I think the best answer to this whole debate is that there is no right or wrong answer as to whether color-based pictograms or orientation-based kanji are better for Shogi variants.
However, I would never EVER recommend using orientation-based pictograms, unless they either:
all have some universal defining feature that makes it easier to tell each side apart, like the wedge shape with the kanji Shogi sets.
are mnemonic representations of the piece moves and are also color-based, like with the mnemonic Shogi variant sets.
I think the best answer to this whole debate is that there is no right or wrong answer as to whether color-based pictograms or orientation-based kanji are better for Shogi variants.
However, I would never EVER recommend using orientation-based pictograms, unless they either: