Ya, very good point. Spend a lot more attention to wire placement if you’re gonna install a heater element on top of your battery. You cant just disregard the general best practices and think only along one individual spec without taking responsibility for ALL the consequences, especially the unaccounted for ones.
One of the reasons i try to put my worst ideas up here, i probably forgot something 
Right. I appreciate the chart and data. I have some other sources, I’ll post in that thread.
I got to admit it’s a bit amazing. We have been making battery packs for years. Yet there is a research and knowledge gap. Something we as humans do not yet have a sufficient understanding of, to describe mathematically via calculations, how much current can you send through a stirp before it overheats. So we have to resort to experimental data.
Essentially what I’m saying is that the formulas for convection and conduction are dependent on the geometry of the object, length, thickness, and width etc… It seems to be the greatest misconception perhaps ever, at least from these equations that the length does not matter. The length plays a major role in heat transfer.
For convection, the “h” term is dependent on the geometry. Formulas exist for certain geometries, not all geometries have known formulas. But we have known formulas for vertical and horizontal plates (could be used for strips), as well as cylinders (could be used for wires).
The thickness of these geometries has very little impact on heat transfer, perhaps another great misconception. The two “Wt” vertical plates has essentially zero heat transfer via convection, and there is essentially zero heat transfer via convection within a closed box. The “Lt” vertical plates has a non-insignificant role in heat transfer via convection. The surface area which has the greatest heat transfer is the horizontal plate “WL”, as this is the largest surface area, and will have the greatest heat transfer via convection.
For conduction heat transfer, from the cells to the strip, to the strip to insulation (admittedly not exactly zero, but almost), insulation to box wall, the thickness plays zero role in heat transfer via conduction.
Based on these concepts “stacking” one strip onto another does not double the cooling capacity and therefore the current.
Based off of these concepts, a wire (approximately a cylinder in shape), would have zero heat transfer via conduction between two flat hard objects, if you consider it to be “almost” a perfect circular cross sectional area. There is very little surface area in contact with the box wall, or the cells, but only a tangential surface area.
Another ampacity table often referenced is Matador’s ampacity table. In my opinion, this table should not be used. If you look into how he calculated the numbers, he took a table from NEC 310 for insulated copper wire, and found a linear best fit between resistance and one over the current. Then extrapolated this for all conductors. There is several problems with this. The difference in geometries, and that ampacity for insulated copper wire is the temperature at which the insulation melts. Something different than what we are looking for, for strips.
