

I would expect the car to achieve better on longer journeys with motorways being the main routes, but, if economy is top of the wish-list, this is perhaps not the car for you.

Either way, I loved it! It was just so nice to drive. The ride and seats made the Levorg really comfortable. And the feel of all the controls, from the steering wheel, to the indicators, the wiper controls and the rest gave confidence that the car was well-constructed. The dashboard and central console looked smart and the leather-trimmed front seats were very comfortable. Everything else was upholstered with a classy-looking fabric.

The Levorg comes with loads of technology on board means owners get a good deal of in-built safety. The car as tested did not have Subaru’s latest ‘EyeSight’ system, but this becomes available in the spring of 2017.

I did get to test EyeSight out, however, in a Subaru Outback. EyeSight has two cameras, in stereo, as it were, like the human eye, located either side of the interior rear-view mirror. As in many other systems, usually using radar in their case, employed by other manufacturers, their purpose is to detect potential crash scenarios looming up and to prevent them by automatically braking the car if the driver and not reacted in time.

Subaru’s claim is that this works at up to 31 mph, apparently impressive compared with the more usual industry claim of around 19 mph. I tested the system twice in a controlled environment provided by Subaru. The car was driven towards a stationary object (a full-size representation on a plastic sheet of another Subaru). I drove at about twenty mph and, as the alarm and red warning lights triggered, my car started to brake before my natural reaction also to brake kicked in. The car stopped inches short.

Now do it without any braking from you, said my instructor in the passenger seat. This time, without any pathetic little last-second input from me, the result was the same: no damage. BUT, my driving colleague then had a go. He accelerated to more than 25 mph, perhaps 30 mph, and our car hit the ‘car’ in front. Therefore, yes, I suppose Subaru’s claim that the system works at up to 31 mph, but not if you are simply too close to the object in front for the laws of physics not to kick in.

EyeSight is, nonetheless, a device will at the very least reduce the severity of an impact. Another clever feature that can prevent an impact lessens the throttle opening should a driver of an automatic for example select drive instead of reverse, or the other way round, and if another car or object is detected to be in line for a bump. This can include bikes, passers-by and so on, as long as we’re talking about things at least about a metre in height.