John, do you measure the vertical components of stress (or deformation) tensor? Do you use the absolute type of measurement? We use a reference frame associated with the plumb line, which is absolute stable. Our sensitivity is order of magnitude higher than tides as you can see here:

http://www.astro.cz/galerie/v/uzivatele/Pavel_Kalenda/Fig10_9_13C.gif.html
and here:
http://www.astro.cz/galerie/v/uzivatele/Pavel_Kalenda/13C_120817.gif.html

This are the anomalies before Kurily and/or Kamchatka EQs. Compare them. You can see the development of the tilt and especially the development of the noise, which is proportional to the deformation rate and/or stress rate. You must use our theory of the non-linear development of the stress before EQ and you must recognise the stress waves both in tilt (deformation) and noise (deformation rate) domains.

On the other hand, I recognised the same stress waves before Loma Prieta EQ based on the creep measurement in California. But, the monograph about this event stated, that there were no observable anomalies before this EQ. This results only from a wrong theory and model of EQ preparation. Not from the measurement itself. You must use other eyes.