You’re correct. As I mentioned, toxin is biologically expensive. It’s also important to note that if that random mutation for slight toxicity never happened, both the deer and the plants likely would’ve carried on as before and both would have gotten along fine. Or if some other trait had happened like spiky leaves, the deer might have evolved tougher lips instead. Or if the deer and plants had split off into two groups, one with toxin+immunity vs tough lips+spikes, they could both coexist in the same ecosystem since neither is competing with the other. Eventually they stop breeding with each other and become different species.
There’s also random/disaster selection, where a trait may not be beneficial at all or even detrimental, but some random event happens. Like say half the deer are nocturnal, while half aren’t. A flash flood wipes out the valley during the day while the nocturnal ones are up in the hills sleeping, leaving only nocturnal deer. And also sexual selection, where a trait has no purpose except to attract a mate (see: peacock).
If we’re getting into K vs r, I like to mention insects, which tend to have an absolutely HUGE r rate. They breed by the thousands, and have tiny lifespans. This is why there are around 950,000 species of insect and only 30,000 species of fish, even though fish came first. Also why you get so many freaking weird looking bugs.
Edit: A note on fish, while fish usually have a large number of offspring at once, only a few make it to breeding age, and they tend to have much longer lives than insects do.