Well then (looking REAL close at the pic I guess I can see you have it mounted horizontally, just cropped the pic).
Yes, this is DEFINITELY a problem with vertical linearity. Most better analog control monitors and many digital control monitors have a dial/menu option for "V.LIN" to control this. Basically the idea is to show a cross-hatch pattern, and then adjust the V.LIN so that each square is EXACTLY the same height from top to bottom of the screen.
Here you've got expansion in the first 10% of the screen and compression in around 20-30% of the screen, thus the crooked diagonal line.
If purely caused by magnetism, I'd also notice vertical lines being bent sideways, so it seems to be a signal problem to the yoke itself as part of the vertical drive, not of side interference.
However, if there is a signal generating noise at the precise frequency nearby, it could make it's way into the components and be reflected by the yoke. (Remember: A yoke is simply windings of copper wires and when electricity is supplied to it, a magnetic field is generated "bending" the electron beam where it needs it. Clearly the horizontal is working perfectly, but the vertical is the problem)
Being that your game produced a good image but the JPac didn't, it seems the monitor is A-OK, but rather the signal or noise nearby is the culprit.
I know a fair amount about monitors but am far from anything beyond causal tinker-er. You'll need some input from some good techies. Post this query to the Sci.Electronics.Repair usenet newsgroup or do some google newsgroup searching for vertical linearity problems with either JPac, arcade monitors, your particular monitor #, etc..
Good luck!