Put this way, to my judgement a skilled driver should be much safer than an equally skilled motorcycle rider.
In the end though, you're right, and we don't even need anecdotes to prove it - the stats are voluminous and clear. Yes, you can reduce your risk somewhat with good bike and skills maintenance, but even with that, motorcycling is much more risky to life and limb than driving a car. If safety is higher in your valuation than whatever your reasons for riding (there are many possible), riding is never rational. Safety is fairly low in my evaluative framework, and my reasons for riding persuasive, so it's not a hard choice in my personal case.
Humans are quite fragile in relation to the things we do.
Not sure whether a graph of skill vs safety would show increasing or decreasing difference though.
On your specific claim, I don't know about US statistics, but when I looked in Aus (about 5 years ago), the proportion of deaths involving drunk/inexperience/unlicensed riders was much greater than for drivers. Individual choices and skill had a much bigger effect for motorcyclists.
Nonetheless I'm not remotely claiming that death rates are not greatly higher than for drivers. They absolutely are, in all categories. But the headline multiplier may not apply to any given individual.