More on the Alexandria Engine

    In a previous post I noted that I had been experimenting with the Alexandria engine and noted that it might be useful for uncovering an opening surprise when preparing for over the board play. 

    The following game was recently played online and using Fritz, I tested Alexandria by letting it analyze the game at 15 seconds per move with the idea of then checking the analysis against Stockfish’s. 

    To my surprise, Alexandria suggested no improvements! What I ended up doing was analyzing the game with Stockfish and then going through it again with Alexandria running in orfer to compare the results. 

    Unlike other engines that will allow you to see multiple lines, Alexandria only shows its top choice and it only uses one core. Its evaluations in Centipawns seems considerably higher than Stockfish. 

     As mentioned previously, while Version 8 is obviously very strong based on its Computer Chess Rating List results, it no longer seems to be available. The version 5 that I have does not seem to be practical for analysis. While it may be able to discover some hidden possibilities in the opening, a better engine for that purpose is probably Berserk which is also highly rated on the CCLR rating list and can be downloaded from the same site as Alexandria HERE. 

     Last year I played two games against a 2300+ otb Master who is also an ICCF International Master because I wanted to test the highly aggressive Berserk to see if it could discover any opening innovations. It didn’t and once we got out of the opening I felt my positions were slipping a bit even though Berserk showed things were equal, It was then that I switched back to Stockdish and we played two uneventful draws. 

    I think what all of this shows is that while it may be true that other engines might be useful for discovering opening innovations for otb play, Stockfish is still unsurpassed and I cannot see using any other engine fot post-game analysis.

[Event "Online G/10"] [Site "?"] [Date "2025.??.??"] [Round "?"] [White "Tartajubow"] [Black "Anonymous"] [Result "1-0"] [ECO "C44"] [Annotator "Stockfish/Aleandria"] [PlyCount "67"] [EventDate "2025.??.??"] {C44: Ponziani Opening} 1. e4 e5 2. Bc4 Nc6 {After the more frequently seen 2.. ..Nf6 I like to play the Urosov Gambit (3.d4) which is not entirely sound but it can give white attacking chances. Also reasonable was 3.f4.} 3. d4 {This is definitely not good! The Vienna (3.Nc3) would still offer white the opportunity of entering a gambit line.} exd4 {Alexandria's choice.} (3... Nxd4 {seems preferable because white then has little choice but to offer the trade of Ns with 4.Nf3} 4. Nf3 Nxf3+ 5. Qxf3 Qe7 {White does not have quite enough compensation for the P.}) 4. Nf3 Bc5 5. c3 d3 6. O-O {More common is 6.b4 which is somewhat better.} d6 7. Qxd3 Nf6 {It was surprising to discover that my database has 31 games which reached this position, but only one of the players was a Master which is not surprising. White's best move is 8.b4} 8. Ng5 {The threat to capture on f7 is no threat at all, nut I didn't see any other way of continuing that offered any play at all. However, black had been moving almost instantly and I thought this might gove him pause for thought.} O-O { No pausing to think; he played this instantly.} (8... Ne5 {is much stronger. After} 9. Qd1 Nxc4 10. Qa4+ Nd7 11. Qxc4 Ne5 12. Qe2 Bg4 13. Nf3 Nxf3+ 14. gxf3 Bh3 {Black is netter, but sy yjr non-engine level the "black is winning" comment by Fritz is not a foregone conclusion.}) 9. Qg3 {This earns a ? from Stockfish and both engines recommend that the N tuck its tail and return to f3. In these types of positions it's rarely a good idea to capture the f-Pawn.} Ne5 10. Bb3 h6 11. Bc2 {The N is bait hoping to get some kind of an attack.} (11. Nxf7 {is the recommended move.} Nxf7 12. Bxh6 Nh5 13. Qg6 Qf6 14. Qxh5 Qxh6 15. Qxh6 gxh6 {White's hoped fpor attack has not materialized and he only has two Ps for the B so black's position should prove to be a winning one.}) 11... hxg5 {Taken without hesitation, but there was a better move. That said, there is nothing at all wrong with 11...hxg5 which also leaves white with a losing position.} (11... Nh5 12. Qh4 Bg4 13. Bd1 Bxd1 (13... hxg5 {and the Q is trapped.}) 14. Rxd1 g6 {White is in serious trouble.}) 12. Bxg5 Ng6 13. Re1 Nh5 {The problem with this move, which technically speaking is not a bad one, soon becomes apparent} (13... Qe8 {is the top choice of Alexandria and the second for Stockfish which showed a slight preference for 13...Qd7. In either case, black has a much larger advantage that after the text. Almost 4 Pawms according to Stockfish which is long way from A,exandria's 7 Ps.]}) 14. Bxd8 Nxg3 {It should be obvious that white never had a real chance at conducting a K-side attack and now has scant compensation for the N.} 15. Bxc7 {I think black missed this. White now has two Ps for the N.} Nh5 16. b4 Bb6 17. Bxd6 { Suddenly white has 3 Ps for the N. Engines still consider black to have a decisive advantage though. In Shootouts whitr losr 0-5, but the endings were long and difficult and far beyond the skill level of the two players in this game.} Rd8 18. e5 Nhf4 19. Nd2 {[%mdl 32] The plan is to replace the B on d6 with the better piece, the N.} Ne6 {19...B6 has better winning chances. The threat is to win with ...Nxe5} 20. Rad1 Bc7 21. Nc4 Bxd6 22. Nxd6 {White has 2 Ps for the N and has managed to achieve equality.} b6 23. g3 {To support the advance f4, but more exact would have been 23.Bb3} Rb8 {[%mdl 8192] Pointless.} (23... Ng5 {takes advantage of white's imprecise last move.} 24. Re3 Nxe5 { Now that the Ns support had been removed black again has gained a significant advantage.} 25. f4 Ng4 26. Re2 Ne6 {and black is back to having a N vs, 2 Ps.}) 24. f4 {[%mdl 2048] White has strong initiative and it;s black who is in jeopardy.} Ne7 25. f5 Ng5 26. Kf2 Bb7 {An obvious tactical blunder. Black, who had started the game moving almost instantly, had gradually begun using more and more time and now had less than two minutes left to my four. Best, but insufficient, would have been 26...Kf8} 27. Nxb7 Rxd1 28. Rxd1 Rxb7 29. Rd8+ Kh7 30. h4 {The N has nowhere to go.} Nc6 {A good try. Now I had to spend some precious time trying to decide whether or not to move the attacked R. U finally secided to complicate the issue, mostly because of my opponent's lack of time.} 31. e6 (31. Re8 Rd7 32. Ke3 Nxe5 33. hxg5 Nc4+ 34. Kf4 g6 35. f6 Rd5 36. Re7 Kg8 37. Rxa7 {would ultimately win, but it will take some time which isnt; available in this blitz game.}) 31... Nxd8 32. f6+ $1 Kh6 33. e7 Nh3+ 34. Kf1 {In this hopeless position black lost on time. A lucky escape for white. Accuracy: White = 64%, Black = 36%.} 1-0