Chess report
Almost all of the damage comes from middlegame moves played at reflex speed, usually two or three in a row, in positions where nothing was forcing you to hurry.
Every one of my daily and rapid games is pulled from chess.com, replayed through a local Stockfish at a fixed depth, and reduced to a per-move centipawn loss. Those evaluations are then aggregated in plain code and handed to Claude twice over the same games: once for the handful of recurring patterns costing me rating points, and once for what holds up. Neither is a list of individual moves. Every move either pass cites is checked back against the evaluation store before it reaches this page, and a finding whose citations don't resolve is thrown out rather than published, including a flattering one.
All my slower games are downloaded from chess.com and replayed through a chess engine, which scores every move by how much it cost compared to the best one available. Those scores are added up and handed to an AI twice over the same games: once to ask what habits keep losing me points, and once to ask what I can rely on. Every move either answer points at is checked against the real games first, and anything it can't back up is thrown away instead of printed here, compliments included.
What I do well
You get out of the opening in good shape almost every game, and your d4 systems as White are the most reliable part of your chess.
You reach the middlegame in good shape high confidence
Your first twelve to fifteen moves are the cleanest part of your game. You come out of the opening with a position that is equal or better from your side in nearly every game, whichever colour you have.
When it shows up
Opening phase, moves 1 to 15, both colours, all 13 games.
What the engine counted
Opening: 156 moves, 0 blunders, 2 mistakes, avg 22.7cp. Moves 1-10: 130 moves, 0 blunders, 1 mistake, avg 19.1cp. Compare your own middlegame at 54.8cp and your moves 21-30 at 56.1cp: your opening loss rate is well under half your middlegame rate, and every one of your 8 blunders came after move 12. Reading the eval trajectories, at move 12 you were level or better in 10 of 13 games, and the worst you ever stood that early was -2.0 (G7).
What to do with it
Opening theory is not where your rating points are going, so stop buying courses for it. Your study time belongs in the first fifteen moves after the book ends, which is exactly where the eval starts to move. One practical use of this: you can play your openings at a brisk pace without worrying, and bank the time for moves 15 to 30.
The moves behind it
- Move 7 Nf3: engine's own move in a quiet London setup, holding +0.4
2026-08-21 · Queens Pawn Opening Accelerated London System · the engine’s own move - Move 12 dxe5: correct central capture, position held at +0.2
2026-09-06 · Queens Pawn Opening Zukertort Variation...3.Bg5 Be7 · the engine’s own move - Move 10 exd4: engine's move, keeps the game on rails as White
2026-09-15 · Queens Pawn Opening Accelerated London System · the engine’s own move - Move 12 Nxe4: as Black against the Smith Morra, correct capture at +1.3
2026-10-02 · Sicilian Defense Smith Morra Gambit · the engine’s own move - Move 6 d5: engine's move, equal at +0.1 out of the opening as Black
2026-09-03 · Sicilian Defense Smith Morra Gambit · the engine’s own move - Move 6 Nf6: sound development, position level at -0.4
2026-10-15 · Sicilian Defense Old Sicilian Variation · within 8cp of the engine’s move
Your d4 systems as White hold up high confidence
The London and Zukertort setups you play as White are your most dependable repertoire, both in the numbers and in the results. You handle those structures better than anything you play as Black.
When it shows up
White, 1.d4 systems (Accelerated London, Zukertort, Symmetrical), against opponents rated 1188 to 1263.
What the engine counted
White: 227 moves, 3 blunders, avg 33.3cp. Black: 212 moves, 5 blunders, avg 45cp. By opening, your Accelerated London is your best split at 26.4cp over 91 moves, with Zukertort 3.Bg5 at 31.1cp and the Symmetrical at 32.3cp, against the Smith Morra at 49cp and the Taimanov at 51.6cp as Black. Results follow the same line: as White 5 wins, 1 draw, 1 loss (including wins over a 1263 and a 1207); as Black 1 win, 1 draw, 4 losses. Your two cleanest games in the sample are both White d4 games (G13 at 17cp average, G9 at 24.4cp average with no mistakes at all).
What to do with it
Keep this repertoire and deepen it rather than widening it. The gap between your colours is where the rating is: put the study time into what you do against 1.e4 c5 2.d4 in particular, since the Smith Morra games account for one loss, one draw and a 49cp average over 89 moves.
