SMIRF Engine: An interesting conclusion. I would tend to agree, especially if the Bishops start out pointing at the opposing King.
This raises some questions though. How can you decide if the advantage is too great? Are we talking about a SMIRF position evaluation of more than 0.3 for example? What about other positions, like where the Cardinal is next to a Bishop?

Beren the 32nd: Well, the advantage exceeded 1/2 pawn unit, but still sank, when calculating longer and deeper. Normally the advantage should be around 1/4 pawn unit, which seems quite normal and acceptable.

Because it is too risky to move Queen or Archbishop that early cross the whole board, it seems to be sufficient to separate the bishops. Moreover there have been only few such extreme starting arrays, so the added rule hopefully should be sufficient to limit possible starting advantages. Now there still are 12118 different starting arrays.

Thus the current SMIRF download also gives now the modified SMIRF beta with engine 1.59.

SMIRF Engine: Can you give an example for such a position, please?

And why should there be an additional advantage for W in a mirrored position?

Caissus: Exactly as Beren the 32nd already has stated, there have been some positions, where a pair of Bishops was targeting the opposite King and King's Pawn. Actually I have not stored the numbers, sorry.

But I did not understand your second question.

SMIRF Engine: Sorry, have to answer in German at the German board.

SMIRF Engine: I have been given a position with neighbouring Bishops as white in this game: http://brainking.com/en/ShowGame?g=1930943
Do we know if Fencer was planning to implement the new rule or not? I don't feel right about playing on with an unfair advantage, unless I'm allowed to count it as a slice of good luck. Other games on this site involve luck don't they? ;-)

Beren the 32nd:
In your mentioned case there is no significant advantage (SMIRF's opinion), it seems to be quite regular (at the starting position):

01:51.7 (11.01=) +0.176 1.Nh3 (d3) Nc6 2.Nc3 Nh6 3.d3 d6 4.e4 g6 5.Bxh6 Axh6 6.Bxj7 Rxj7 7.Cb3
00:55.1 (10.07=) +0.281 1.Nh3 (d3) Nc6 2.Nc3 Nh6 3.e4 e5 4.g3 d6 5.d3 g6 6.Be3 Bxh3 7.Axh3 Bxj2 8.Rxj2 Nd4
00:46.5 (10.03=) +0.281 1.Nh3 (Nc3) Nc6 2.Nc3 Nh6 3.e4 e5 4.g3 d6 5.d3 g6 6.Be3 Bxh3 7.Axh3 Bxj2 8.Rxj2 Nd4
00:42.7 (10.03+) +0.178 1.Nh3 (Nc3) Nc6 2.Nc3 Nh6 3.e4 e5 4.g3 d6 5.d3 g6 6.Be3 Bxh3 7.Axh3 Bxj2 8.Rxj2 Nd4
00:29.1 (10.01=) +0.176 1.Nc3 (e4) Nh6 2.Nh3 e5 3.e4 Nc6 4.g3 d6 5.d3 g6 6.Be3 Bxh3 7.Axh3 Bxj2 8.Rxj2 Nd4
00:07.7 (09.01=) +0.281 1.Nc3 (e4) Nh6 2.Nf3 Nc6 3.h3 e5 4.e3 f5 5.d3 Bf6 6.Nxe5 Nxe5 7.Axe5 Rxe5 8.Bxj7
00:02.6 (08.04=) +0.375 1.Nc3 (e4) Nh6 2.Nf3 Nc6 3.h3 e5 4.e3 Bf6 5.Nxe5 Nxe5 6.Axe5 Rxe5 7.Bxj7
00:02.4 (08.03=) +0.375 1.Nc3 (Nh3) Nh6 2.Nf3 Nc6 3.h3 e5 4.e3 Bf6 5.Nxe5 Nxe5 6.Axe5 Rxe5 7.Bxj7
00:00.6 (07.01=) +0.434 1.Nh3 (e3) Nc6 2.Nc3 Nh6 3.e3 e6 4.d4 Cb6 5.Bxj7 Rxj7 6.d5
00:00.2 (06.06=) +0.518 1.Nh3 (e3) Nc6 2.Nc3 Nh6 3.e3 e6 4.Bi6
00:00.1 (06.01=) +0.518 1.Nh3 (Nc3) Nc6 2.Nc3 Nf6 3.d3 e5 4.Bxi7+ Kxi7 5.Ne4
00:00.0 (05.01=) +0.469 1.Nc3 (e4) Nc6 2.Nf3 d6 3.h3 Bg4
00:00.0 (03.03=) +0.377 1.d4 (d3) Nc6 2.e3 Nxd4 3.exd4 Nh6

SMIRF Engine: It is interesting that you say there is no advantage in this case. I will not look too closely at your analysis, as only one move of the game has been played so far, and (again!) I don't want to have an unfair advantage, this time by knowing about SMIRF's understanding of the position.
This still leaves the question about the rule. Can Fencer or anybody tell me if it will be implemented?