position 43 II - cd4 fg5 gh4 gf4

This is also lost in Brazilian version. In here idf64.org/tables-of-draw/ you have correct table, first one. To see why position is lost you can look and the 14th chapter of this study:

lidraughts.org/study/EIVFmLf5

We have used the FMJD tables as they are also found on results.fmjd.org/viewpage.php?page_id=2. The page takes a long time to load, so be careful.
Last time I checked it showed the same list of positions for Brazilian and Russian, so I assumed the positions would be fine for both variants.

I've taken a look with the Aurora Borealis engine, and it confirms that these openings are indeed clearly losing (or very bad) under Brazilian rules: 41 XIII , 25 XIII, 42 XIV, 43 II

They will be taken out of the opening table for Brazilian.

For 1 IX and 4 XV the move sequence is invalid under Brazilian rules. But the resulting starting position is fine. So I would say we might as well leave those in.

Hi Roep! FMJD is very stupid, they wrongly copied IDF tables and add some mistakes :)

after FMJD done it, wrong randoms were corrected, but FMJD didn't do anything about it

In addition to what had been written above, I propose to examine a possibility for choosing different sets of ballot openings in lidraughts tournaments. Many ballot openings are hard for those players who are rated below 1500, and I believe that a set of 150 positions which is available at idf64.org/tables-of-draw/ , would suit them better.

The following openings have now been taken out of the table for Brazilian:
5-IX, 19-V, 25-XIII, 41-XIII, 42-XIV and 43-II

I like this idea @kalnap! It will be some work to get all positions entered digitally (as FEN), but they are much smaller tables.

That IDF table with 150 positions has been added today @kalnap. You can now create tournaments with random positions from "IDF Drawing Table 64 - Basic Positions", find them in the board editor, etc

1