Im halfway guessing that the figure 8 has to be turned sideways for this.
Correct.
The cardioid pattern fills in the null from the side of the figure-8 mic. The figure-8 mic is in phase for stuff coming from one side and out of phase for stuff coming from the other side.
In decoding it, you add the side (figure-8) to the mid (cardioid) for one channel and subtract it for the other channel.
Stuff coming in from straight in front of the pair comes in on the cardioid (mid) mic and doesn't significantly appear in the signal from the side mic at all. It ends up in the center because L=M+S and R=M-S, so both have a signal from the cardioid mid mic.
For sound coming in from the positive side of the figure-8, the left side gets M+S, so the volume of the sound is increased on the left. The right side gets M-S, so the sound is decreased on the right.
For sound coming in from the negative side of the figure-8, the left side gets M+S, and since the sound, as reproduced by the mid mic, has negative (inverted) polarity, the sound is diminished (adding a negative is like subtracting a positive). Likewise, because the right side gets M-S, the sound on the right side is increased (subtracting a negative is like adding a positive).
So it fundamentally depends on the sound from one side being out of phase with the sound from the other side, which only occurs on a figure-8 capsule (and, I think, on a hypercardioid capsule, but not evenly enough to be useful).
Thumbs up on
the VP88, BTW. That's a sweet mic.