Most adult improvers analyze their games backwards. They lose, they feel annoyed, they paste the game into an engine, they watch a red bar spike on move 23, they mutter “should’ve seen that,” and they close the tab. Twenty minutes spent, almost nothing learned. The engine told them where the game went wrong but not why their thinking went wrong — and the “why” is the only part that transfers to your next game.
Game analysis is the single highest-leverage habit available to a working adult with limited study time. You already paid the tuition by playing the game; review is where you collect on it. But the tools have quietly gotten so strong that they now do your thinking for you, which feels productive and teaches you remarkably little. This guide walks through the major analysis tools, what each is actually good for, and a four-pass method that keeps you — not Stockfish — in the driver’s seat.
Why engine-first review fails adult improvers
An engine evaluates positions; it does not explain decisions. When Stockfish says a move loses 1.8 pawns of advantage, that number is true and almost useless on its own. You didn’t blunder because you lacked a 3500-rated calculator in your head. You blundered because you stopped checking your opponent’s threats, or you had no plan and grabbed a pawn, or you were short on the clock and moved on feel.
Those are thinking-process errors, and an engine is blind to them. It will happily show you a brilliant ten-move line you could never have found over the board, which is entertaining and trains nothing. The skill that actually raises your rating is catching your own recurring decision failures — and you can only catch those if you do the thinking before you turn the engine on. This is the same reason a high puzzle rating often fails to show up in real games: the format trained pattern recall in isolation, not decision-making under live conditions.
The tools, and what each one is genuinely for
There is no single best tool — there are tools that fit different jobs. Here is how the main options actually stack up for someone reviewing their own games.
Lichess Analysis Board
Free, fast, and the most under-used training environment online. The Analysis Board lets you walk moves without the engine on, write text comments on positions, and only then toggle a local Stockfish. Its “Learn from your mistakes” feature replays your inaccuracies as puzzles you must solve yourself — the closest any free tool comes to teaching the why. For most adult improvers this is the workhorse.
Chess.com Game Review
The most polished experience: move classifications (Brilliant, Blunder, Miss), a “key moments” digest, accuracy scores, and a coach-style narration. The polish is also its trap. The classification labels feel like understanding but are really just an engine delta with a friendly sticker. Excellent for spotting where things broke and surfacing your most critical moment; weak at making you reconstruct your own reasoning. Full reviews require a paid membership.
Stockfish (standalone, via a GUI)
The raw engine everything else is built on. Running it yourself in a desktop GUI gives you depth control, multi-line (MultiPV) output, and zero hand-holding. This is overkill for casual review but valuable when you want to interrogate one critical position deeply — set it to show three candidate lines and compare the plans, not just the numbers. If you ever wondered what “accuracy” and “centipawn loss” really measure, it is worth understanding what those metrics actually tell you and where they mislead before you anchor on them.
A plain board and a notebook
Not a joke, and not nostalgia. Reviewing a game on a physical board with the engine closed forces the recall and visualization that build over-the-board strength. Many coaches insist on at least one engine-free pass for exactly this reason.
| Tool | Best job | Cost | Risk |
|---|---|---|---|
| Lichess Analysis | Full self-review + replay-your-mistakes | Free | Plain UI, easy to over-rely on engine toggle |
| Chess.com Review | Fast “where did it break” + key moment | Paid (full) | Labels feel like insight; they aren’t |
| Stockfish GUI | Deep dive on one critical position | Free | Overkill; trains nothing if used passively |
| Board + notebook | Building visualization and recall | Free | Slow; no error-checking safety net |
The four-pass method: think first, verify second
The order matters more than the tool. This method works in any of the environments above and takes 15–25 minutes per game. The principle is simple: the engine stays off until pass three.
Pass 1 — Replay your emotions (2 minutes)
Click through the game once and mark — out loud or in a comment — every point where you felt something: confident, confused, rushed, surprised. Emotion is a reliable flag for the moments where your decision-making was actually tested. You are building a list of positions to investigate, not judgments yet.
Pass 2 — Re-decide, engine off (8–12 minutes)
Return to each flagged position and genuinely re-solve it as if it were a puzzle. What were the candidate moves? What was your opponent threatening? What was your plan? Write a one-sentence answer for each. This is the pass that does the work, and it is the one almost everyone skips. It mirrors the candidate-move process you should be using over the board — review is just that same process, slowed down with no clock pressure.
Pass 3 — Verify with the engine (3–5 minutes)
Now turn it on. For each flagged position, compare what you concluded to what the engine says. You are not looking at the eval bar across all 40 moves — you are checking your reasoning at the three or four moments that mattered. When the engine disagrees, ask which of your assumptions was wrong, not just which square was better.
Pass 4 — Write one transferable lesson (2 minutes)
Close the game with a single sentence you could apply next week: “I stop checking my opponent’s threats once I have an attack,” or “I trade into endgames I haven’t actually calculated.” One sentence, logged somewhere you will re-read it. A stack of these sentences becomes the most valuable study document you own.
How many games, and which ones
You do not need to analyze everything. Review your losses and your hardest draws — wins teach less because nothing forced you to think. Two or three deep reviews a week beats skimming twenty games with the engine narrating. Depth compounds; breadth evaporates. If you are unsure how to fit this into a realistic weekly routine, it slots naturally into a study plan built around your actual available time rather than an idealized one.
Where the patterns live
Here is the payoff that no single tool gives you: the same lesson sentence will start showing up again and again. Three weeks of pass-four notes and you will see it — you don’t have a thousand random weaknesses, you have three or four recurring decision failures that cost you most of your points. That recurring shape is your archetype, and naming it is what turns scattered review into targeted improvement.
That is exactly what MyChessPlan was built to surface. Take the free archetype report to identify your dominant decision pattern in a few minutes, and if you want a structured plan that turns it into weekly training, the $14.99 premium plan maps your archetype to specific drills and a review routine. The tools above tell you what happened in one game; the goal is to learn what keeps happening across all of them.
The bottom line
The best analysis tool is the one used in the right order. Lichess for most reviews, Chess.com for a fast critical-moment scan, a standalone engine when one position deserves a deep look — but in every case, your own thinking comes first and the engine only verifies. Do that consistently and game review stops being a chore you skim and becomes the most efficient rating-builder you have.