How Many Chess Games Should You Analyze Per Week? (Honest Answer)

Every chess coach on YouTube tells you to analyze every game you play. Every adult improver with a job, kids, and a 9pm Lichess habit knows that’s impossible. So what’s the actual number? After running thousands of chess.com game analyses through MyChessPlan and watching what correlates with real rating gains, the honest answer is: 2 to 3 deep reviews per week, plus one 100-game pattern view per month. Less than that and you don’t move. More than that and analysis crowds out the playing reps you also need. Here’s why that number works, and why “analyze every game” is bad coaching for adults.

The myth of “analyze every game”

The “analyze every game” advice comes from junior chess academies in the 1980s, where 12-year-olds had 4 hours of free time after school and a coach physically present. In that environment, post-game analysis is the whole point — the playing was just raw material for the lesson. The advice was correct in context. It’s mostly nonsense in 2026 for an adult improver playing 30 rapid games a month between meetings.

There are three reasons the advice fails for amateurs. First, attention is finite — analyzing a 50-move game with any care takes 25–45 minutes, and after the third one in a session your retention collapses. Second, most of your games don’t have anything to teach. A 12-move miniature where the opponent hung a queen is not pedagogically interesting; it’s a free rating point. Third, single-game analysis blinds you to the pattern view — the diagnosis of how you lose across many games — which is where the actual improvement leverage lives.

The honest framing: analysis is not a moral duty. It’s a tool. Use enough to extract the lesson, then stop. The opportunity cost of over-analyzing is not playing, not drilling tactics, and not sleeping — all of which compound rating better than your fourth review of the week.

The 50/50 play-vs-analyze rule

The cleanest heuristic comes from sports psychology, not chess: for every hour you spend competing, spend roughly an equal hour reviewing. In a chess context, this maps to: if you play 4 hours of rapid a week, spend 2–4 hours on analysis-adjacent work. That doesn’t all have to be your own games. It includes watching annotated GM games, working through a Chessable course, or studying an endgame book. The “analyze” half is broader than just clicking through Game Review.

For the typical adult improver — 5–10 hours of chess a week — the 50/50 rule decomposes like this:

• 3 hours playing rated rapid or classical (your competitive reps).
• 1 hour deep analysis of 2–3 of your own games (one win, one loss, optionally one critical draw).
• 1 hour skim analysis — clicking through 5–10 of your other games at speed, just spotting turning points.
• 1 hour pattern study — tactics, openings, or endgame drills targeted to your archetype weakness.
• 30–60 min watching annotated games from a player 200–400 points above you (Daniel Naroditsky speedruns are gold for the 1000–1800 band).

That’s the 50/50 rule made concrete. Notice that “deep analysis of your own games” is only one slice — the smallest one. Most adult improvers who feel guilty about not analyzing enough are actually doing roughly the right amount of own-game work and missing the other categories entirely.

Deep vs scan analysis (when each works)

Not all analysis is the same. Two formats do completely different jobs, and conflating them is why most adults feel like their analysis isn’t moving their rating.

Deep analysis: 25-40 minutes per game, 2-3 games per week

Deep analysis is the coach-style review. You write your thoughts before checking the engine. You categorize the loss into tactics, strategy, time, or opening. You find the one move where you’d realistically have spotted the better continuation if you’d been more careful, and you write that takeaway in plain English. The output is a single sentence per game — “I keep trading queens when I’m a pawn up because I want simplification, but the resulting endgame is harder than the middlegame.” That sentence becomes a drill target for the next week.

Deep analysis builds calculation discipline. It rewires the moment of decision at the board. It is irreplaceable, and it is also expensive. The cost is why you cap it at 2–3 games a week. Beyond that, retention drops below the threshold where the lesson actually transfers to your next game. Our GM-style deep analysis guide walks through the candidate-moves method that makes each session worth the time.

Scan analysis: 3-5 minutes per game, 5-15 games per week

Scan analysis is different work. You click through the game once, at speed, looking only for the turning point — the move where the evaluation flipped by more than 0.8 pawns. You don’t read the engine line. You just note when it happened (move 14? move 31?) and what kind of mistake it was (tactical miss, slow positional drift, time pressure).

Scan analysis builds the pattern view. It tells you, after 20 games, that 60% of your decisive mistakes happen between moves 25 and 35 — which means your problem isn’t openings or tactics, it’s middlegame stamina. That’s a diagnosis you can’t get from one deep review, no matter how careful. It’s only visible across the distribution of many games.

The two formats are complementary. Deep analysis without scan analysis means you fix one mistake per week without ever knowing if it’s the most important one. Scan analysis without deep analysis tells you where the wound is but never closes it. You need both.

The 100-game shortcut

The most useful single artefact for an improving adult player is a 100-game pattern report — a structured view of your last 100 rated games with phase-by-phase centipawn loss, opening-exit evaluation distribution, time-trouble flag rate, conversion rate from winning positions, and a frequency count of recurring tactical and positional motifs you miss. This is what a $400 coaching package builds for you over four weeks of homework. It’s also what MyChessPlan generates in 60 seconds.

The 100-game view is not a replacement for deep weekly analysis — it’s the input that tells you which weekly analyses to do. If your report says you exit the opening at -0.8 in 55% of games and your conversion rate from +2 positions is 38%, the deep review priorities for the next month write themselves: one opening repair session, two endgame conversion sessions, and the rest of your time goes to playing and tactics. Without the 100-game view, you guess.

