I have been having the same issues with the same processor. 16F883, PBP 2.50b, and the USB programmer on ver 4.21.

What I discovered over the last 36hrs:
I make a code change in the program, compile and then write it to the processor. Then upon testing the change I find it get the same results as if I made no code change. VERY frustrating.

After a while (5 days) I thought I was losing my mind, I decided to put an old program into the chip to see what was happening. And I discovered the update did not get into the chip even though it verified the program after writing.

I went to a program that was months old(just the hex file no re-compile), that I knew did not work, and pushed it out to the 883. To my amazement it went to the chip, then I was able to push my latest code (with re-compile) out and it too worked.

What the issues seems to be so far, after calling MELabs and ordering an new programmer which did not help. They said their hardware was probably not the issue. And they were correct.

Guess#1:
The default installation path for PBP is very long in characters and you get close to the 62 character max for compiling if you give your revisions any detail at all, so you the path described in PBP needs to be verified. If you created a mapped drive to the default MCS directory that might be the issue?

I have no idea if that is the issue but now I push done the dummy hex file every few re-programs if I have any doubt if the programmer or PicBasic Pro is passing the hex file correctly to the programmer.

My new programmer came with the updated 4.24, so I updated my old programmer and put the new one on the shelf.

Somehow the new hex file is not getting to the programmer software or to the directory and the programmer is pushing down the previous code, which is why I am getting a good verify.

Kevin