Chess simul: Zedekiah's Cave, Jerusalem

Zedekiah's Cave is Jerusalem's ultimate hidden gem. Inconspicuous from the outside, its interior is a vast network of interconnecting chambers, a spring and ancient limestone blocks made all the more impressive by colorful atmospheric lighting. During the Second Temple period the site was a quarry for stones used in Jerusalem's construction, and since the nineteenth century it has served as a ceremonial hall for local Freemasons. Yesterday it was the site of a simultaneous chess exhibition where forty players gathered to take on Grandmaster Nitzan Steinberg.

My first experience at a chess simul was in September 2024. Jerusalem hosted Israel's largest ever such event at Safra Square, and I played against GM Alexander Zlatin. That game was a Chigorin French Defense, and I not only held my own for most of it but had winning chances twice before carelessly blundering away my advantage. Yesterday's game was a Steinitz French, and this time opponent had the upper hand for all but one brief moment. Interestingly, both games' missed opportunities involved me failing to capitalize on opponent trapping his own minor piece.

Amir Afsai, Board 21
Photo: Mark Livshitz

Shortly after five o'clock, the players took their seats at fifteen tables arranged in a U and draped in black and white checkered tablecloths. At the center of the U was a giant chess set and at the top were chairs and microphones for a soon-to-arrive musical ensemble led by Michael Halevi Elkabetz. In the space between each two chessboards, three LED tealights flickered, adding to the venue's ambience. I sat at the centrally-located Board 21 with my notebook and Hydro Flask and began adjusting the pieces on my board.

Alon Cohen, CEO of Jeruchess and the driving force behind the event, officially welcomed us to Zedekiah's Cave and instructed us as to the etiquette we were expected to follow, including not making our moves until GM Steinberg arrived at our board. There was no need to warn against the use of phones as no signal could penetrate to our depth. Steinberg walked from board to board, shook each player's hand, and made the first move. True to his YouTube persona, he sprinkled insightful commentary and witty remarks into his interaction with the players, making for an experience as entertaining as it was instructive.

---

Position 1: Black to move

Assigned the Black pieces, I replied to White's 1.e4 with the French 1...e6, and after 2.d4 d5 3.Nc3 Nf6 I was relieved White didn't continue 4.Bg5 and instead advanced 4.e5. My knight repositioned with 4...Nd7, and after 5.f4 I struck with 5...c5. We developed knights with 6.Nf3 Nc6, and Lichess' opening database indicates that 7.Be3 is the Boleslavsky Variation of the Steinitz French. The game was still roughly equal, but my 7...Qb6 gave White a slight edge of +0.5 and was the first crack in my position.

1. Why did I choose the move 7...Qb6?
It seemed to me that as long as White couldn't push a pawn to c3 there was no rush to relieve the central tension between c5 and d4. Meanwhile, 7...Qb6 added pressure to d4, attacked b2 and gave my queen access to the checking square b4.

2. Why is my move not ideal?
I was expecting 8.Na4, but for some reason I calculated that 8...Qb4+ would prompt 9.Nc3, not 9.c3. My queen got pushed around while White achieved a strong center -- exactly the situation I had sought to prevent.

3. Why is the better move better than my chosen move?
The engine wanted me to chop with 7...cxd4 and only then deploy my queen to b6. After 7...cxd4, White would have to rechop with a piece. I would gain a semiopen c-file and achieve a disconnect between White's c- and e-pawns, and my pieces would enjoy greater freedom of movement while the dark squares to White's kingside would become a potential avenue of attack.

---

Position 2: Black to move

After 8.Na4 I needed to give a check not with 8...Qb4+ but with 8...Qa5+, allowing 9...cxd4 and White can't rechop with 10.cxd4 because of my pin. After 9.c3 my queen got pushed back with 9...Qa5, and now White could chop with 10.dxc5 because of the threat of pawn-forking with 11.b4 in the event of 10...Nxc5 or 10...Bxc5. Judging from his reaction to my 10...Qc7, Steinberg was probably expecting me to fall into his trap. I saw through it, but my move missed an idea and dropped my advantage from -0.5 to +2.5.

1. Why did I choose the move 10...Qc7?
I knew I couldn't rechop with 10...Nxc5 or 10...Bxc5 on account of 11.b4 pawn-forking two minor pieces. With my queen on c7, I thought at least she would be off the a-file and then I could reclaim initiative on the queenside beginning with an attack on c6.

2. Why is my move not ideal?
There was nothing inherently wrong with 10...Qc7 -- it was the engine's second choice -- but it missed a tactic stemming from the absence of escape squares for White's a4-knight.

3. Why is the better move better than my chosen move?
The engine's top move, and the only move that kept the game equal, was 10...b5. If 11.cxb6 then 11.axb6 and White's knight on a4 falls. If 11.b4 then 11...Qxa4 12.Qxa4 bxa4 13.Bb5 and I'm up a point of material but with a cramped position.

---

Position 3: Black to move

White correctly solidified his queenside pawn structure with 11.b4, and I correctly challenged c5 with 11...b6. White's pawns penetrated into my territory, first with 12.b5 Na5 and then with 13.c6 Nc5. After exchanging knights with 14.Nxc5 bxc5, White resolidified with 15.a4 and I pushed 15...a6. White's rook repositioned with 16.Rb1, and I advanced 16...c4 to reduce the scope of White's light-square bish and to fix the weakness of White's backward pawn on c3. White developed with 17.Be2, and we exchanged with 17...axb5 18.axb5.

Now it seemed to me I could claim compensation in the form of White's weak queenside pawns, and with 18...Nb3 it seemed to me I would soon start munching on them. White castled with 19.0-0, and I infiltrated with 19...Ra2. After 20.Re1 my queen attacked b5 with 20...Qa5, but White spoiled my plan by saccing his rook for my knight with 21.Rxb3 cxb3 22.Qxb3. Material was equal, but I had failed to chop White's queenside pawns. 22...Ra3 prompted 23.Qb1, and at last I chopped a pawn with 23...Qxc3. White attacked my queen with 24.Bd4 and his advantage dropped from +2.5 to +1.7, but it surged to +4.6 after I offered a queen trade with 24...Qb3.

1. Why did I choose the move 24...Qb3?
If I had any hope of chopping White's b-pawn, the pawn needed to be cut off from its defenders. If I could exchange queens and then rooks, then I could castle and maybe my bish pair and rook would suffice to get the job done.

2. Why is my move not ideal?
With my dark-square bish still undeveloped and my rook stuck on h8, my priority needed to be mobilizing all my forces. There was a move that mobilized my dark-square bish without the loss of a tempo, but I overlooked it.

3. Why is the better move better than my chosen move?
The engine's top move was 24...Bc5, effectively forcing a bish trade with 25.Bxc5 Qxc5+ and allowing me to castle and finally liberate my rook from the corner.