Todd, There was a quake in the 1890s along the northern NAF SE of Istanbul...maybe Princess Islands and SE of there. In 1912 there was a major quake onshore west of Marmara Sea (Ganos quake) that the French sea floor data suggest extends relatively far east into Marmara Sea.

But, yes, sounds like the fault in the rest of northern Marmara Sea has not broken since 18th century, CAUTION: I have not read the papers on earthquake history: I may get some of this wrong.

See my response to Canie, above. The seafloor location of what many consider the main active strand of the NAF is now known. But, its dip is important and we, after additional processing, image its dip in its upper 1 or 2 km.

Imagine that the Northern NAF is located 10 km off of western Istanbul coast (it is 10 km south of the coast), in the vulnerable area near the airport (bad site conditions, bad construction (perhaps on building less than 5 stories?) (another caution: I've heard these things second or third hand: I am not an engineer).

If the fault dips north, the deep part of the fault would be closer to being beneath the coastline than if it dips south...try 70 deg, 10 km depth, draw a triangle, and do the trig (then post it so I know also). Or, the fault could be vertical. This last is part of the last part of my draft SCEC abstract...which will likely be deleted because the abstract is 2x too long.
Chris