This is a true dichotomy. If the bank is not expected to prove the sigs match, then the only other option is to burden the account holder to prove that the sigs do not match. You are the first person to claim that in a dispute the account holder has the burden of proving that their signature does not match the transaction signature (which they don't have a record of). You are also an unreliable source. This claim is so profoundly absurd, it will need to come with substantial convincing evidence to sway anyone with a reasonable idea of how the US legal system operates.

So now I'm looking at claims from two unreliable sources, but your claim is the one that is highly inconsistent with US legal culture. The legal system in the US does not, as a rule, burden the defense to supply proof. Although there are some instances where the defense must submit proof, your claim is just too far fetched because the scenario does not have the characteristics of other cases where the defense must prove a negative. Unless you can miraculously come up with some very ironic solid evidence, you're not convincing anyone with this.

Difficulty of forgery is irrelevant because the premise assumes that a forgery happened. If you don't have a forgery, there is no case to speak of. There is no standard of evidence to compare. You're off in the woods, while the discussion you're replying to is comparing the legal protection behind victims of PIN forging to victims of sig forging. If one of the perps had to work harder to pull this off, it doesn't change the courts standard of evidence.

You're using circular logic. Your premise is broken because it assumes the cost of fraud by way of pin forgery is higher for the bank -- which would actually only be true if the bank failed to demonstrate that the pin entered matches the account holders pin. Yet it's actually trivially easy for the bank to prove that the pins match, which means costs you're assuming the banks absorb are incorrectly assigned.