Argument over the shifting of the burden of proof often seems to rest upon whether this is a fallacy at all, and not upon what sort and degree of proof ought to actually suffice to engage such a shift. But this is not the only application of this argument. In a civil action in an American law court, for example, the person doing the suing must simply show that it is more likely than not that some claimed wrong was committed against them. There, the inquiry is not whether something is possible or impossible, but of which side's story is the more probable, a burden borne by the person employing the reigns of the court going into the trial by suing at all. But if that person provides enough proof to make their claim the more likely, this, then, shifts the burden of proof, and the other side may still prevail (and, now, may only prevail) by providing even stronger proof negating the side of the one who sued, and instead favoring their own side of the story. But this level of inquiry ought not to be restricted to the sphere of law.

Indeed, most any argument may come to revolve on a back-and-forth proof-shifting display, each side making points and counterpoints designed to make their result seem the more likely, each argument countered by additional proof from the other side designed to shift things back. Now, consider this example, Cynical Cindy and Gullible Gil are watching a horse race. The leading horses are the green-jersey horse and the gold-jersey horse. They are neck and neck in the last lap, when the green-jersey horse suddenly pulls ahead and wins. Now the argument commences, why'd the green-jersey horse win?
Gil asserts: "The green-jersey horse must have had an extra set of legs which gave it that extra speed burst."
Cindy retorts: "What? We didn't see any extra set of legs."
Gil replies: "It was an invisible extra set of legs."
Cindy huffs: "There is no such thing as an invisible extra set of legs."
Gil shrugs: "Prove it."
Now then, knowing nothing more about the world, or about history and biology and physics, we might credit Gil's explanation. We, as well, might shrug and tell Cindy, "Well he's got you there," and chalk it up to a 50-50 possibility, at least, of one or the other being right; or even giving the argument to Gil because something made the green-jersey horse win, and no other explanation has been offered. Here, then, Cindy has two possible routes of contention; one being insistence that the proposition raised by Gil is impossible, the other being demonstration that there exist other alternative and more probable explanations which render Gil's proposal unnecessary and comparatively improbable. Suppose Cindy begins with the first tack:
Cindy: "Every horse historically recorded to have been observed to have been in this or any race has been observed to have, at most, two sets of legs."
Gil: "How about Sleipner?"
Cindy: "Sleep-what?"
Gil: "Sleipner. Odin's horse. It's all in the Eddas, Loki turned into a mare and rutted with SvaĆ°ilfari, and the next day Loki gave birth from his womb to Sleipner, an eight-legged horse-- which could also walk on water and on air!! And" (Gil continues, folding his arms triumphantly) "it is, in fact, recorded that Odin sometimes journeyed among mere mortals, disguised as an old man, so it follows, obviously, that if he traveled on Sleipner, those extra sets of legs would have to have been invisible, as well."
Naturally, now, a somewhat agape Cindy realises this argument is threatening to devolve into an unwinnable contest over the historic validity of Norse mythology. She knows that no matter how absurd she finds it, it is impossible to affirmatively and absolutely disprove Gil's inherent claim that Norse mythology is historic fact.

This spiral into absurdity may be avoided by the distillation of what is actually at issue here, which is the assertion of the necessity of additional capacities. Every additional feature or ability can be claimed as a capacity (for even a feature presupposes the ability for the generation and sustenance of that feature). If we claim an additional capacity to be impossible, yes, we do bear the burden of proving that; but if we merely claim it to be unnecessary, unsupported by proofs for which this capacity would offer the only explanation or even the best explanation, then it falls to its proponent to prove the capacity to be necessary, and to prove it as the best explanation.

A claimed capacity may deviate so far from what may be observed that it may lack any foundation in comprehensible reality; it might be internally inconsistent, such could not actually coexist with other asserted capacities; and most damning of all it might offer an incomplete explanation, while a more complete explanation for the proofs offered is one exclusive of the capacity claimed. Denial that a thing exists differs from denial of the necessity that the thing exists, or, even, of the likelihood that the thing exists. One need not prove the impossibility of a thing to prove it so improbable that it ought to be treated as fully accounted for by a more probable explanation, unless evidence is provided disproving the more probable explanation. And, in the same stroke, if more likely explanations may be discounted, then the claimed capacity is only as likely of an explanation as all other equally unlikely explanations.

And so, Cindy's response ought to be a shrug and an alternate positive explanation --
Maybe the green-jersey horse was on steroids; or maybe the gold-jersey jockey was paid to throw the race and slowed down so much that it only looked like the green-jersey horse sped up in comparison; or maybe the green-jersey horse had a better diet, or a better training regimen, or had been holding back up to that point to build excitement. Any one of these is a possible explanation; affirmatively disprove all of those, and all other more mundane possibilities, and your explanation becomes a possibility worth considering -- but even then not a necessity -- because if it is possible that the horse had an invisible extra set of legs, it is as possible that it instead had an invisible set of wings, or an invisible rocket-booster, or an invisible man running behind and pushing.
And that properly shifts the burden of proving probability and necessity to the person claiming an invisible extra set of legs to be the only or best explanation.