The moves behind it
- Move 21 Nxd7: engine's move, converting a growing edge in the London
2026-08-21 · Queens Pawn Opening Accelerated London System · the engine’s own move - Move 24 a5: engine's move, keeping the win on track at +5.0
2026-08-21 · Queens Pawn Opening Accelerated London System · the engine’s own move - Move 15 O-O: engine's move in a balanced London middlegame
2026-09-15 · Queens Pawn Opening Accelerated London System · the engine’s own move - Move 20 Rxh5: engine's move, correct recapture in the London attack
2026-09-30 · Queens Pawn Opening Accelerated London System · the engine’s own move - Move 15 Nxa4: engine's move en route to beating a 1263
2026-09-17 · Queens Pawn Opening Symmetrical Variation · the engine’s own move - Move 20 Bf5: engine's move, accurate piece placement at +0.6
2026-09-06 · Queens Pawn Opening Zukertort Variation...3.Bg5 Be7 · the engine’s own move
Your judgement at your own normal pace medium confidence
When you spend roughly your usual amount of time on a move, your decisions are sound. The damage comes from the two ends of your clock: moves you fire off instantly and moves you grind on for ten times your pace.
When it shows up
Any phase, any colour. Moves taken at around 0.8x to 2x your median time in that game.
What the engine counted
Normal-pace moves: 97 moves, 0 mistakes, 1 blunder, avg 23.6cp. Snap moves: 133 moves, 9 mistakes, 3 blunders, avg 47cp. Long thinks: 162 moves, 8 mistakes, 3 blunders, avg 38.5cp. The normal bucket is the only one of the four with no mistakes in it, and its average loss is half the snap rate. Long thinking does not rescue you: G3 ply 26 cost 278cp at 15x your pace, G10 ply 30 cost 400cp at 33x, G4 ply 75 cost 339cp at 5.8x.
What to do with it
Trust your first considered idea instead of either banging the move out or grinding. When you notice you have been on a move far longer than usual, that is the signal to run a short fixed check (what did their move change, any checks or captures, is anything of mine loose) and then play. More time is not the lever for you.
The moves behind it
- Move 23 Qb3: engine's move at 0.89x your pace in a live middlegame
2026-09-17 · Queens Pawn Opening Symmetrical Variation · the engine’s own move - Move 14 Nc2: engine's move at 1.37x pace, position held at +0.7
2026-09-15 · Queens Pawn Opening Accelerated London System · the engine’s own move - Move 26 a6: engine's move at exactly your own pace
2026-08-21 · Queens Pawn Opening Accelerated London System · the engine’s own move - Move 19 Ba6: engine's move at 1.02x pace in a position you were worse in
2026-09-18 · Sicilian Defense Taimanov Variation · the engine’s own move - Move 14 Qxg6: engine's move at 1.57x pace, correct queen recapture
2026-09-30 · Queens Pawn Opening Accelerated London System · the engine’s own move - Move 18 Bh7: engine's move at 1.88x pace, holding the edge
2026-11-12 · Queens Pawn Opening Zukertort Variation · the engine’s own move
Clean once the pieces are nearly gone low confidence
In real endgames with few pieces left you stop giving things away, and you played a level rook endgame well enough to win it.
When it shows up
True endgames and moves 46 onwards. Shows up in two games only (G3, G12).
What the engine counted
Endgame: 41 moves, 0 blunders, 1 mistake, 0 inaccuracies, avg 7.3cp. Moves 46+: 27 moves with no errors at all, avg 1.4cp. Against your own middlegame at 54.8cp and moves 21-30 at 56.1cp, that is a different standard of play. In G3 the eval sat at 0.0 for roughly ten moves of a level endgame against a 1133 before you converted it to +7.3. Important qualifier: this is not a queens-off effect, since queens off (40cp) and queens on (38.5cp) are the same for you. It is specifically the thinned-out endgame.
What to do with it
You can simplify into an endgame a pawn up without fear, and trading into one is a legitimate way to play a position you do not understand. Do not spend study time on endgame technique yet. Bear in mind that simplifying while queens are still on does not help you, so the trade to aim for is the one that leaves a genuine endgame.
