To the best of my knowledge, the internal castings are different on a B18, they didn't have an extra 3mm of metal there just hanging around through a decade of production waiting for the B20.
I have heard of rare cases where the internal voids were slightly out of position (core shift) and it caused problems in a large overbore like this. Might want to have the wall thickness checked sonically before boring it, if they're not fairly even you might not want to do it.
John, As you mentioned, shifting of the water jacket cores does occur during the casting process in some blocks resulting in thin areas. I sonic tested a B18 block bored to B20 size with a core shift, and in the thinnest area of a bore the iron was only .050" thick, and it was .180" thick on the other side. This engine was run many miles but when carefully measured w/a dial bore gage, and it was out of round by close to .004" front to back in the bore where were the shifting had occurred, but not side to side where a cylinder bore normally wears.
Also checked four other B18 blocks and one of the four had a core shift. Neither of this block or the one mentioned above could have been bored successfully to B21 size and would be very thin in spots at a B20 bore size.
Blocks with a core shift will probably work OK when driven normally at a B20 size, but with the possibility of some blow by and a loss of some compression, but with a high-HP normally aspirated engine or with a blower or turbo it will balloon the thin areas on the power stroke and probably crack or split the cylinder wall in time as seen below.
Split cylinder wall.
Have bored a lot of cylinder blocks over the years and, you can actually hear the difference in sound when the cutter passes over a thin spot. In these areas, the cutter will actually force the thin areas of a cylinder wall outward as the cutter passes by resulting in an out of round bore - a rigid cylinder hone will also do the same thing and it is noticeable during the honing operation.
Another alternative is to offset bore the affected cylinders if the head gasket and clearance between the piston wrist pin boss and the rod will allow for this, or one or both can be trimmed by a small amount to get the required clearance. If this last approach is taken all of the pistons or rods need to be balanced. If both are trimmed everything needs to be balanced.
This image below shows a cylinder wall that was to thin and cracked and started leaking shortly after it was first run. Worst case scenario if large enough chunk of the wall breaks out it might hydraulic and ruin everything.
So the best thing to do before boring a B18 to a B20, is a JohnMc mentioned earlier, is to have the block sonic tested. If it does not pass the test find another block and try again. High performance and racing engine shops should be able to check a block for you for under $100.