Sorry to say so, but it does not work like this at all. The idea behind the type of exposure metering that you are discussing is to place the midtone into the mid-point of the useful input range of the photosensitive material. 18% grey is midtone, but the card is flat. To get from a flat card readings that are representative for a 3D real-life object an exposure meter should apply correction. But it does not know what type of the surface it is. The difference in reflectiveness of two equally grey surfaces, one being flat, and the other being a sphere, is about 0.5 EV. 18% divided by square root of 2 (1 EV is a 2x change in light, so 0.5 EV is sqrt(2) ) is 12.7% grey. For a flat grey card to work as it should for exposure metering with a non-flat receptor it must be at least be placed at an angle to the lens axis, not perpendicular.