Mike beat me to the M9 could only be subductions....and intensity vs magnitude.The difference between a M8 or 9 at the distance and a M5 nearby....which might produce the same intensity at a particular location...is that a M8 would last for maybe 2 minutes and a M9 might last for 5 minutes. I was also going to say that if you had a large part of the subduction zone fail in 2 or 3 separate M8+ quakes, you could not then have a M9 to re-break the same subduction zone until some centuries of strain accumulation has passed. There is some research my Ray Weldon et al on San Andreas fault that suggests that a fault can fail in a very different pattern.....but that may be a special case and I am unaware of M8+ quakes occuring on the same part of the same subduction zone in the same decade...the big question is, is it possible for a subduction earthquake to release only part of the accumulated stress/strain. When I posted that there was a zero chance of a M9 quake on the same part of the Sumatra subduction zone around the holidays...I did say "zero" was rounded to the nearest integer...the article in EOS about the West Andaman fault...which I read and posted on right after I made the first "zero" post, suggested that the subduction zone only failed as far downdip (NE) as its intersection with the West Andaman fault...so if the deeper part farther NE is seismogenic (slab in contact with upper plate..and locked), then that might still be able to fire off a subduction quake....but the seafloor motion would not be as dramatic so would not expect a big tsunami...
Chris