The 20-Game Review: How to Audit Your Own Ranked Month

From Doran Ranked Notes — tips from the EloBoostDoran team.

Ask a hardstuck player why they’re stuck and you’ll usually get a story about one game: the AFK, the smurf, the promo their support ran down. Ask a player who climbs steadily and you’ll get something duller — a pattern. The difference isn’t talent. It’s that the second player reviews their games in blocks large enough for the noise to cancel out, and small enough that the findings still describe the player they are today.

Twenty games is that block — roughly a month of ranked for most people, and the smallest sample where your habits, not your luck, start to dominate the numbers. This is the review we run with students before anything else, and you can run it on yourself in about forty minutes, once a month, with nothing but your match history and a notebook.

Why 20 games, and why blocks of five

One game tells you what happened. Twenty games tell you what keeps happening. Inside a single loss, everything looks causal: you died at level 3, so the level-3 death “lost the game.” Across twenty games, you find out whether dying early is a habit or an accident — and habits are the only things worth practising.

Work through the twenty in four blocks of five, oldest to newest. Blocks show direction. If your early deaths cluster in the first block and fade by the fourth, you already fixed that leak mid-month. If they cluster in the last block, something new has crept in — a new champion, a new duo partner, worse sleep, a patch change. Flat across all four? That’s a structural habit, and it’s your headline finding.

One ground rule: you are not looking for your win rate. A 55% month can hide a serious leak that strong teammates papered over, and a 45% month can hide real improvement variance hasn’t paid out yet. You’re auditing behaviour, and behaviour is what you control.

What to track: four numbers and one note

Don’t build a spreadsheet with thirty columns — you’ll abandon it by game six. Track four numbers per game, all readable from a standard match-history page, plus one handwritten note.

1. Deaths before 14 minutes. Not total deaths — early ones. A death at 25 minutes is often a teamfight outcome; a death at 6 minutes is a decision you made alone, with full information about your lane. Early deaths are the purest measure of your lane discipline, and they compound: each one donates gold, experience and map pressure at the exact stage where the game is most snowball-sensitive.
2. CS at 10 minutes. Not as a vanity score — as a consistency check. Compare it against your own average on that champion, not against a pro benchmark. A player who hits 78 on their main in calm games and 61 in chaotic ones isn’t bad at last-hitting; they’re bad at farming while things go wrong, which is a wave-management and recall-timing problem, not a mechanics problem.
3. Your first death’s cause. One short phrase per game: “face-checked river,” “overstayed after kill,” “traded into stacked wave,” “gank, no ward.” Causes, not timestamps. After twenty games you’ll sort these into piles, and the biggest pile is the most honest coaching you’ll ever get for free.
4. Game length of your losses. Short losses and long losses are different diseases. Losses under 25 minutes usually mean the early game collapsed — laning, jungle tracking, first objectives. Losses over 35 minutes usually mean the mid-to-late game leaked — side-lane deaths, thrown barons, bad teamfight positioning. Count how many of your losses fall in each bucket and you’ll know which half of the game is costing you LP.

The one note: after each game, a single sentence — what you’d do differently in the one moment that mattered most. Not “play better.” A moment, a decision, an alternative. Twenty sentences a month is a diary of your decision-making, and its patterns are invisible any other way.

A worked example: Maya’s month

(The figures below are illustrative — one hypothetical player’s month, shown so you can see the shape of the audit. Your numbers will differ; the method won’t.)

Maya is a hypothetical Platinum IV top laner, 11 wins and 9 losses across her twenty games. If she’d only looked at the record, she’d have called it a fine month — slightly positive, climbing slowly. Her audit told a different story.

Her early deaths: 14 across twenty games — but 9 sat in her losses, and 6 of those 9 came in blocks three and four. Direction: getting worse, not better. Her CS at 10 averaged 74 in wins and 66 in losses — until the cause column corrected her reading of that gap. Of her 14 early deaths, the biggest pile, five of them, read the same way: “traded while wave pushing away, jungler unspotted.”

That single sentence reorganised her whole month. Her problem wasn’t farming, aggression, or mechanics. It was this: when her wave was pushing away from her tower — the moment she most wanted to trade — she traded without knowing where the enemy jungler was, and the wave state meant any death cost her two stacked waves, which is why her CS cratered in exactly those games. One leak, three symptoms. The CS gap, the early deaths and half her losses were the same mistake wearing different clothes.

This is what the review is for. Maya didn’t need to “work on everything.” She needed one rule, practised for two weeks.

