I use the Y and M combinations for the same reasons you state on my LPL, and I'm able to get a good grade 4, going beyond that pretty much requires just magenta filtration.
I've not tried Ralph's method, but when I first got the enlarger I made a set of comparison prints using the full range of MG filters, and another using the combination filtering from Ilford's data sheet.
In general, I found the color head filtration was about 1/2 grade softer than the MG filters.
Since grades are kind of arbitrary with VC papers anyway I didn't go any further in trying to calibrate, I just add or reduce contrast as needed, and I've rarely had a need to go beyond what amounts to grade 3.5.
All that said, I'd try some tests with just magenta filtration and see what happens. Getting up to 3.5 with the combination filtering should be fine for most work, and your tests will be a good basis for venturing beyond that when you need to.