Those are formulaic estimations, not measurements--formulas are incredibly inaccurate. For instance, plugging in 22.5 liters of 1.090 with the hop schedule of the second recipe and an 11 liter partial boil gives me 193 IBUs if I choose to use Rager for the estimate, but just 88(!!) IBUs if I use Tinseth.
That's not a small difference, and shows that you really can't trust formulas to get you near the correct IBU value.
Even when you're in "normal" IBU ranges formulaic estimates are lucky to get within 20-30% of the actual IBU value--Brad Sturgeon and Paul Childress did a great talk on this for BBR back in July of 2007. Once you get outside the range of normalcy, they tend to be much farther off.
Basically, you should pick one formula and stay with it so that you have some sort of known guide to use when tinkering with a recipe. That's quite valuable. You just shouldn't assume that the number it spits out is very closely related to the actual number of IBUs in your beer.