Week 36: how to get to 1000 Elo, one winning streak at a time
Week 36 is the week I got back over 1000. From September 10 to 15, 2024, I played 27 rapid games, won 17, lost 9 and drew 1, and my rating went from 945 to 1007 (+62). The first ten of those games were all wins, the longest streak of this stretch of the diary, and they took me to 1025. So, how to get to 1000 Elo? In my case it took a streak and a lot of my opponents' blunders, and the engine still found big mistakes inside the wins.
The week in numbers
| Games | 27 (#844–#870) |
| Wins / losses / draws | 17 / 9 / 1 |
| Rating | 945 → 1007 (+62) |
| Average accuracy | 78.7 % |
| Blunders (my moves) | 49 (1.8 per game) |
| Most played openings | Queens Pawn Opening Zukertort Chigorin Variation (5), Queens Pawn Opening Zukertort Variation (3), Queens Pawn Opening (2) |
The streak ran from game 844, a checkmate with Black against Shishka526 (893) at 99.2 % accuracy, to game 853. I crossed 1000 at game 850, and it wasn't pretty: an Englund Gambit against MarshallZyn (986) that I won on time, with 8 blunders of my own and 73.2 % accuracy.
After the streak the week was ordinary. From game 854 to game 870 I went 7 wins, 9 losses and 1 draw, and my rating drifted from 1025 back to 1007. With White I finished 10–3; with Black, 7–6 with the draw. Five of my nine losses were recorded as "game abandoned".
How to get to 1000 Elo: what the streak was made of
Looking at the numbers, the streak was not a sudden jump in my level. My average accuracy for the week was 78.7 %, close to the weeks around it. What changed was the result column: 8 of my 17 wins ended in checkmate and 5 by resignation, so my opponents' mistakes were punished and the games were finished. My most played setup was again 1.d4 with Bf4 (Chess.com's "Zukertort Chigorin" five times and "Zukertort Variation" three times).
The three key moments the analysis picked for this week all come from games I won during the streak. That is the honest part of this diary: winning ten in a row felt like a breakthrough, but the analysis shows I was winning while still missing the strongest moves.
Three mistakes hidden inside wins
Game 845: trading my best piece instead of adding a second attacker
With White against rayvskp (939), Black had castled queenside and the engine had me at +4.81, "white is winning".
I traded my bishop for the knight. Position.py describes Ne5 as a double attack: it hits the knight on c6, the queen on d7 and the undefended pawn on f7 at once. The engine's line is 9.Ne5 Qe8 10.Nxc6 bxc6 11.Bxa6+, a check on the opened diagonal. After my trade and Black's recapture with the queen, the score dropped to +0.81. I still won the game by checkmate.
Game 846: opening the centre with my king still at home
With Black against Extrakz (938), I was better (-1.34) after the opening, but my king was still on e8 and my knight on g8.
The pawn break d5 lets White trade in the centre: the engine's reply cxd5 also attacks my bishop on e6. After it the position is "slightly better for white" (+0.55). The quiet Ne7, followed by castling in the engine's line, kept "slightly better for black". I won this game by checkmate too, but the advantage I had after the opening was gone by move 10.
Game 847: the poisoned pawn on b7
With White against naitsabesrl (944) I was already slightly worse (-0.72) when my queen saw a free pawn.
Qxb7 attacks the knight on c6 and the rook on a8, so it looks active. But the engine's reply Rab8 "attacks the queen on the queen's side, which has no defender", and after 11.Qa6 Rxb2 Black's rook was in my position. The evaluation went to -2.80. Bb5 (or c4, which the engine rates the same) kept the game close. My opponent gave it back with 12...Nxd4, and they later abandoned the game, so it went into my streak as a win at 51.8 % accuracy.
The lesson of the week
In all three positions I chose the move that takes something: a knight, a central pawn, a pawn on b7. The engine chose the move that improves a piece first — a knight to a strong square, a knight off the back rank, a bishop to an active diagonal. At 950 that difference didn't cost me these games, because my opponents made bigger mistakes. But it explains why the streak ended and the week finished at 1007 instead of 1025. If I want to stay above 1000, my question before capturing should be: is there a move that adds an attacker or develops a piece, and keeps the capture for later?
