Final variables and downcasting

[del boy]

Hello,
I do not understand why when we try to assign a FINAL variable of type primitive data type (eg. int ) to a lower type (eg. byte)
the compiler WILL NOT throw an error, only IF the value of the final variable is within range.... if you look at case 2 below if the variable
g is NOT final will have compiler error: POSSIBLE LOSS OF PRECISION, same with case 3 when final var is larger than data range of byte.
MY QUESTION IS: WHY with final variable in the first case we DO NOT have to explicitly downcast ???
Many thanks in advance!

   //CASE 1 -  int is FINAL & within range off byte: Will compile OK
final int g = 10;
byte c = g;

//CASE 2 -  int is NOT final and within range of byte : Compiler error: POSSIBLE LOSS OF PRECISION
int g = 10;
byte c = g;

//CASE 3 -  int is final and out of range of byte : Compiler error: POSSIBLE LOSS OF PRECISION
final int g = 1024; 
   byte c = g;  

[Thomas Paul]

Because the compiler knows that the final variable can never be changed to something else so it will never be beyond the range.