Hello i'm a brand new member and really new to this whole thing so maybe this is a dumb point but here goes.
Mick, I think you missed one of the key points of Chris' video in your rebuttal. His point of reference for the location of the object is not the intersection between the camera views, but instead a trajectory of the object being recorded that places it at a location on each line at each time point. This has a few key assumptions, mainly that the object is going at a constant velocity without turning. That's why his representations for location of the object are places where the distance between each line of bearing is equal, not the intersections. Focusing on the intersection between point 1 and 2 in your simulation makes no sense to me. How could an object be at the intersection of line of bearings 1 and 2 but then Jump to the line of bearing of point 3? it makes no sense. Each line of bearing represents some location where the object must be at that point in time. Assuming it's moving (a decent assumption given its elevation) It takes 10 seconds to get from a point on line 1 to line 2 to line 3.
The error in estimating the position using the intersection of the bearings from times t and t+delta are small compared to the other errors that exist in the model, and don't change the results significantly. The intersection is best viewed as an estimate of the position of the object at time t+delta/2 - one line will be off by some amount, the other line will be off by some other amount the other way. Better schemes can be introduced when/if there's sufficiently reduced uncertainty in the inputs to warrant it, and right now there isn't. Pinning down the rate of turn is way more important, as it has far more significant numerical repercussions.