This is somewhat typical of what I would see with locked timing ( timing stays with in a couple degrees throughout the power curve). There are always exceptions.
The difference in power between peak torque and peak horsepower is very little to nothing. Where I see the shift when doing timing sweeps is below peak torque and above peak horsepower at the ends of the curves Of course this is somewhat dependant on the motor combo and amount of ignition retard.
This is easier to read on these graphs because the ignition only retards a couple degrees through the power curve so you can pretty much read the graph directly. That's not always possible.
If you optimized this curve and compared it to locking at 30 you would pick up 1.6 tq at the bottom edge 3500 rpm and zero at the top end 7200. Timing was swept from 28 to 34 degrees. 34 lost on both ends and 28 made the same as 30 on the bottom and lost on the top. Again this doesn't represent every motor but I don't recall ever seeing anything that was 20 on each end. That doesn't mean it can't happen.
The link is timing measured on the engine from about 1500 (idle) to little over 7000. The power curve starts at 3500 and goes to 7200.
If it matters this is a race engine not a street engine. It's also a chevy and I'm sure that does matter. lol
Do you have the graphs where you saw 20 and 20 that you can share? It would be interesting to see what they look like.