removing current from mast & coax with a choke and mast isolation forces more return current to flow into whatever is connected to the coax braid above the isolation point as per Kirchhoff's current law
You have a misunderstanding of Kirchoff's Current Law as it applies to the feedpoint of antennas. You seem to think that the feed point is a single node and the two antenna elements are the connections to said node. However, the feedpoint of even the simplest of antennas does not have one node, but in fact two.
Take a center fed dipole that is fed with a balanced feed line. The current from one leg of the balanced feed line goes through one node and out to that element of the antenna. The current from the other leg goes through the other node, and into the other element of the antenna. In this case balance is achieved, but only because the source itself was in balance with regards to current to begin with. As you can see, if the source is in perfect balance, and there is nothing present such as a matching system, your attempt at implementing Kirchoff's Current Law appears to work. However...
Now lets change the balanced feed line to an unbalanced feed line, such as coax. Now we have a center conductor with has a given amount of current flowing on it, and the inside of the shield, which has some other amount of current flowing on it. This is possible because the feed line is not balanced, hence the term unbalanced feed line. In this case we have the current that is flowing on the center conductor that passes through one node and out onto one of the antenna elements. We also have the other current that was traveling on the inside of the coax shielding that travels through the other node to the other antenna element. As these currents were not equal to start with the currents flowing on the antenna elements will also be different, even though the antenna itself is a balanced design. This is why using a balun on such an antenna setup provides a benefit.
Now to show another example of where your implementation of Kirchoff's Current Law goes wrong, lets look at a 5/8 wavelength antenna matching system. This is the schematic of an Imax matching section.
With one exception (the one on the far left), every other black dot in this schematic is a node. Each node will have its own voltage and current at any given time. The voltage and current present at any node will be based not only on the source voltage/current from the feed line, but said effects of the neighboring components as well. You cannot treat this entire circuit as a single node, which is what you seem to be doing. You have to treat each point in the circuit as a separate node or the math that goes along with Kirchoff's Current Law (and Kirchoff's Voltage Law as well) simply doesn't work.
NOTE: (If this note seems confusing feel free to ignore it) The exception would be the wire that directly connects two black dots. This link has no components, so both sides of this link would be the same node. They are essentially the same point electrically. However, there is a caveat here as we are dealing with an RF signal, which is AC, not DC. Because of this, if there is a distance between the two sides of this connection, the two sides will be slightly out of phase with each other even if the difference is only slight. As such, depending on the design, this distance, should it exist, would cause the two points to be treated as separate nodes.
This is just one possible matching system, there are others that would be even more complex. You cannot simply ignore the feed line and its possible imballance as a current source and, should one exist, the components in a matching system, and assume that an antenna works as a single node for Kirchoff's Current Law. That is simply not how antenna feed points work. Unfortunately, every time I see someone refer to Kirchoff's Current Law, they make this very mistake.
The DB