The answer to question one is that different images contain different numbers of colour pixels in different block. So a picture of a bird flying across a blue sky can be compressed more than a picture of a bird against a background of trees etc, as it requires less information to record all the blue pixels than if the pixels are different colours.
The second question is a bit strange, as I've repeated your problem with a picture on my computer with the following results. I opened a 5.4mb Jpeg file resized it and saved as a 1.57mb file. Reopened that and save again and it's now 1.58mb. Reopened the 1.58mb file and save again and it shows as 1.58mb file. It might be that the estimate for file size isn't accurate in Windows. The normal wisdom is that the more you open and resave a Jpeg file the more information that is lost. So, if you were to do this twenty times the smaller the file would become, as recompression discards more informatiom. If you did the same with a Raw or Tiff file then it should remain the same as no information is discarded.
The third question regarding the value selected by you in the slider decides the amount of compression applied to the Jpeg. So the lower the value or the more you move the slider to the left the greater the compression applied to the file. The biggest problem with this is you have no control over what information is lost. In normal operations this shouldn't make much differences to the finished image, except it will be smaller in dimensions. Hope that helps.
