You could avoid triple 20 entirely, you can go for 8 triple 19s and a triple 15, for example.
Can you point to data supporting that claim?
I would think a simple model for throwing is:
- get the left-right direction right
- get the throwing angle/speed combo right
and do not rule out that some players are better at one of those than at the other, showing smaller variation. If so, those players would be better at hitting the (almost) vertical triples than the (almost) horizontal ones or vice versa.
I guess I would want to know the point-spread function (https://en.wikipedia.org/wiki/Point_spread_function) per player.
If you know that, you could devise an optimal strategy, weighing the number of points each target is worth (even if you can hit triple-10 with 100% precision, it most of the time isn’t a good idea to aim for it) and situation of the game (you’re maximizing the probability of winning, not the winning margin, so if you’re far, far ahead, that triple-10 may become enticing)
I personally don't have any data to support this, but it seems extremely intuitive and so I think the burden should be on you to show 19 is equally easy to hit if for some reason you think it is.
20 also has the advantage of getting you closer to 0 if you don't shoot a perfect game, which is pretty much always, so 20 is BETTER unless 19 is significantly easier to hit.
Note that perfection does not necessitate hitting a high percentage of triple 20s, since one is also constrained by having to throw a total of 501 points.
One way to achieve true perfection is to hit triple 20 + triple 19 + double 50 with every 3 darts, giving all 9 darters.