The moves behind it
- Move 43 Kd6: engine's move, correct king activation at -0.4
2026-10-15 · Sicilian Defense Old Sicilian Variation · the engine’s own move - Move 53 Rf7: engine's move, holding the rook endgame at 0.0
2026-10-15 · Sicilian Defense Old Sicilian Variation · the engine’s own move - Move 62 Kxd5: engine's move, correct king capture converting the endgame
2026-10-15 · Sicilian Defense Old Sicilian Variation · the engine’s own move - Move 26 Kd7: engine's move, correct king move holding a worse endgame
2026-09-03 · Sicilian Defense Smith Morra Gambit · the engine’s own move - Move 31 hxg5: engine's move, correct pawn capture while worse
2026-09-03 · Sicilian Defense Smith Morra Gambit · the engine’s own move - Move 34 g4: engine's move, salvaging the position at +1.9
2026-09-03 · Sicilian Defense Smith Morra Gambit · the engine’s own move
What I don’t do well
Reflex moves in the middlegame high confidence
When you move far faster than your own pace in that game, your move quality roughly doubles in cost. The damage is not one slip, it comes in clusters of two or three instant moves that turn a fine position into a lost one.
When it shows up
Middlegame, moves 11 to 30, in positions that look quiet or already decided. Both colours, but worst as Black.
What the engine counted
Snap moves (far below your own median time in that game): 133 moves, 3 blunders, 9 mistakes, 26 inaccuracies, 47cp average. Moves played at roughly your normal pace: 97 moves, zero blunders, zero mistakes, 23.6cp average. That is the sharpest split in the whole sample: the normal-pace band is clean, both fast and slow bands are not. By phase the middlegame is 54.8cp against 22.7cp in the opening and 7.3cp in the endgame, and moves 11 to 30 hold 5 of your 8 blunders.
The moves behind it
- Move 15 Nd7: 0.11x your pace, -140cp, in a position you had held level for ten moves
2026-11-05 · Sicilian Defense Smith Morra Gambit · cost 140cp · inaccuracy - Move 24 Nf8: 0.17x pace blunder that ends a game you were +1.1 in at move 20
2026-11-05 · Sicilian Defense Smith Morra Gambit · cost 327cp · blunder - Move 18 O-O: 0.16x pace, -291cp, from a comfortable +0.8
2026-10-02 · Sicilian Defense Smith Morra Gambit · cost 291cp · mistake - Move 21 Rfd8: 0.1x pace, -437cp blunder three moves later, game over
2026-10-02 · Sicilian Defense Smith Morra Gambit · cost 437cp · blunder - Move 18 Qxd5: 0.17x pace mistake
2026-09-06 · Queens Pawn Opening Zukertort Variation...3.Bg5 Be7 · cost 164cp · mistake - Move 19 Qxc4: second instant mistake immediately after, the cluster shape
2026-09-06 · Queens Pawn Opening Zukertort Variation...3.Bg5 Be7 · cost 153cp · mistake - Move 16 Ba6: 0.01x pace, -134cp, essentially no thought at all
2026-09-17 · Queens Pawn Opening Symmetrical Variation · cost 134cp · inaccuracy
Going slack the moment you are winning high confidence
Once the engine says you are two or more pawns up, your move quality falls off a cliff. You keep winning against weaker opponents anyway, but against someone your own strength the advantage leaks away over fifteen moves and the game flips.
When it shows up
Your own evaluation above about +2, queens on or off, from roughly move 20 onward.
What the engine counted
Your wins contain 10 mistakes and 29 inaccuracies across 240 moves, so the errors are not confined to losses. G1 is +4.5 by move 12 and still takes 75 plies with seven inaccuracies. G4 is the cost: +3.4 at move 16 as White, then 3.1, 1.2, 2.5, 1.9, 2.0, and finally -4.1 at move 40 against a 1236. That is a slow bleed, not a single tactic, and it is your only loss from a clearly winning position.
The moves behind it
- Move 25 Ra2: 0.13x pace, hands back most of a +3.4 lead
2026-10-14 · Queens Pawn Opening Zukertort Variation · cost 105cp · inaccuracy - Move 32 Re3: mistake that keeps the bleed going at around +2
2026-10-14 · Queens Pawn Opening Zukertort Variation · cost 154cp · mistake - Move 21 O-O-O: snap castling at +7, cost 141cp, no need to hurry
2026-11-12 · Queens Pawn Opening Zukertort Variation · cost 141cp · inaccuracy - Move 28 Bxf4: 0.04x pace blunder while comfortably winning
2026-09-30 · Queens Pawn Opening Accelerated London System · cost 311cp · blunder - Move 31 gxf3: 0.04x pace mistake at around +7
2026-09-17 · Queens Pawn Opening Symmetrical Variation · cost 290cp · mistake - Move 28 Qxc7: quick queen grab at +6.2 costing 156cp
2026-08-21 · Queens Pawn Opening Accelerated London System · cost 156cp · mistake
Already worse by move 15 as Black high confidence
As Black you tend to be a pawn or two down in evaluation before the middlegame starts, so the rest of the game is damage control. Four of your five losses are with Black, and in each one the slide begins in the opening.
