I am trying to understand the explanation for how the LINEST function
regresses that is given he http://support.microsoft.com/kb/828533

On the QR Decomposition example worksheet there are a couple of steps that I
don't understand how to go from A to B that I would like to know if there is
any further explanation anywhere that I cannot find.

First is the section where the X premultiplied by P matrix is re-written
with the comment Rewrite: effectively 0 -- 0: What is happening here?

Finally at the conclusion of the example the regression coefficients are
backcalculated from the matrix as 1.25, 0 and 0. Yet the results of the
LINEST for this example give different coefficients (-3.111, 0 and 1.222).
Why are those numbers different.

Full disclosure - I am not a mathmetician by any stretch, just a ChemE who
skipped most of my math classes at Uni, so explanations that don't involve me
looking up big long words in big heavy books would especially be appreciated.