Please note that 538's final forecast (as of the morning of Election Day) had Trump winning KY by about 19 points, yet he actually won by 30 points.
If trends from the recent past continue, it would suggest that Trump would win KY in 2020 by a margin in the mid-20s.
It's not that simple though. Parties and candidates have ceilings, the national environment (which the FiveThirtyEight average obviously overestimated for the Democrats that year) has an impact, and the specific candidates running make a difference. So I wouldn't be shocked if, say, a Kamala Harris who lost the popular vote nationwide managed to underperform Hillary, let alone Obama, let alone Bill Clinton in Kentucky. But I would also not be shocked if a Joe Biden (who was the one polled here) who won the popular vote by 5 or more points managed to significantly outperform Hillary and more or less match Obama's 2008 margin here. It's all relative, and "trends" in and of themselves are not the end-all, be-all answer to anything. Kentucky has not shifted so far right since 2008 that a margin similar to McCain's would be unthinkable for any candidate under any situation; Texas, on the other hand, has shifted so far to the left since then that a margin similar to McCain's there would be pretty much unthinkable. The demographic and voter changes driving these trends, after all, obviously have to be taken into consideration.
