Some simulator data should prove helpful here to make things less abstract:
Simple Big Money
Let's take a simple Young Witch Big Money bot which buys 2 YW and money. Let it play against a bot that opens YW/Herbalist (one of the worst banes). The YW+Herbalist beats YW-no bane 58-40.
Now replace Herbalist with Wishing Well (top of the bane list) and the YW+bane beats YW-no bane 80-17 (it buys up to 4 Wishing Wells before Silvers).
Big Money + important $5
Now, if there's a Wharf we really want:
YW+Herbalist (into Wharf) is on par with YW-no bane (into Wharf).
YW+Wishing Well (into Wharf) beats YW-no bane (into Wharf) 72-26
So at least for Big Money games it's probably a safe bet to always open YW/bane instead of YW/Silver.
To expand on this, I think this is still a little problematic to follow blindly, but...
Doing the same thing with YW-money, courtyard bane, vs standard courtyard BM, it's more or less tied, with a miniscule edge to courtyard BM. However, both of these things should change their play a good bit. The YW bot needs to buy CY at some point, and CY dude should be buying more CY, sooner.
Still don't try to use embargo as a bane.
If there are important, non-terminal 5s, things get hairier. Lets look at Lab, which is pretty good with the filtering of YW. Here, you want the bane before your silvers. And you want a couple banes before your labs. Even if they're really bad banes (again, not embargo).
Oh, and by the way, if the opponent doesn't buy the bane at all.... the YW bot beats every single pre-programmed single card strategy I can think to test it against. All the other cursers, jack, ambassador*(I should point out that ambassador is played a little poorly, and this buys only one), masquerade (if it only gets 1 masq), you name it. It even beats the witch bot on a 5/2 start