Don's comments on M15 did not make much sense as written. I think he just meant that they might be using a different scale. As for larger than M9.5...the Chile quake of 1960 was about a Mw 9.5...and its rupture lnegth was about 1000 km. The Dec 26 2004 quake was a M9.2 or 9.3 and its rupture length was longer than the Chile quake....but the average slip was (much?) less, and the width of the fault that broke was less also (I think). I looked at a map of the Sumatran subduction zone and it looks like there is about 1000 km of it south of the equator until you get south of Sunda Straight...and then it continues farther...south of and beneath Java...OK, pulled out my Atlas...there is 10 deg of longitude of this south of Java....over 1000 more km...and the subduction zone continues even farther. OK...let's say that you have an earthquake that breaks 2000 km of the subduction zone...from the equator south and east to about Bali. This may be a totally unrealistic earthquake, but I can't say impossible. Let me guess that the average downdip width is 200 km. There is a simple equation that relates fault area to magnitude. It would be better to add slip...I don't know if Wells and Coppersmith or some other paper includes slip....does someone (Roger or Don?) want to calculate the average slip that would be needed over a 2000x200 km rupture to make a M10.0 earthquake?
I heard that the Chile quake had slip as large as 100 m....that seems like too much...I wonder if it was feet?

Please don't take this post too seriously: I do NOT think that there is going to be a M10 quake on this subduction zone ever, and last spring posted my best guess of odds for a M8+ on the subduction zone for different periods of timke....emphasis on "guess".
Chris