A reasonable cadence for the pattern view: re-run it once a month, or any time your rating moves 100+ points in either direction. Your weakness profile shifts as your skill changes — the Time-Pressured 1200 may be an Opening-Confused 1450 four months later, and training to fix the old wound while ignoring the new one is the most common reason “analysis” stops paying off.

A realistic weekly schedule by life stage

The 50/50 rule scales — but only honestly if you scale it to the time you actually have, not the time you wish you had. Three realistic schedules below.

The busy adult (5 hours/week total)

1. 2 sessions of rapid play (3 games each, ~90 minutes total).
2. 1 deep analysis on Sunday morning (45 min, your worst loss of the week).
3. 2 quick scans mid-week (5 min each, just spotting turning points).
4. 1 monthly 100-game refresh (60 min, replaces one deep session that week).
5. 30 min tactics spread across the week on phone (Chess Tempo or Lichess Puzzle Storm).

The committed improver (8-10 hours/week)

1. 3-4 sessions of rapid or classical (~4 hours).
2. 2 deep analyses per week (one win, one loss; ~75 min total).
3. 1 hour of scan review across 8-12 other games (mid-week, 5-7 min each).
4. 90 min targeted drilling on the weakness flagged by your 100-game report.
5. 1 hour annotated GM-game watching (Naroditsky, Hess, or Aman speedruns at your band).

The serious adult improver (12+ hours/week)

1. 5+ hours play, mostly classical with one rapid session.
2. 3 deep analyses (cap at 3 — beyond that retention drops).
3. 2 hours scan analysis.
4. 3-4 hours opening + endgame study, structured around your archetype.
5. Bi-weekly 100-game refresh instead of monthly.

The schedules above assume you actually want to climb. If your goal is recreational chess — and that’s a perfectly valid goal — analyze whatever feels fun and ignore the schedule. The sport doesn’t owe you a structured improvement curve unless you’re chasing one.

Common mistakes when deciding how much to analyze

Five mistakes I see repeatedly in MyChessPlan user threads:

1. Analyzing every game superficially. Five 5-minute reviews extract less learning than two 30-minute reviews. Depth beats breadth on the deep-analysis side.
2. Skipping wins. Wins teach what works under pressure. A win where you held +0.3 the whole game is more informative than a loss where you were down a piece by move 12. Aim for a 1:1 wins-to-losses ratio in deep review.
3. Analyzing only with the engine on. If you read the eval before forming your opinion, you learn nothing — your brain rationalizes the engine line as obvious. Always write your thoughts first, then check.
4. Ignoring scan analysis. Scanning 10 games for turning points is half an hour and produces the pattern view that tells you which deep reviews matter. Most adults skip it because it feels less “serious.” That’s exactly backwards.
5. Re-running the 100-game view too often. Once a month is plenty. Re-running it weekly because your rating moved 30 points is noise-chasing — Glicko swings 50 points in a single session for many adult players.

Frequently asked questions

Should I analyze blitz and bullet games too?

Generally no — at least not deeply. Blitz and bullet games are dominated by time pressure and pre-moves, so the “mistake” you’re analyzing is usually just clock collapse, not a real chess decision. Use them for opening repertoire reps and pattern recognition, not for deep review. The exception: scan-analyze your blitz games to spot opening lines where your win rate as Black drops below 40% — those are repertoire holes worth fixing.

Is chess.com Game Review enough or do I need a real engine?

For 99% of improvers under 2000, chess.com Game Review (or Lichess analysis) is more than enough engine power. Stockfish at depth 22 already plays at roughly 3500 strength — you do not need depth 40 to find your blunders. The bottleneck for amateurs isn’t engine accuracy; it’s the human work of writing a clear takeaway sentence and turning it into a drill. Our chess.com analysis guide covers the workflow.

How long until 2-3 deep reviews per week shows up in my rating?

Most adult improvers who run the 50/50 schedule honestly for 8–12 weeks gain 80–150 rating points, with the bulk arriving in weeks 6–10 (the lag between learning a pattern and reliably executing it under clock pressure is real). If you’re seeing nothing after 12 weeks, the problem usually isn’t the analysis volume — it’s that you’re analyzing the wrong games. Re-run a 100-game pattern view to recalibrate which weakness to attack next.

What if I only have 1 hour of chess time per week?

Don’t analyze. Spend the hour playing one rapid game and one tactics session. At that volume, analysis has negative return — you’ll forget the lesson before next week’s session anyway. Save the analytical work for life stages where you have at least 3–4 hours of total chess time. There’s no shame in just enjoying the game.

Does the 50/50 rule apply to over-the-board (OTB) tournament players?

Yes, but with a twist. OTB games are richer per-game (longer time controls, higher stakes, more memorable moments) so you can extract more from each one. A reasonable OTB schedule: deep-analyze every classical game you play, scan-analyze your rapid sidegames, and run a 100-game online pattern view monthly to keep your archetype diagnosis current. The 5 archetypes guide covers how OTB and online weakness profiles often diverge.

The honest answer, summarized

Two to three deep reviews per week, plus a monthly 100-game pattern view, plus 5–10 quick scans of your other games. Less than that and you stagnate. More than that and the analysis crowds out playing, drilling, and resting — all of which compound rating in ways analysis can’t. The “analyze every game” advice is junior-academy folk wisdom that doesn’t survive contact with adult time constraints. Pick the schedule that fits your actual life, run it for 8–12 weeks, and trust the process.

If you want the 100-game pattern view without the manual scan work, that’s exactly what we built MyChessPlan to deliver. Run yours here. And if your deep weekly reviews keep flagging the same losing patterns, the recurring-pattern guide is probably your next read.