Mujannah vs Mujannah (table I)

Move #14 (#26) of the secondary line is a wrong input from me, it was supposed to be 14. Axb5! (not 14. Nxb5?): I couldn't do it right due to modern Chess rules.

If Red/White really plays 26. Nxb5? then al-Lajlaj suggestes 26... Nxd3+ 27. K~ Rxb5 with better game for Black. (I excluded this option from the diagram to avoid further confusion)

Very cool, thanks for sharing!

A few things to notice which perhaps are not clear watching the diagram:

- al-Lajlaj analysed what happens after 13. h3 a6 to which White/Red replies with 14. a3, however it is not mandatory, W/R can also ignore Black's movement on the Farzin-side and go ahead with the 'normal plan' Rb2--Rbg2 on the Shah-side (or Rb2--Rf2 as well);

- although so, we should consider that a Black's pawn established in b4 could be quite annoying for White/Red;

- check carefully, many variations are possible around moves #7 and #8 (actually #19 and #20 counting the early 12 moves), personally I am lost there;

- we assume that White had to play 13. Nd5 (not 13. Ad3?) and finally he still gets a better game (see next diagram).

25. Nd5 Rd7 26. Rxc6 Rxd5 +=

I'm glad to receive a thanksgiving from Coolthing!

There are still a lot of openings and variations analysed by al-Lajlaj, so there's a lot of work to do! Not sure I will complete it, but I'll try to...

chesscom also provides analyses like "great move" and "mistake" for variants?

I'd love!!, but actually it is not possible: there are no excellent programs for Chess variants, this applies also to Shatranj/Chaturanga. Plus in order to have a good program you need some good programmers inputting the game theory which humans have developed along the centuries! Unfortunately a great part of Shatranj theory was lost, only a few Arabian manuscripts survived, and we are the guys studying them, not others! So, no much hope...

However the absence of an excellent program is somehow good to avoid cheating.

I'm sure that with the help of the AI something can be done but I guess it would require to invest Big money into that.

If your question were who marked these moves as 'good' or 'bad'? Well it's me! I just copied what found in "A History of Chess" by Murray, which is the translation of what written by al-Lajlaj 1000 years ago.

And yet another variation:

This one perhaps is less accurate: there is too often the phrase "any move" which is not easy to fix... Still interesting in the end. Again many sub variations at move #7 (which is move #19 of the game).

Well, my first impression was that al-Lajlaj didn't do a good job with this second variation. Black moving the Rook back to h8 doesn't looks great to me! Perhaps al-Lajlaj wanted to analyse what happens when White doubles their Rooks on g-file and Black doesn't reply very accurately. This would explain also why there are a couple of "any move"s on the Black part and it also gave me ideas for a possible title (I'll edit the previous post).

Perhaps when I have more time (and money), I will attempt to make Leela Shatranj Zero (based off of Leela Chess Zero). In the meantime Fairy Stockfish is superhuman, and I am also developing a Stockfish-like engine to surpass it.

Though, modern engines (Fairy Stockfish, my engine) assume the ferzes are on e1/e8
it is very easy to tell when one of these programs is playing - they have very unique tabiya

Ah wow great! I forgot we have programmers inside the team!!

If I had the ability to make a Shatranj program I would make it working with both conventions: d1/d8 and e1/e8, then add a "mirror position" button... Just an idea.

Questions: is Fairy Stockfish free? Or somehow downloadable for free? I would be curious to try it (when I repair my PC or buy a new one). I'm sure it is a GREAT tactician, of course, but I doubt it can find good openings. Btw I would be really happy to be wrong! Well, and can it recognise stalemate and 'baremate' as winning??

Debatable ending in this analysis because there is a small typo and a sort of inconsistency in the second to last move.

24... Ad6! cannot be so, it's just a gift to White, however not even 24... Ae6 seems to be the best option for Black. So by my choice I marked 24... Kxg8 as interesting move ("!?").

Apart from that, the other moves seem sound.

Another one, rich of variations:

Mujannah vs Mujannah (4)

Rf2 v Rf7 'classic' Rook formation on both sides

Phew! Many variations in this one!

In the main line after 22. Rfg2 Black plays 22...Nf6 or 22...Nf5; other defences will be analysed in the next two diagrams.

There's a bit of inconsistence in the last Black's move, not sure why... I guess Black's g-pawn doesn't capture the Alfil in h3 to avoid White entering with Rg7.

You could download it from https://fairy-stockfish.github.io/ or try online at https://fairyground.vercel.app/

Even without nnue, it can play a decent enough tabiya (just as my non-nnue engine can). Of course any NNUE engine can find good tabiya - but my personal belief is that the Fairy-Stockfish NNUE tabiya is difficult to play as a human.
My engine and Fairy-Stockfish can both play reasonably with the vertical mirror, but both engines will suffer more from this if using NNUE.
Both my engine and Fairy Stockfish should be fully aware of and compliant with the rules (unlike Fairy-Max or Pulsar)

Yes. Your work is enlightening and interesting, thanks for sharing.

Mujannah v Mujannah (5 & 6)

After both players completed their tabyiat (in 12 moves) the next 9 moves are just the same as for variation #4 main line.

In other words #5 and #6 are just sub variations of #4, therefore I'll start from move 22 which is move 10 of the previous diagram (21 early moves aren't shown here).