So currently we have our FMA speeds based on two calcs, one where Bradley blitzes several soldiers and one where Ed dodges a bullet from point blank range .
Unfortunately, the calculations are both inaccurate and need to be corrected.
The calculation where Ed dodges a bullet from point blank range, uses a rifle bullet (~800 m/s) to get projectile velocity. It was actually a revolver. Specifically, it was a snubnose revolver, meaning generally lower power rounds and a loss of velocity. I'd guess it was a .38 special as that's pretty much the quintiseential snubby. That means a highball velocity of 350 m/s (it would actually be slower because revolver cartridges are designed for slower deflagration which means they require longer barrels).
According to the blog, Ed's head is 0.093m from the revolver and he moves 0.2625m. 0.093m / 350 m/s = 0.000266 s.
0.2625m / 0.000266 s = 986.84 m/s or Mach 2.877 (Supersonic+).
Next, King Bradley's feat is up. The math here is actually correct; the issue is the values used. The low end calc is only subsonic, but we use the high end which is based off the idea none of soldiers were able to see Bradley move. To this end, the original calc used 1/220 seconds (the value where fighter pilots could perceive an image in an experiment) as the perception time.
However, using this value is incorrect, and makes for an rather large highball. The issue is that the value the calc uses is absolute perception speed in perfect darkroom conditions (perception speed is highly dependent on contrast and lighting) with subjects who genetically and training-wise have the best vision in the world.
In contrast, several other studies have shown that perception speed in decent lighting for the average young adult is at best around 13 ms (~1/75th of a second).
According to the blog, Bradley moves around 25.7 meters in total.
25.7 m / 0.013 = 1976 m/s or Mach 5.76 (Hypersonic).