...It would have run out of fuel 3000 feet ago, and is coasting.
You may actually have something there, but not much. IR seekers go for the heat, and with no countermeasures happening, it will lock on to an engine, and go there. There have been a couple of MANPAD attacks on Commercial jets in Afghanastan, and both hit the engine, not the fuselage, and those jets were well within range of the missile.
Sorry but the Afghanistan shoot downs were at much lower altitudes and speeds... as were the 26 total civilian aircraft that have been brought down by "stinger" missiles. They were taken out shortly after takeoff or on approach for landing. The missiles were fired close enough to lock onto a single heat source and not be "confused" by a multiple target. The Tier One Stinger Missiles that hit their engines were much less sophisticated than a Tier Two MANPAD which have greater range and longer engine ignition times.
The resolution of the IR sensing array is too course to differentiate the location of any one engine at the closing speed and the missile's simplistic computer will target the center of the averaged IR image. This is particularly true of an "All-Aspect" targeting missile. This means that the missile will lock-on to the center of the overall IR signature of the target, not a specific engine.
The Mistral and several of the Stingers are "All-Aspect" and will lock on to the signature. If the missiles fuel has been expended and it is "coasting" then all steering is aerodynamic and the small fins of the missile do not allow gross or quick maneuvers. By the time the Mach 2.5 missile is close enough to differentiate the four heat sources, it is far too late to lock on to any particular one and maneuver for an accurate hit on an engine. The evidence actually points the possible missile actually hitting the aircraft IN FRONT of the CWT.