Yes, "no lines out of point 73" - you can't have a white dot that isn't on the grid, and doesn't have a line, or colour edge, leading it back to the grid. You should be able to locate the points in the error message via a display somewhere in your editor which shows the XY coordinates of your cursor on the image, ie 1770, 2318 and 2343, 2184 and see what is being complained about.
I often put unconnected dots on the grid, when I'm doing heightmapping, when a layer on my stack is ASTER height data, converted to BoB height map format, to capture inflection points in the height - ridgetops and valley bottoms - as they pass across the grid. It is a quick and easy form of heightmapping. But if I need to capture a hilltop with a white dot, in the middle of the tile, then I must connect the dot to some existing border or other, which leads back to the grid, via a blue fake line.
You have to think of mtt2 as a robot that crawls along the grid, and only takes excursions into the tile when it encounters a line or colour edge which crosses the grid with a white dot. Then it follows along the line, and every line that breaks out along it, cataloguing the locations and heights of all the dots it finds along the way, til it hits the grid again. At that point it goes back to the first point on the grid past the one it was following before it began the excursion, and continues along the grid til it hits the next crossing line, then repeats. When it has gone round the whole tile, it scans the tile for any white dots it hasn't encountered, and puts them up as errors. This is the point where it can find a series of white dots linked together but not connected to the grid, and it will commence going around the series looking for an exit, indefinitely, and you have to ctrl-C to kill it.