Hyperaccuracy in NYT clueing

One of the reasons the NYT puzzles are getting too easy is the over-accuracy of their clueing.

Saturday's puzzle was by Mike Nothnagel and Byron Walden.  I happen to know that Walden is a mathematician (at Santa Clara, I think).  So when I read the 15-letter clue

Rectangular array that's identical when its rows and columns are transposed, as this puzzle's grid

I wrote in SYMMETRICMATRIX, which is about the only thing it could be, at least mathily.

For the 15-letter answer at 5 down       

They may be seen on a lake's surface

I considered RIPPLEDWAVELETS and REFLECTINGWAVES before settling on REFLECTEDIMAGES, the correct answer, but I didn't actually ink any of that in.  The first reinking I had was clued

Believer in al-Hakim as the embodiment of God

For which I wrote in DRUZE after getting most of the crossings (it's supposed to be spelled DRUSE, I guess).

One down,

Hindu musician's source material for improvisation

got reinked from POEMS to SAGAS and then to RAGAS, but I don't know what at RAGA is.  (And no, I don't intend to look it up, because I never look up things that come out of crossword puzzles.  And why is that?  Because I once asserted to my father that to look up a crossword puzzle clue on the web is to "Reduce crossword puzzle solving to a clerical activity," and I intend to stay true to that sentiment in the spirt of the leading my life according to meaningless rules, personal stricture, and pointless self-denial.)

I've got the puzzle here, but I'm too lazy to post a photo.