11 losses in a row: inside my longest losing streak on Chess.com
Between April 8 and April 11, 2024, I lost eleven rapid games in a row on Chess.com: games #355 to #365 of this journey. My rating dropped from 1051 to 974, 77 points in four days. In 2,116 rapid games it is still my longest losing streak on Chess.com, and I've gone back to every one of the eleven games for this post.
The weekly diary for week 14 already covers the week as a whole. This page goes game by game through the streak itself, because the pattern only shows up when you look at how each loss ended, not just that it was a loss.
How I got there
Game #354, on the evening of April 8, was a win that put me at 1051, equal to my best rating so far in the journey. I'd crossed 1000 for the first time at game #232, about six weeks earlier. The streak started with my very next game.
The eleven games
All games were 20-minute rapid, no increment. Accuracy and blunder counts are Stockfish's.
| # | Date | Colour | Opponent (rating) | How it ended | Accuracy | Blunders |
|---|---|---|---|---|---|---|
| 355 | Apr 8 | Black | Vag89 (1214) | I resigned | 55.6 | 2 |
| 356 | Apr 9 | White | cruiser418 (1079) | I resigned | 75.2 | 3 |
| 357 | Apr 9 | Black | Ogre2016 (1068) | I resigned on move 5 | 57.7 | 1 |
| 358 | Apr 9 | Black | NeverGonnaCheckYouUp (1073) | I resigned on move 8 | 63.9 | 1 |
| 359 | Apr 9 | White | Brigitteu (1063) | Abandoned | 88.7 | 0 |
| 360 | Apr 9 | Black | Far_cool (1004) | I resigned | 63.1 | 3 |
| 361 | Apr 10 | Black | SikkenBerend (1172) | I resigned | 69.8 | 4 |
| 362 | Apr 10 | Black | Andrey19091980 (998) | Abandoned | 76.0 | 3 |
| 363 | Apr 10 | White | machoem (1108) | Checkmated | 83.7 | 2 |
| 364 | Apr 11 | White | siegb (1020) | I resigned | 60.5 | 8 |
| 365 | Apr 11 | White | percusantos (984) | Abandoned | 66.7 | 5 |
Put together: 32 blunders in eleven games, an average accuracy of about 69, five games with White and six with Black, and eleven different opening labels. Opponents ranged from 984 to 1214. Eight of them were rated within 50 points of me. Five games came on a single day, April 9.
There's no single opening to blame. What repeats is something else.
Pattern 1: I was winning in more than half of them
In six of the eleven games (#355, #356, #361, #362, #364 and #365), the engine had given me a winning chance of at least 80% at some point. Game #355 against Vag89 was the extreme case: I was crushing, missed a pawn on b6 that was walking up the board, and lost to a promotion. Game #356 against cruiser418 had the same shape. On move 12 I had a bishop sacrifice on h7 that the engine scored at +7.45, and I played a quiet bishop move instead. Ten moves later, at +9.88, I missed a queen capture on f7 with check. By move 30 I was lost.
Losing a winning position once is a chess problem. Losing six of them in four days is something else.
Pattern 2: two games over almost before they began
Games #357 and #358 were played half an hour apart on the morning of April 9, and neither reached move 9.
In game #357 Ogre2016 opened 1.e3 and brought the queen out to b3 on move 3. On move 4 it took my b7 pawn and attacked the rook in the corner. I pushed the c-pawn. The queen took the rook, and I resigned on the spot, on move 5, with more than 19 minutes on my clock. The engine's move was knight to d7: with both the bishop and the knight gone from the back rank, my queen on d8 defends the rook on a8.
Game #358 was a Danish Gambit. On move 7 I pushed d6 instead of developing the knight to f6, White's bishop took my g7 pawn and attacked the rook on h8, and I resigned after move 8, again with more than 19 minutes left.
In both games the position after the loss of material was bad, but resigning with nineteen minutes left in a 20-minute game is not a chess decision.
Pattern 3: leaving games I didn't need to leave
Three of the eleven games, #359, #362 and #365, are recorded by Chess.com as abandoned. I can't tell you from the data whether each one was a choice or a disconnection. What the engine can tell me is the position I left behind. In game #359 against Brigitteu, my most accurate game of the streak at 88.7, the last position is rated clearly better for White, for me (+1.67), and I had almost 15 minutes on my clock.
Game #364 against siegb ended with my resignation after 45 moves. I went back to the final position: the engine rates it as only slightly better for Black (-0.69). That game also had my worst numbers of the streak, eight blunders, including this one:
My queen on c2 was one of the two pieces defending my knight on e4, which Black's knight on f6, bishop on f5 and rook on e8 were all attacking. When I took the pawn, the knight on e4 had one defender left, and Black's bishop took it. The evaluation went from +2.54 to -5.78. The engine's order was the check on f6 first, then the pawn on c4.
The game that sums up the streak: #365
The last loss was against percusantos (984), the lowest-rated opponent of the eleven. On move 9 I pushed my a-pawn to a3 to attack Black's queen on b4. Black answered by taking my knight on d2 with their knight.
The queen was still there to take. The engine scores taking it as completely winning (+6.45), even after Black's knight jumps to b3 and wins my rook in the corner. I took the knight instead, and the queen escaped to a5. The evaluation fell to +1.20. Ten moves later I was lost, and the game ended abandoned. I was attacking the queen with my own pawn on move 9 and didn't take it on move 10.
That's the move I keep coming back to, because it doesn't need any chess knowledge to find. It needs attention.
How it ended
Game #366, later the same day, was a win with Black against Olga3m777 (972). It was a quiet game: no mistakes and no blunders by the engine's count, three inaccuracies, and an accuracy of 93.6, my best of that week. On move 15 my bishop took the pawn on h3, which the tool describes as a double attack that also threatens mate.
The bishop took the rook on f1 a move later, and White resigned after my 21st move. One win in a row, after eleven losses. I won the next game too.
What the streak cost, and what I take from it
The rating bottomed out at 974 after game #365, my lowest since early March. I was back to 1001 three days later, at game #375, but I didn't get above 1051 again until game #612, at the end of June. So four days of tilt cost me roughly two and a half months of progress, measured in rating.
What I take from it is simpler than any chess lesson. The engine found the same three things in almost every game of the streak: winning positions not finished, material grabbed in the wrong order, and games given up early, either by resigning or by leaving them. None of that is new knowledge. When I see a second loss from a winning position on the same day, the rule I took from it is to stop and look at that one moment before playing again.
The week diaries before and after are week 14 and week 15, and the full climb from 631 to 1385 is on the journey page. One of the streak games, #361, repeated a Giuoco Pianissimo knight move I'd got away with in my upset win in game #311.
A streak like this is easy to see in a rating graph and hard to see from inside. Having the engine name the moment in each loss, in plain words, is what I built Chess Coach to do: here's how the analysis works, and why Chess.com's game review doesn't always explain the moves.
Frequently asked questions
How long was my worst losing streak on Chess.com?
Eleven rapid games in a row, #355 to #365, between April 8 and April 11, 2024. My rating went from 1051 to 974. In 2,116 rapid games no other losing run was as long.
Was the streak caused by bad openings or by tilt?
The data points at tilt more than openings. Eleven games, eleven different opening labels, and in six of them the engine had given me at least an 80% winning chance at some point. I resigned two games before move 9 and resigned another in a position the engine rates only slightly worse for me.
How did the losing streak end?
With game #366, a win as Black against Olga3m777 (972). It was my most accurate game of that week: 93.6, with no mistakes or blunders by the engine's count.