Most computer hardware and software does math in binary, not decimal. In
binary, most terminating decimal fractions are non-terminating binary
fractions that can only be approximated in binary. When you have to
approximate the inputs, it should be no surprise if the output is also only
approximate.
Almost all binary computer math (including Excel) follows the IEEE standard
for double precision, which gives decimal accuracy of about 15 significant
figures. As a result, decimal digits beyond the 15th usually not what you
would expect. Excel will not display more than 15 decimal digits (it will
pad the display with meaningless zeros if you ask for more), but you can
often get a peak behind the curtain when you cancel leading significant
figures by subtraction.
The two calculations you described use binary values that have the following
decimal equivalents
1.145000000000000017763568394002504646778106689453125
-1.100000000000000088817841970012523233890533447265625
-----------------------------------------------------
0.0449999999999999289457264239899814128875732421875
0.1449999999999999900079927783735911361873149871826171875
-0.1000000000000000055511151231257827021181583404541015625
---------------------------------------------------------
0.044999999999999984456877655247808434069156646728515625
If you do the math, you will see that it is exactly correct, given the
unavoidable initial approximatisons.
If you note where the 15th significant figure is on each of the two answers,
you will see why they display differently.
Jerry