When it shows up
Playing Black, especially facing the Smith Morra Gambit and in your own Sicilian setups.
What the engine counted
Black: 212 moves, 5 blunders, 45cp average. White: 227 moves, 3 blunders, 33.3cp average. By opening, Smith Morra 49cp over 89 moves and the Taimanov 51.6cp, against 26.4cp in your London System as White. Trajectories: G7 is -2.0 by move 12 and -3.1 by move 16, G10 is -1.2 by move 12 and -3.4 by move 16, G2 starts -1.6 at move 4. You lose the opening before anyone has blundered.
The moves behind it
- Move 8 Bb4: opening inaccuracy after a 12x pace think, already -1.1
2026-09-18 · Sicilian Defense Taimanov Variation · cost 57cp · inaccuracy - Move 10 O-O: 68x pace spent on castling and the eval still drops to -1.6
2026-09-18 · Sicilian Defense Taimanov Variation · cost 64cp · inaccuracy - Move 10 Nd7: 117x pace on an opening move, -131cp
2026-09-09 · Caro Kann Defense Exchange Variation · cost 131cp · inaccuracy - Move 12 Nf6: -144cp in the opening, position now -1.2 before any tactic
2026-09-09 · Caro Kann Defense Exchange Variation · cost 144cp · inaccuracy - Move 2 Nc6: 35x pace on move 2 and the eval moves to -1.6
2026-09-03 · Sicilian Defense Smith Morra Gambit · cost 115cp · inaccuracy - Move 12 Rd8: 0.09x pace opening inaccuracy in the Morra
2026-11-05 · Sicilian Defense Smith Morra Gambit · cost 141cp · inaccuracy
Long thinks that never get blunder checked high confidence
Your long thinks go wrong almost as often as your instant moves. The time appears to go on choosing between plans rather than on checking whether the move you picked loses material to a reply.
When it shows up
Middlegame moves where you sense something is wrong and stop to work it out, often at five to thirty times your own pace.
What the engine counted
Long moves: 162 moves, 3 blunders, 8 mistakes, 27 inaccuracies, 38.5cp average, against 23.6cp for normal-pace moves. Several of your single largest errors in the sample were your slowest moves of the game, which rules out time pressure as the cause.
The moves behind it
- Move 13 a5: 15.4x pace, -278cp mistake
2026-10-15 · Sicilian Defense Old Sicilian Variation · cost 278cp · mistake - Move 20 Qb5: 11.3x pace, -201cp mistake
2026-10-15 · Sicilian Defense Old Sicilian Variation · cost 201cp · mistake - Move 15 b4: 33x pace, -400cp blunder
2026-09-09 · Caro Kann Defense Exchange Variation · cost 400cp · blunder - Move 9 Nb3: 20.7x pace, -175cp in the opening
2026-09-17 · Queens Pawn Opening Symmetrical Variation · cost 175cp · mistake - Move 33 Rxg4: 7.9x pace, -173cp
2026-11-05 · Sicilian Defense Smith Morra Gambit · cost 173cp · mistake - Move 38 Kg2: 5.8x pace, -339cp blunder
2026-10-14 · Queens Pawn Opening Zukertort Variation · cost 339cp · blunder - Move 10 Bxe4: 18.8x pace and still an inaccuracy
2026-09-30 · Queens Pawn Opening Accelerated London System · cost 65cp · inaccuracy
Shuffling the same piece instead of fixing the position medium confidence
In middlegames without a clear target you move a piece out and straight back, or play the same knight move twice, while the engine keeps asking for the same simple improvement. Each repetition costs half a pawn and they add up.
When it shows up
Middlegames where no tactic is available and you have no plan, typically moves 15 to 30.
What the engine counted
The repeats are visible across four separate games: the same knight move to b4 twice in one game and twice in another, a rook moved to a6 and returned to a8 next move, and in G2 the engine asked for the same pawn push on three different moves and you never played it. These are not one-off errors, they are the same decision made repeatedly, and they are the mechanism behind the 50.1cp and 56.1cp averages in moves 11 to 30.
