That's sort of what I expected. The article gave the impression that there was one panchromatic pixel for each 1 red+1 green+1 blue. i.e. 1/4 of all the pixels were panchromatic, and there are equal numbers of panchromatic as for any one colour. (sure makes it easy to form square grids out of sets of 4, not sets of 3).

The pictures you included show that 1/2 of all the pixels are panchromatic, and the others are shared evenly between RGB, so thera are 3 times as many panchromatic cells as any one colour (except that green is over-represented due to a combination of sensitivity ans spectrum issues) . This is about what I guessed, but it's nice to see the actual diagram confirming it.