I finally did observe (3 times!) the frozen-MyHD-app-on-FF-jump bug lately. It was during heavy FF/REW action on two MyHD-captured H2M-edited (and H2M-nullstripped) Olympics files. I had the video card set up to clone its output to S-Video, so I was able to observe the desktop in the frozen state and confirm that the PC was completely unresponsive to mouse and keyboard, requiring a press of the reset button to recover.
I forget who-all have reported seeing this before, but this experience brings to mind a question for you that I didn't ask before: Have you ever seen this bug when playing a CBR recording? That is, a MyHD recording that's either unedited or edited with all null packets retained?
Also, on VBR files where the bug occurred, was the play-time indicator wildly inaccurate, showing a "current time" many times longer than the full program length?
I suspect that this bug does not occur except with VBR files and that the degree of inaccuracy of the play-time marks on the OSD and control panel is also related to the propensity for this bug to jump up.