The moves behind it
- Move 16 Nb4: knight to b4 while the engine wants the rook to e1
2026-09-15 · Queens Pawn Opening Accelerated London System · cost 54cp · inaccuracy - Move 18 Nb4: same knight move again, same rejected rook improvement
2026-09-15 · Queens Pawn Opening Accelerated London System · cost 64cp · inaccuracy - Move 21 Nb4: the same knight move to b4 repeated in another game, -117cp
2026-10-14 · Queens Pawn Opening Zukertort Variation · cost 117cp · inaccuracy - Move 23 Ra6: rook to a6, -299cp
2026-09-18 · Sicilian Defense Taimanov Variation · cost 299cp · mistake - Move 24 Raa8: rook straight back to a8 the next move, another -110cp
2026-09-18 · Sicilian Defense Taimanov Variation · cost 110cp · inaccuracy - Move 27 Ra5: rook wandering to a5 instead of the engine's a6
2026-10-15 · Sicilian Defense Old Sicilian Variation · cost 146cp · inaccuracy - Move 27 cxb4: third move in the game where the engine asked for the same pawn push and you played something else
2026-11-05 · Sicilian Defense Smith Morra Gambit · cost 136cp · inaccuracy
How I should improve
- Reflex moves in the middlegame. Set a floor, not a ceiling: in the middlegame, no move gets played until you have named your opponent's most annoying reply out loud. If you have moved quickly twice in a row, force a pause on the third regardless of how obvious it looks.
- Going slack the moment you are winning. When you reach +2, switch modes deliberately: trade pieces, not pawns, get every piece to a safe square, and ask before each move whether it reduces your opponent's options. Treat a winning position as the point to start spending time, not stop.
- Already worse by move 15 as Black. Pick one answer to 1.e4 and know the first eight moves cold, including what to do when White throws a pawn at you in the Morra. Do not improvise development as Black. The aim is not a trick, it is reaching move 12 near level.
- Long thinks that never get blunder checked. End every long think the same way: name the move, then spend thirty more seconds only on your opponent's checks, captures and threats against it. If you cannot answer all three, pick a different move.
- Shuffling the same piece instead of fixing the position. Before moving, ask which of your pieces is doing the least work and whether a rook belongs on an open or soon to be open file. If you are about to repeat a move you already made, treat that as a signal that you have no plan and go find your worst placed piece instead.
By the numbers
- 6–5–2Win / loss / draw
- 439Moves scored
- 8Blunders
- 20Mistakes
- 72Inaccuracies
- 39Avg loss per move (cp)
| Phase | Moves | Blunders | Mistakes | Inaccuracies | Avg loss (cp) |
|---|---|---|---|---|---|
| middlegame | 242 | 8 | 17 | 54 | 54.8 |
| opening | 156 | 0 | 2 | 18 | 22.7 |
| endgame | 41 | 0 | 1 | 0 | 7.3 |
| Think time | Moves | Blunders | Mistakes | Inaccuracies | Avg loss (cp) |
|---|---|---|---|---|---|
| quick | 47 | 1 | 3 | 7 | 49.6 |
| snap | 133 | 3 | 9 | 26 | 47 |
| long | 162 | 3 | 8 | 27 | 38.5 |
| normal | 97 | 1 | 0 | 12 | 23.6 |
| Color | Moves | Blunders | Mistakes | Inaccuracies | Avg loss (cp) |
|---|---|---|---|---|---|
| black | 212 | 5 | 11 | 41 | 45 |
| white | 227 | 3 | 9 | 31 | 33.3 |
| Result | Moves | Blunders | Mistakes | Inaccuracies | Avg loss (cp) |
|---|---|---|---|---|---|
| loss | 146 | 6 | 8 | 31 | 57.4 |
| draw | 53 | 0 | 2 | 12 | 32.5 |
| win | 240 | 2 | 10 | 29 | 29.2 |
| Moves | Moves | Blunders | Mistakes | Inaccuracies | Avg loss (cp) |
|---|---|---|---|---|---|
| moves 21-30 | 92 | 4 | 5 | 21 | 56.1 |
| moves 11-20 | 126 | 1 | 11 | 27 | 50.1 |
| moves 31-45 | 64 | 3 | 3 | 10 | 48.7 |
| moves 1-10 | 130 | 0 | 1 | 14 | 19.1 |
| moves 46+ | 27 | 0 | 0 | 0 | 1.4 |
| Queens | Moves | Blunders | Mistakes | Inaccuracies | Avg loss (cp) |
|---|---|---|---|---|---|
| queens off the board | 137 | 4 | 6 | 22 | 40 |
| queens on the board | 302 | 4 | 14 | 50 | 38.5 |
