The Immortal Mansuba of Abu Na'im al Khadim: a modern chess study and Murray's proof game
Some mansubat [shatranj studies] are like chess problems. If no ferz [Q], fil [B], the bare king rule or a pawn double step, are involved, then probably the rest is like modern chess.
Solve this one, it' s a mate in 3.
It's no53 in Murray's History of Chess , p. 284. In p. 309 writes:
"'This happened to Abū’n-Na‘ām, and he used to boast of it,’ AH. From both al-‘Adlī and aṣ-Ṣūlī, H. As already stated on p. 274, I accept the evidence of H as to the origin of this and the two preceding problems in preference to that of AH.
The claim that this position represents the termination of an actual game must be qualified. The position has clearly been edited to satisfy the artistic canons of Muslim chess, for Red’s attack is so strong that he must have had a mating attack the preceding move. The problem supplies a good example of this embellishment. It has been a favourite, both in West and East."
[MSS: AH61 / C12 / F56 / Man.62 / BM10=194 / H16=55 / V55 / Alf.58=63=64=97]
The Murray's saying about the fact if this mansuba was relied on an actually played game, concerned me as a proof game!
This endgame pattern has appeared [as seen already in Murray's writings] in many versions, since arabic manuscripts. I've tried and managed to find three MSS, referred in Murray. Specifically these:
- BM = MS. British Museum, Arab. Add 7515 (Rich). 655/1257. "There is nothing in the MS. to show the name of its author, but he has made liberal use of al-‘Adlī’s work, and quotes from al-Lajlāj with approval. Aṣ-Ṣūlī is on the whole ignored", by Murray, [in Qatar Digital Library].
- H = MS. John Rylands Library, Manchester, Arab. 59. 817/1414. "Nuzhat al-arbāb al-‘aqūl fī’sh-shaṭranj al-manqūl (‘The delight of the intelligent, a description of chess’), by Abū Zakarīyā Yaḥyā b. Ibrāhīm al-Hakīm. Al-Ḥakīm’s work is based upon the works of al-‘Adlī and aṣ-Ṣūlī, and carefully discriminates between the problem material which was taken from each of these lost works." by Murray, [in Manchester University Digital Library].
- Man = MS. John Rylands Library, Manchester, Arab. 93. 850/1446. Kitāb ’anmūdhaj al-qitāl fī la‘b ash-shaṭranj (‘Book of the examples of warfare in the game of chess’), by Shihābaddīn Abū’l-‘Abbās Aḥmad b. Yaḥyā b. Abī Ḥajala at-Tilimsānī alḤ-anbalī (B. 725/1325, D. of the plague, 776/1375) [in Manchester University Digital Library].
The Murray's diagrammed in book version, the above puzzle, [no53, p. 284] was found only in Man62.
Returning to this shatranj proof game, I must say that checked all possible variations but took in account also the fact that this game would have been played by aliyat [shatranj masters]. And looking at this version, Murray was probably right.
What could be black's last move?
If the d7 rook, was moved last and wasn't there, then black could have played instead 1... Rh8#. If the h7 rook was missing, then 1... Rd8+ 2. Re8 Rxe8#, as h7 square would be garded by the other black rook. If N wasn't there then 1... Rdg8+ 2. Kf8 Rhh8#. The pawn is irrelevant, and so K remains. K can only have been moved by force from g5 square after white's f4+. And before this from g4 after f3+ or from h5 after Ng3+. But before these black doesn't face a possible forced move and so Rh8# (mate in 1) is possible!
Maybe better to check from last move to first...
But was this version the actually played??
Firstly some brief images of shatranj history
From Shahnama (Book of Kings) by Firdawsi, 1650, Persian MS 909 in Manchester University Library
Abu Naim al Khadim [written also Abu'n Na'am] was one of the first recorded aliyat [shatranj masters] of the 9th century [in wiki], and surely had played the game against Rabrab and al-Kufi, before caliph al Ma'mun (reign 813-833) [in A World of Chess, J-L Cazaux, R. Knolton, p. 17 and Murray p. 231].
Al-Adli was the greatest aliya during the rule of al-Wathiq (227/842) and al-Mutawakkil (232/847) [Murray p. 198]. He wrote Kitab ash-shatranj (Book of the chess), a work that was reproduced in later MSS.
"It was under the following caliph, al-Muktafī (289/902–295/908), that the historian Abū-Bakr Muḥammad b. Yaḥyā aṣ-Ṣūlī first came into note as a chess-player of consummate skill." Murray, p. 199. As-Suli wrote 2 works under the title Kitab ash-shatranj, that were reproduced in later MSS. He also names Jabir, Rabrab, Abu'n-Na'am, al-'Adli and ar-Razi as having been of the first class of shatranj players (aliyat). [Murray p. 231].
The other versions of Na'am mansuba in Arabic MSS
Checking other versions of the above mentioned mansuba in arabic MSS, I've found three [one found in 2 MSS], mentioned also by Murray.
