Thinking through the whole process of a discharge led me to suspect that these 2 properties of WST added together may have contributed to its unsuitability for the 9mm Luger cartridge:

1. load density: WST occupies a lot of physical space. In order to load it enough for the speed I want, the case is pretty close to full. This means the load density is high and we all know that high load density produces higher pressure. On a Pressure/Time curve, I believe it pushes the peak higher; i.e. the curve peaks higher on the Y-axis. Other fast powders tends to be loaded with low load-density; e.g. Titegroup (TG) or Bullseye (BE). I do not need to fill the case up with BE or TG to get the speed I want. So, the peaks of their Pressure/Time curve is lower.
2. burnt-rate: The burnt-rate of WST is considered fast. This means its Pressure/Time curve is shifted to the left on X-axis, compared to slower powder; that is to say the pressure spike is swift and it take a lot shorter time to reach the peak. Now, if you compare this property to powders that fill the case up for 9mm, they are not considered fast powder. So their Pressure/Time curve is not shifted so much to the left along the X-axis.

I think [1] is what compounded the discharge characteristics of WST in 9mm Luger cases. I bet the powders that are published with data for the 9mm Luger do not possess these 2 properties together; i.e. they either have one of these properties or the other but not both.