If P, then Q.
P.
Therefore, Q.
If I eat nightshades (P), then I get migraines (Q).
I eat nightshades (P).
Therefore, I get migraines (Q).
But denying the antecedent, though inferentially invalid, makes even more sense to me, and conforms most to my experience:
If P, then Q.
Not P.
Therefore, not Q.
If I eat nightshades (P), then I get migraines (Q).
I do not eat nightshades (not P).
Therefore, I do not get migraines.)
This is false. You still may get migraines. For instance, you deny the antecedent, but a horse kicks you in the head. Bang, headache, logic is not applicable.
Modus tollens seems even worse to me:
If P, then Q.
Not Q.
Therefore, not P
If I eat nightshades (P), then I get migraines (Q).
I did not get a migraine (not Q).
Therefore, I did not eat nightshades (not P).
But Popper used modus tollens to reject the hypothetico-deductive model and get out of the inductivist dilemma, which utilizes the inferentially invalid “form” of affirming the consequent:
If P, then Q.
Q.
Therefore, P.
To my nightshade example, this is the worst choice to me because it’s invalid and contrary to my experience:
If I eat nightshades (P), then I get migraines (Q).
I get migraines (Q).
Therefore, I eat nightshades.
This is also false. You may get a migraine from other than eating nightshades.
To make your dillemma and your belly-aching stick, you need to reword the original condition:
IFF P, then Q.
If and only if you eat nightshades, then you get a migraine.
Had you started with this, your arguments and ensuing discussion would have made sense.