#1 BM10(=H16). Found also in Alfonso's Libro de los juegos, problem #64. From al-Adli's work. Maybe the most popular version. FEN= " 5K2/7r/4k1rn/4p3/4P3/5N2/8/3R1R2 w - - 0 1 "
Here there is a possible previous black's move. That is Rxg6, where on g6 could be a pawn or a farzin, threatening the h7 rook.
#2 BM194. Probably from al-Adli's work. FEN= " 5K2/7r/4knr1/4p3/4P3/5N2/8/2BR1R2 w - - 0 1 ". Here Murray would be right. No possible previous black's move found.
_
#3 H55. From as-Suli's work. FEN= " 5BK1/r6r/5kn1/5p2/5P2/6N1/8/4R1R1 w - - 0 1 " [in FEN the white Bishop is a farzin]. Here also there's a possible previous white's move. Rhxh7 where on h7 could be anything, and especially a N. While the existence of farzin makes things really difficult for black.
_
So what was the possible actual position of this Na'im's shatranj game?
Taking in account firstly that al-Adli played just few years after Na'im. Secondly the Murray's quote [p. 309] on this mansuba: "I accept the evidence of H as to the origin of this and the two preceding problems in preference to that of AH." And finally that H MS has two possible versions of this mansuba, both with a possible previous black's move, one of al-Adli and one of as-Suli....
....I vote for the version of H16=BM10, same as #1.
_
This of course comes with a lot of doubt, as I've checked only the diagrams of arabic MSS and english and german reference [couldn't of course check arab and persian texts]... but it could be a good hypothesis. So....
_
Na'am's mansuba in Europe
In Alfonso X's Libro de los Juegos, 1283. Here, this mansuba is found in 4 versions.
Chess problem # 64, f.42V: same position as al-Adli's [H16] // Chess problem # 59, f.40R // Chess problem # 65, f.43R // Chess problem # 98, f.60V
In this website photos of the book can be found. While Los libros de acedrex dados e tablas: Historical, Artistic and Metaphysical Dimensions by Sonja Musser, University of Arizona is one of the best on the topic.
In Repetición de amores y Arte de ajedrez, by Luis de Lucena, 1496. Same position as al-Adli's
_
Gioachino Greco enriched the position a little.
From Gioachino Greco by W. Lewis, 1819, p. 144
Some more on Abu Naim al Khadim
In my previous research I bumped into some Na'im's mansubat that it would be a pity not to share. The first three could be set in a puzzle form, but have in mind that win can come with the bare king rule too, that is to take all your opponents pieces, while your opponent can't take your last one in the next move. Don't also forget that farzin [Q] moves only 1 square diagonally, and fil [B] 2 squares diagonally with a leap.
#1
Murray, p.307: "The solution of this problem, which is a classical position in the modern treatment of the ending R v. Kt, varies in the different MSS... The position occurred in a game between Rabrab (red = white) and Nā‘īm al-Khādim, and Rabrab made an exhaustive study of the ending (AH)."
No10: MSS: AH18 - C120 - Man2 - BM143 - AE72 - V0 (end of text only)
#2
Murray, p.307: "From a game, Rabrab (white) v. Nā‘īm."
No12: MSS: AH20 - C122 - BM139 - V5 - AE90 - H24
#3
Murray, p. 307: "From a game, Abū’n-Na‘īm (black) vs Rabrab."
No13: MSS: AH21 / C7 / H21 / V6 / BM146=183 / AE70
Till now Rabrab vs Na‘īm 3-0 
.
I found other three that cannot be set in a puzzle form, so just for presentation.
#4
"...and Red-white, confining the Black K in the corner, queens his P and brings it across and mates (AE). AH concludes with the personal note (? by aṣ-Ṣūlī): ‘This happened to me when playing a man at odds. Abū’n-Na’ām boasted that he had played a similar game. There is not, however, one in the least like it among the problems of Abū’n-Na‘ām.’" Murray, p. 310, No64: MSS: AH72 / AE108 / C148
#5
Murray, p.317, No168: "Red plays and Black wins. 1. a5+ Kxa5 2. Kxc5 Qd6+ 3. Kxc4 Kb6 4. Kd5 Qc5 wins. Or 3. Kxd6 Kb6 4. Kd7 (or c5, Kb5), c5 wins. By an-Na’ām, who was proud of it,
MSS: AH." AH183 / C90 / Y40 / BM83 / AE107
Till now, arabic MSS represent Na’ām bragging about his successes, while found only three games against Rabrab, where Na’ām lost.
#6
I couldn't find the next one mansuba neither in Murray nor in some MS 
. I bumped into it in an archived Jeremy Silman's review on Chess Gems by Igor Sukhin (2007) and also was represented in Chess Horizons Jan 2008, p. 33, again as a Chess Gems review.
By Chess Horizons:
"...another mansuba of Abu Naim Al-Khadim, of which he was quite proud [again??]."
...I hope you've enjoyed this journey I took in arabic MSS that started since my previous blog