| Opening | Moves | Blunders | Mistakes | Inaccuracies | Avg loss (cp) |
|---|---|---|---|---|---|
| Sicilian Defense Taimanov Variation | 34 | 1 | 1 | 9 | 51.6 |
| Sicilian Defense Smith Morra Gambit | 89 | 2 | 5 | 21 | 49 |
| Queens Pawn Opening Zukertort Variation | 78 | 2 | 2 | 13 | 42.5 |
| Queens Pawn Opening Symmetrical Variation | 36 | 0 | 2 | 6 | 32.3 |
| Queens Pawn Opening Zukertort Variation...3.Bg5 Be7 | 22 | 0 | 2 | 2 | 31.1 |
| Sicilian Defense Old Sicilian Variation | 72 | 1 | 3 | 8 | 29.4 |
| Queens Pawn Opening Accelerated London System | 91 | 1 | 3 | 10 | 26.4 |
The games in this report
| Date | Color | Result | Opening | Opponent | Your moves | Blunders | Mistakes | Avg loss (cp) |
|---|---|---|---|---|---|---|---|---|
| 2026-11-12 | white | win | Queens Pawn Opening Zukertort Variation | SabrinamarinA (850) | 38 | 0 | 0 | 29.2 |
| 2026-11-05 | black | loss | Sicilian Defense Smith Morra Gambit | nicjosephs (1251) | 34 | 1 | 2 | 62.4 |
| 2026-10-15 | black | win | Sicilian Defense Old Sicilian Variation | emmanuelfrank (1133) | 72 | 1 | 3 | 29.4 |
| 2026-10-14 | white | loss | Queens Pawn Opening Zukertort Variation | nicjosephs (1236) | 40 | 2 | 2 | 55.2 |
| 2026-10-02 | black | loss | Sicilian Defense Smith Morra Gambit | nicjosephs (1222) | 21 | 1 | 1 | 46.9 |
| 2026-09-30 | white | win | Queens Pawn Opening Accelerated London System | nicjosephs (1207) | 43 | 1 | 2 | 33.6 |
| 2026-09-18 | black | loss | Sicilian Defense Taimanov Variation | nicjosephs (1221) | 34 | 1 | 1 | 51.6 |
| 2026-09-17 | white | win | Queens Pawn Opening Symmetrical Variation | emmanuelfrank (1263) | 36 | 0 | 2 | 32.3 |
| 2026-09-15 | white | draw | Queens Pawn Opening Accelerated London System | nicjosephs (1206) | 19 | 0 | 0 | 24.4 |
| 2026-09-09 | black | loss | Caro Kann Defense Exchange Variation | nicjosephs (1205) | 17 | 1 | 2 | 76.8 |
| 2026-09-06 | white | win | Queens Pawn Opening Zukertort Variation...3.Bg5 Be7 | nicjosephs (1188) | 22 | 0 | 2 | 31.1 |
| 2026-09-03 | black | draw | Sicilian Defense Smith Morra Gambit | nicjosephs (1202) | 34 | 0 | 2 | 37 |
| 2026-08-21 | white | win | Queens Pawn Opening Accelerated London System | nicjosephs (1201) | 29 | 0 | 1 | 17 |
How to read this
Thirteen games and 439 scored moves is enough for the pace and phase patterns, which are large and consistent, but thin for anything opening specific: the Taimanov figure rests on a single game and the Caro Kann game is not in the opening table at all. Two things the aggregates rule out rather than support, and that are worth not worrying about: queens on the board versus off makes essentially no difference to you (38.5cp against 40cp), so the G4 and G2 collapses are about the positions, not the queen trade, and your endgames are genuinely clean (41 moves, zero blunders, zero mistakes, 7.3cp), with the one endgame error in G12 being a single data point. The think time figures are relative to your own median in each game, as the pack states, and the sample mixes daily and rapid, so a long think may partly reflect when you happened to sit down rather than deliberate effort. Evaluations are engine judgements, not practical ones: several positions marked lost were still being defended against opponents around 1200.
- This report covers 13 games. The store holds 1,851 games going back to 2010-06-28, of which 649 have been through the engine. Only daily and rapid games are in scope, and the engine pass runs in bounded batches, so most of the blitz and bullet history has never been scored.
- Every number here is over my own moves only, never my opponents’.
- “Blunder”, “mistake” and “inaccuracy” are fixed centipawn-loss thresholds: 300, 150 and 50 centipawns respectively. A centipawn is a hundredth of a pawn.
- The report itself is regenerated on a schedule; this page is rebuilt from the stored report, so it changes when the analysis does rather than after every game.