Additionally, what of partial "answers?" I suggested Illness in the Ranks above. A resolved Illness in the Ranks turns a Monastery Mentor into a 3-mana 2/2 with Prowess. This, in my opinion, should be enough to beat the deck. They have spent, at minimum 2 colorless and a colored. I have spent one colored. If, with my two mana virtual advantage, I cannot at least begin to overcome the Mentor player, I'm curious to ask, what better can I do? Does Monastery Mentor need to be answered for less than one mana? (Also, I've provided at least one Mentor answer for <1 mana above)
I think it's possible that the difference between the way others see the problem and the way you do is that most don't think Illness in the Ranks (or Dread of Night) actually answers Monastery Mentor. Illness in the Ranks or Dread of Night both answer Monastery Mentor's tokens - things that the Mentor deck gets for free. If you spend a card and a coloured mana to deal with the things the opponent got for free you are certainly still quite far behind. Sure, you could spend and additional mana and an additional card to kill the Monastery Mentor, but in that situation you have effectively traded down because the opponent got a 2 for 1.
An answer like Supreme Verdict means that one trades down on mana (spending 4 to deal with a 3 mana threat whilst still trading 1 for 1). Virtue's Ruin is significantly better, but it can be countered and is pretty awful against the rest of the format.
With respect to considering a good definition, I think something to the effect of "a favourable trade can be thought of as expending fewer resources answering a threat than the opponent invested in the threat" would be a good starting point.
In this example, the opponent didn't spend any resources on the Monastery Mentor tokens, which you have now spent one mana and one card answering.