Obsidian has a valid point. The Gent has a valid point. The Curious Soapmaker has a point. Explaining the mathematical distinction between the three viewpoints will only be more confusing, so I'm not going to try.
In the end, any of these methods will work about the same, as long as you stick with a superfat that has a reasonable saponification value. There will be too much error when using superfatting with a fat that has an unusually low sap value, such as jojoba or lanolin. But you want cocoa butter, and that will work plenty fine.
The easiest and simplest is Obsidian's method. Calculate the recipe as a regular CP or HP recipe using the lye discount you want. Hold out the fat you want as superfat. Make the recipe with the rest of the fats and all of the lye as usual. At the end of the cook, the soap ~may~ possibly be a bit lye heavy. Small matter. Add in the superfat. If there is a bit of excess lye remaining, it will saponify happily with a bit of the superfat. The fat remaining in the soap will be the fat you want as superfat at the correct lye discount % you wanted.