Good morning everyone. I mapped that frequency graph; let me try to explain.
The graph shows ONLY the frequency response (given a flat input) of the "tri-comp" network you can easily see on the schematic. It's the final stage of the Rusty, post gain & EQ stages. The purpose is to achieve reduced intermodulation and a more even response, by overdriving narrower bandwidths (multi-band distortion/compression); all three bands share the same clipping threshold or "ceiling", and since there's no post-EQ the treble ends up dominating the sound unlike most other designs.
What is NOT shown is the frequency-shaping going into it from the input-gain and EQ stages. If you map out the input-gain stage (I don't have that anymore), it's something like what a Rat would do with increased treble pre-emphasis as the control is turned up, but I remember the high-pass frequency being even higher than a Rat, something like 3KHz (a Rat is 1.5KHz; a Tube Screamer like 750 Hz). Then, going into a fairly conventional PASSIVE eq stage, you end up with a fairly deep mid-scoop around 400 Hz, I believe.
For what it's worth, I suspect the mid-band of the tri-comp "makes up" for this mid-scoop, reducing and narrowing it a bit by providing extra mid content as it overlaps with low and high bands as you can see in the graph. The EQ controls, being passive, have pretty broad responses and a lot of interaction compared to what discrete active stages would provide.
Long story short, because the EQ is before the the tri-comp clipping, the EQ essentially plays a role in determining which frequencies get clipped. It is hard with the Rusty to get a high level clean bass relative to the overdriven highs because like I said they share the same "ceiling", and so the output is always a bit treble-heavy...aggression!