Iannello and Godfrey - McMurdo
Another way to look at the same question is to postulate that the Cocos were used as a destination in Shah's simulation, and the aircraft simply continued past the Cocos and ran out of fuel. This short note takes a quick look at how that scenario plays out.
The track, great circle, from the 19:40 range ring to the Cocos is shown below (green line). The track was extended past the Cocos maintaining the great circle trajectory,
The small "star" North and West of the "4" range ring marker denotes the Cocos.
For simplicity a fixed ground speed of 480knots and a fixed track of 169 degrees was used to create the path. The details associated with this choice are shown in the spreadsheet below. A late FMT was used in this path as was used by Iannello and Godfrey.
As is my custom, I stopped the calculations at 00:11. The 00:11 location is 27.1S and 101.1E. The Iannello and Godfrey 00:11 location using McMurdo station as a destination is 26.08S 100.36E. I would characterize the BFO errors associated with the Cocos destination as comparable to the errors obtained by Iannello and Godfrey using McMurdo as a destination.
It should be noted that if the great circle is extended to ~45S the corresponding latitude is ~105E.
Update: 24 November 2016
A recent publication by Mike Chillit suggests a terminus near the Batavia Seamount based on flight path and drifter considerations. That paper can be found at the link below.
Mike Chillit - Batavia Seamount
A screen capture of a graphic from the linked paper shows the suggested terminus below.
The Batavia Seamount location (25.75S 100.33E) is added to my flightpath derived above in the graphic below. The Batavia Seamount is approximately 80nm North of my derived terminus.
A close-up of the above with the Iannello and Godrey McMurdo based terminus (pin labeled I&G) is shown below.
Drifter summary from referenced Chillit post.