Ilford recommends 13 8x10 prints per liter of processing chemistry. With 2 liters in each tank of the CAP 40, that works out to 26 prints. I've found that you can go to 32 prints for test prints before color crossover becomes too much of an issue.
That's why I said you can make 22-24 prints (8x10 or equivalent number of square inches of material) before you have to change to making test prints for the next run. The density and exposure times are still accurate at prints 26 - 32, but the color balance in dark areas gets too red.
Using Ilford's recommendations or my experience - you can make 5-6 16x20's in one 2 liter chemistry batch. If you've done your test prints from 26 - 32, then you'd start the next session by making 2-3 test prints to reconfirm color balance of the 16x20's to be made during the printing session and then go directly to printing the display 16x20's for the next 5-6 prints.