The original sin of electrophysiology is putting metal into aqueous solution. Didn't your mother ever tell you not to do that?
Therefore generation upon generation of electrophysiologist is cursed to endure the vagaries of these recordings. In particular the properties of metals submerged in saline are poorly understood. One of the few properties we can actually measure is impedance, and so we attribute mystical significance to it.
Impedance is a complex quantity consisting of a real, resistive component (like a resistor), and an imaginary, reactive component (like a capacitor). The impedance of your recording system is dominated by one bottleneck: the metal-electrolyte junction, that is, the uninsulated tip of the electrode in saline. Nothing in the entire circuit matters at all except for the tiny two-dimensional surface at the tip.
You might ask: what about the ground wire? Shouldn't that matter too? The answer is: no, not if you've set it up correctly. A nice ground electrode will have a large, uninsulated surface area and a low impedance, essentially a short circuit. Perfect. For intracellular recordings, which will necessarily have a nonzero DC offset roughly equal to the resting potential, you will also need your electrode to pass direct current. That is why Ag-AgCl wires are commonly used for that application: presumably the chloride ions can freely move between the solution and the wire's coating, without limit, in either direction. For extracellular recording, the DC offset is not relevant, and so you can get away with a much simpler ground, commonly a stainless steel screw. (Stainless means non-oxidizing, which would make it a particularly poor choice for passing DC current.)
The primary requirement of a good signal electrode tip is small exposed area. The electrode will spatially average the voltage over its tip surface, and the spatial extent of neuronal signals is tiny (perhaps tens of microns)