Finding your one recurring leak

When your twenty games are logged, rank your findings by a simple rule: frequency first, drama never. The leak that appears eight times at small cost beats the catastrophe that happened once. One baron throw is a memory; eight unspotted-jungler deaths is a month of LP.

Three leaks account for most of what we see in these audits, so check for them explicitly:

• The unspotted trade — fighting, pushing or recalling without a current read on the enemy jungler. Shows up as early deaths clustered in one lane phase and one wave state.
• The overstay — staying for one more wave, one more plate, one more camp after the play is over. Shows up as deaths within a minute of a won fight or a taken objective.
• The dead-time death — dying in the 60–90 seconds before a major objective spawns, handing the enemy a free dragon or baron setup. Shows up as mid-game deaths in losses that were even or ahead.

Pick exactly one. Not the three — one. Improvement compounds when practice has a single target, and a month is long enough to genuinely move one habit and far too short to move four. The others go on next month’s list; they’ll wait.

Turning the finding into a two-week focus

A finding without a practice rule is just trivia about yourself. Convert your leak into three things:

A trigger. The exact game situation where the leak happens. Maya’s: “my wave is pushing away and I don’t know where their jungler is.” Triggers must be recognisable in the moment, at game speed — if you can’t spot it live, it’s not a trigger yet, so narrow it until you can.

A rule. One sentence, binary, checkable after the game. Maya’s: “Wave pushing away + jungler unspotted = no trades past the river line until he shows.” Rules should feel slightly over-cautious; a rule you never break is calibrated wrong in the other direction, but a rule you break twice a game is a wish.

A score. For the next ten games, mark each game pass/fail on the rule alone — ignore the result. Ten games of passes while losing is progress the LP graph can’t see yet; ten games of LP gained while breaking the rule is a loan you’ll repay. This is exactly the kind of structured self-coaching we build sessions around in LoL coaching: one leak, one rule, one honest scorecard — because that loop, repeated monthly, is what steady climbers actually do.

After two weeks, run the next 20-game review. The old leak should have shrunk in the cause column. If it hasn’t, the rule was wrong, not your effort — rewrite the rule, narrower, and go again.

The 40-minute monthly checklist

• Pull your last 20 ranked games on your main role. Blocks of five, oldest first.
• For each: early deaths, CS at 10, first-death cause in a phrase, and (for losses) game length.
• Write the one-sentence note per game if you didn’t at the time — from memory of the key moment, honestly labelled as reconstruction.
• Count the cause piles. Circle the biggest. Check the three common leaks explicitly.
• Check direction across blocks: is the leak growing, shrinking, or flat?
• Write the trigger, the rule, and the pass/fail scorecard for your next ten games.
• Schedule next month’s review before you close the notebook.

FAQ

Do I need replay software or a stats site to do this?

No. Everything above — early deaths, CS at 10, game length, your death causes — is visible in standard match history, and the cause column honestly should come from your own memory and notes rather than an automated tool, because the act of naming the cause is where the learning happens. Stats sites can speed up the counting, and watching one or two replays of your worst recurring moment is genuinely useful. But the audit itself is deliberately low-tech: if it needs an evening of setup, you’ll stop doing it monthly, and monthly is the whole point.

What if my champion or role changed during the 20 games?

Split the audit, don’t abandon it. If you swapped mains at game 12, audit the two stretches separately for champion-specific numbers like CS at 10 — those don’t transfer cleanly. But still read the cause column across all twenty, because the interesting leaks — overstays, unspotted trades, dead-time deaths — follow the player, not the champion. In our experience a role swap often makes the review more valuable, not less: whatever survives a change of champion is unmistakably a you habit.

How accurate is a 20-game sample, really?

Honest answer: it’s a compass, not a laboratory. Twenty games won’t tell you your “true” win rate on a champion, and you shouldn’t make big decisions — quitting a role, rebuilding a pool — off one month’s numbers alone. What it does reliably is rank your recurring behaviours: a mistake that shows up six times in twenty games is real, whatever the sample-size caveats. Treat the percentages as approximate, treat the repeated causes as fact, and let two or three monthly reviews in a row — not one — drive the big calls.

The short version

Your last twenty games already contain your coaching plan; you just have to count them. Track early deaths, CS at 10, first-death causes and loss lengths, in blocks of five. Find the pile that repeats, name its trigger, write one rule, and score yourself on the rule — not the LP — for ten games. Do it monthly and the “mystery” of being hardstuck usually turns out to be one or two habits that were never hidden at all.

Doran Ranked Notes publishes practical ranked tips for LoL, Valorant and TFT. We’re the team behind EloBoostDoran (eloboostdoran.com), where we also offer coaching and boosting for players who’d rather not climb alone.