As posted last January, I wore my Observatoire continuously for a month after it arrived, winding it morning and night as suggested in the paperwork. Total deviation over 30 days was +9 seconds -- pretty good! Something on the order of a third of a second per day.
The Besancon certificaiton of the watch showed a considerably broader deviation than that, so it's not "official" -- and while I like the idea that the Peseaux movement was capable of tenths of a second per week, I understand that Kari's rendition of the movement is not intended to achieve super accuracy. That said, I do like (and expect) my watches to achieve a good standard of accuracy, and the Observatoire has served me well in this regard.
My overall "champion" is still my JLC Geographique, though -- the timing slip that accompanied it from its most recent service had no number greater than a 1 on it, and it seems to keep virtually perfect time whether on the wrist or on the winder...
Best,
Gary G