All games this week
| Game # | Date | Colour | Opponent (rating) | Result | Opening | Accuracy | Link |
|---|---|---|---|---|---|---|---|
| 844 | Sep 10 | Black | Shishka526 (893) | Win | Queens Pawn Opening | 99.2 % | game |
| 845 | Sep 10 | White | rayvskp (939) | Win | Queens Pawn Opening Zukertort Chigorin Variation | 86.7 % | game |
| 846 | Sep 10 | Black | Extrakz (938) | Win | Giuoco Piano Game | 84.4 % | game |
| 847 | Sep 11 | White | naitsabesrl (944) | Win | Queens Pawn Opening Zukertort Chigorin Variation | 51.8 % | game |
| 848 | Sep 11 | Black | SalievSh (951) | Win | Giuoco Piano Game Giuoco Pianissimo Variation | 95.6 % | game |
| 849 | Sep 11 | White | alitafreshi555gmailcom (971) | Win | Queens Pawn Opening Zukertort Chigorin Variation | 89.5 % | game |
| 850 | Sep 11 | White | MarshallZyn (986) | Win | Englund Gambit | 73.2 % | game |
| 851 | Sep 12 | Black | stelserg (990) | Win | Ruy Lopez Opening Cozio Defense | 75.8 % | game |
| 852 | Sep 12 | Black | 1212sanel (1009) | Win | Kings Fianchetto Opening | 92.4 % | game |
| 853 | Sep 12 | White | tanmayplayschezz (999) | Win | Queens Pawn Opening Zukertort Variation | 92.3 % | game |
| 854 | Sep 12 | Black | Lonely56 (1030) | Loss | Kings Pawn Opening Leonardis Variation | 52.6 % | game |
| 855 | Sep 12 | White | kuzuu04 (1015) | Win | Queens Pawn Opening Zukertort Chigorin Variation | 82.9 % | game |
| 856 | Sep 12 | Black | Kamilshukry (1046) | Loss | Scotch Game Classical Variation | 55.4 % | game |
| 857 | Sep 12 | Black | Rick_Astley_1966 (1004) | Loss | English Opening Reversed Sicilian Variation | 80.6 % | game |
| 858 | Sep 13 | White | DCraven27 (987) | Win | Queens Pawn Opening Zukertort Variation | 97.1 % | game |
| 859 | Sep 13 | White | mirz1985 (1038) | Loss | Queens Pawn Opening | 60.0 % | game |
| 860 | Sep 14 | White | armaanagrawall (1015) | Loss | Queens Pawn Opening Zukertort Variation | 87.4 % | game |
| 861 | Sep 14 | Black | lugman611 (992) | Win | Center Game Accepted Kieseritzky Variation | 91.1 % | game |
| 862 | Sep 14 | Black | thedogfromuptheroad (1000) | Win | French Defense | 86.5 % | game |
| 863 | Sep 14 | White | ganesh2307 (1018) | Win | London System | 53.6 % | game |
| 864 | Sep 14 | Black | Paul_Hay09 (1044) | Loss | Kings Gambit | 61.6 % | game |
| 865 | Sep 14 | White | baba_ayoub (1016) | Win | Queens Pawn Opening Horwitz Defense | 75.6 % | game |
| 866 | Sep 15 | Black | kalamambo (1045) | Draw | Kings Pawn Opening | 78.4 % | game |
| 867 | Sep 15 | White | PiotrIV (1048) | Loss | Queens Pawn Opening Horwitz Defense | 69.6 % | game |
| 868 | Sep 15 | Black | Goutham18 (1003) | Loss | Giuoco Piano Game Evans Accepted McDonnell Defense | 70.3 % | game |
| 869 | Sep 15 | White | aurora_photon (986) | Win | Queens Pawn Opening Zukertort Chigorin Variation | 94.3 % | game |
| 870 | Sep 15 | Black | nexttimeplaysober (1016) | Loss | Giuoco Piano Game | 86.6 % | game |
Week before: week 35 · Week after: week 37 · The full journey · Review your own games like this
Frequently asked questions
How did I get to 1000 Elo again?
By winning my first ten games of week 36, from game 844 to game 853, which took me from 953 to 1025. My opponents averaged 997 that week, and many of my wins came from punishing their blunders rather than from clean play.
Is a winning streak a sign you have improved?
Not on its own. The three biggest mistakes the engine found this week all came from games inside my streak, and after the streak I went 7–9–1 and finished the week at 1007.
What is a poisoned pawn?
A pawn that looks free but costs you more than it is worth. In game 847 my queen took the pawn on b7 and Black's rook came to b8 to attack the queen; the evaluation went from -0.72 to -2.80.