The ghosted double line U? watermark is quite quizzical especially since it never appeared on the 1 cent Washington stamp of this design, not to mention the apparent head-to-tail pairing of two letters. I would attribute it to some sort of contact contamination from its environment over the past century and move on to the other facets.
Yes, I agree, this is the real watermark, an upside down "P" with a weak leg extending upward.
With regard to "plate position 43": Scott's diagram of a pane showing watermark 190 is only representative to show how approximately 90 letters can spread randomly across 100 definitive-sized stamps and shift from position to position. A stamp may have a portion of the USPS watermark resembling "Scott diagram position xx", but this would not have any correlation to an actual plate position. The paper could also be fed upside down, reversed, etc., as your stamp is down compared to the Scott diagram.
Then it gets down to whether you have a genuine coil of something trimmed down. You state it is flat plate, which would point toward Scott 441, rather than the rotary 448, as the thread developed, unless I missed something in your data.