I think the whole "comfort" thing is overblown.
I do think some players could be affected by being jerked around wildly, but minor movements or lefty-righty slot swaps are not as upsetting as many seem to feel it is.
One thing that makes constructing an "optimal line-up" is deciding what data or combination of data to use. Do you use strict overall numbers (career, last year, last 2-3 years?), or lefty righty splits, or batter vs specific pitcher numbers (with significant sample sizes involved, which is rare), or numerous other factors including hot or cold streaks. It's easy to conclude that every person views which of these is more important differently.
I, for one, probably view lefty-righty splits as more important than JF and many posters. I'm not trying to claim I know more than JF or that I am right and he is wrong. I realize I could be wrong on my views of player comfort with more consistent batting slots, and that "comfort" might be more influential than lefty-right numbers or other factors.
My own philosophy is that the lefty-righty-lefty line-up construction is over-used. I'd try to put the highest OBP guys up 1-2-3 with the guy up 3rd having more power than 1-2, but I'd rely on lefty-righty splits about equally with overall numbers (last 2-3 years mostly but not in a vacuum). So with my philosophy oversimplified, I might mostly take the average between lefty-righty splits and overall numbers over the last two years to guide my line-up creation methodology.
Here's a look at this over simplified data;
2015-2016 average between lefty-righty split and overall numbers with an adjustment made to HanRam & Pablo due to their outlier 2015 season:
vs RHPs
Betts .357/.510/.867
J B J .347/.501/.848
Pedey .367/.447/.814
Ramirez (.329/.457/.786) adjusted to (.345/.475/.820)
Pablo .302/.340/.642 (adujsted to .335/.440/.775)
Bogey .347/.428/.775
Moreland.316/.458/.764
Young .306/.428/.734
Holt .336/.387/.723
Leon .329/.384/.713
Vazq .264/.293/.557
vs LHPs:
Young .367/.525/.892
Ramirez .343/.518/.861 (adjusted to .350/.530/.880)
Betts .342/.503/.845
Bogey .383/.454/.837
Pedey .375/.443/.818
Moreland .309/.440/.749
Leon .331/.409/.740
Vazq .324/.384/.708
Holt .347/.356/.703
Pablo .266/.346/.612 (adjusted to .275/.355/.630)
Beni's sample sizes are too small, but I'll project:
vs RHPs: .360/.460/.820
vs LHPs: .335/.435/.770
So, going by just these numbers, the "optimal line-up" guide might look like this:
vs RHP:
1) Beni .370 OBP
2) Pedey .367 OBP
3) Betts .867 OPS
4) JBJ .848 OPS
5) Ramirez .820 OPS
6) Bogey .775 OPS
7) Sandoval .775 OPS
8) Moreland .764 OPS
9) S Leon .713 OPS
vs LHPs
1) Bogey .383 OBP
2) Pedey .375 OBP
3) Young .892 OPS
4) Ramirez .880 OPS
5) Betts .845 OPS
6) JBJ .781 OPS
7) Beni .770 OPS
8) Leon .740 OPS
9) Holt .703 OPS or Rutledge
Is this "optimal"?