Age of consent and good/bad sexual interactions

Age Of Consent (AOC) exists in order to prevent bad sexual interactions. It is usually explained differently (e.g., religiously protect the children, power differences, risk for trauma, etc.), but these reasons don't work; I will ignore religious accusations of blasphemy, during a sexual interaction both parties may actually enjoy power difference, and risk for trauma is a general concern for which we have developed preventive and remediative methods which are widely available for application not just by a trained professional. So really AOC exists in order to prevent bad sexual interactions. In terms of interactions bad just means that at least one of the two people of the interaction thinks the interaction is bad. This is more general and also opens up the possibility of good interactions, which the usual explanation of AOC skips.

Imagine a situation without AOC, some number of good and bad sexual interactions happen per unit of time. If we introduce AOC, the idea is that it intends to block all of the interactions regardless of their valence.

So after introducing AOC there are some good interactions which would have occurred but have been blocked by AOC, some bad interactions which would have occurred but have been blocked by AOC, some good interactions which occurred despite AOC, and some bad interactions which occurred despite AOC.

Obviously good interactions being blocked is bad, bad interactions being blocked is good. If these numbers were weighed equally (humans have negativity bias so there would have to be more good than bad interactions blocked in order for these two sides to be weighed equally) AOC would easily be debatable. If (many) more good than bad interactions are blocked then AOC is removable.

Normals don't consider this possibility because they think there are no good interactions in this context, and that any positive evaluation of illegal sex by an underage person cannot be true. They only see that it is good that bad interactions are being blocked so they don't see what could possibly be bad about AOC.

Normals are not insensitive to good interactions being blocked, it just doesn't occur to them that this can happen, so stories may help to persuade them that it's possible for AOC to also do harm by preventing good sex interactions. An easy example is the 23 year old who falls in love with a 15 year old but does not act on it due to fear of prosecution and ostracism, thus missing out on what could become a relationship.

In terms of measuring AOC effectivity, perhaps there are pre-post numbers available which show that bad interactions went down after introducing AOC. But then there are also many good interactions which are turned bad due to the AOC's criminalizing effect regardless of whether the younger one wanted and enjoyed it, so the real test would be to see the ratios between the numbers of bad interactions blocked, and good interactions blocked/criminalized.

The crucial point here is that AOC decreases the number of bad interactions, but it also decreases the good interactions and criminalizes ALL good interactions that it doesn't prevent. The fourth category, the bad interactions that it doesn't prevent, are still bad and it doesn't remedy that in order to balance out the baddening effect the AOC has on good interactions (by criminalizing them).

So imagine there are 10 good and 10 bad interactions "scheduled". AOC comes in and blocks 5 bad interactions and 5 good ones. however there are still 5 bad ones left and the 5 good ones that are left have now been criminalized. The final situation is 5 bad interactions and 5 criminalized interactions, with the potential of more than 5 bad interactions if some of the criminalized interactions turn bad due to the difficulty of dealing with the criminalization.

Have I just gone too far in trying to logic this? Can anyone help me spot my error or have I actually made a point?