In the master game database, 1.e4 has been played more than one million times. Every reply Black has ever tried — from 1…e5 to the Sicilian to the Alekhine — is catalogued. Every continuation from every one of those positions is indexed, scored, and waiting for you to click.

That is what an opening explorer does. It turns a century of grandmaster practice into a navigable tree where every node shows you the statistics: how often each move was played, what percentage of those games White won, what percentage ended in a draw, and links to the actual games so you can read how each position resolved. The move a world champion trusted is one click away from the move that caused them to resign.

This guide explains how to read the tree, why master game statistics matter more than any other dataset, and how to use the explorer to build an opening repertoire that is grounded in real evidence rather than memorised theory.

What Does an Opening Explorer Show You?

An opening explorer shows every move played from the current position in the selected database, along with win/draw/loss percentages and a list of model games. Make a move on the board and the right panel refreshes: a new list of continuations, new statistics, and new example games — the next layer of the tree, one position deeper.

The explorer is not a recommendation engine. It does not tell you which move is best. What it tells you is which moves have been trusted by players at a given level, and with what results. That distinction matters. A move that scores 55% for White in amateur games might score only 38% in master games, because at the top level Black knows the refutation. The statistics change at every rating threshold, and master games are the cleanest signal.

Most modern explorers give you several database choices:

  • Master games — games played by titled players (typically 2200+), sourced from over-the-board tournaments going back to the 1950s. Lichess curates roughly two million such games; Chess.com’s explorer draws on nearly three million. This is the dataset most useful for improving.
  • Lichess games — every game played on Lichess, filterable by rating range and time control. Enormous (hundreds of millions of games), but noisier. Useful for finding what people at your exact rating actually play.
  • Your own games — in PGN Base, you can filter the tree by your own imported collection and see exactly where you diverge from master practice in your pet lines.

Below each move’s statistics, the explorer surfaces notable games from the current position. These are the model games — the ones worth clicking into, reading, and replaying. A single master game shows you what the statistics cannot: the pawn structures, the plans, the moments where one side went wrong.

How to Read the Move Statistics

Each move in the explorer shows three percentages: White wins, draws, and Black wins. Reading them correctly takes a moment to learn.

A balanced, mainline position typically shows something like 37–36–27 — meaning White wins slightly more than Black (which reflects the first-move advantage), with a healthy draw rate in between. If you see 50–10–40, you are looking at a sharp gambit line where the game is almost always decisive and results vary wildly. If you see 30–55–15, you are in theoretical endgame territory where draws dominate.

The raw game count matters as much as the percentages. A move played 40,000 times with 38% White wins is a proven mainline. A move played twelve times with 100% White wins is a curiosity — twelve games is not enough to conclude anything. Always weight the statistics by volume.

One detail that surprises new users: the statistics are not engine evaluations. A position where Black wins 45% of master games is not necessarily “equal” — it might simply be one where both sides have plentiful winning chances. Conversely, a position the engine calls +0.8 might show 40% draws in master games because it requires technique most players lack. The explorer shows you results, not evaluations. Those are different lenses on the same position, and you need both.

Why Master Games Are the Right Dataset

Master game statistics strip away the noise that distorts amateur databases. In amateur games, almost any move gets played — including the ones that lose a piece on move eight. Including those games in the statistics would tell you what blunders are popular, not what works. Master games eliminate that problem.

Lichess’s master database covers roughly two million games played in over-the-board tournaments by rated players, sourced from FIDE events, national championships, and elite invitationals dating back to 1952. The database has been updated with around 200,000 new games since 2022. These are not online blitz games — they are tournament games where players had time to think, books to prepare, and reputations on the line.

The consequence is that patterns in master statistics are durable. A line that scores poorly for Black in a hundred thousand master games has been tested by the best-prepared players in the world and found wanting. That is worth more than any single recommendation. It is the aggregate verdict of the strongest players in history.

There is a second reason to use master games specifically: they show you the consequences of opening choices in positions you will actually play, not just in the opening itself. Many amateur games are decided by tactics in move 5–10. Master games routinely last past move 30. That means the master database shows you what happens after the opening ends — the middlegame structures and endgames that flow from each choice, across thousands of examples.

How Kramnik Used the Database to Topple Kasparov

Vladimir Kramnik’s 2000 World Championship match against Garry Kasparov is the clearest historical demonstration of what an opening explorer can do. Kasparov had dominated the Ruy Lopez for years, relying on the Morphy Defense (3…a6) to steer into complex middlegames where his calculation was unmatched. Kramnik and his preparation team studied the master database and found something that had been sitting in plain sight.

The Berlin Defense — 3…Nf6, following 1.e4 e5 2.Nf3 Nc6 3.Bb5 — led to a near-immediate queen trade after 4.O-O Nxe4 5.d4 Nd6 6.Bxc6 dxc6 7.dxe5 Nf5 8.Qxd8+ Kxd8. The statistics told the story: in master games, the Berlin endgame scored far better for Black than any other Ruy Lopez reply. The position was technically equal, but the simplified structure neutralised Kasparov’s attacking genius almost completely.

Kasparov had played 1.e4 his entire career. He had no prepared response. Kramnik played the Berlin in every game where Kasparov opened with 1.e4 and won the match without losing a single game. The Berlin Defence became known as “the Berlin Wall.”

What Kramnik did with a preparation team and specialist software, any player can now do in an afternoon with a free opening explorer. The Ruy Lopez is a good place to start: navigate past 1.e4 e5 2.Nf3 Nc6 3.Bb5 and compare the statistics for the Berlin, the Morphy, and the Marshall Attack side by side. The draw rate for the Berlin at master level — around 43%, versus 36% for the Morphy Defense — is exactly the signal Kramnik was looking for.

A Walk Through the Tree — The Italian Game and What the Statistics Show

Start from the initial position in any opening explorer. Click 1.e4. The explorer immediately tells you it is the most-played first move in master chess by a wide margin: in the Lichess master database alone, it has been played over a million times.

After 1.e4 e5, click 2.Nf3. The explorer narrows the field. Of the responses to 2.Nf3, one move dwarfs all others: 2…Nc6, played in the vast majority of master games. This is the first real lesson the explorer teaches — not all legal moves are equally viable, and the tree shows you which ones have actually been trusted.

After 1.e4 e5 2.Nf3 Nc6, White has two primary tries in modern master play: 3.Bb5 (the Ruy Lopez) and 3.Bc4 (the Italian Game). For much of chess history the Ruy Lopez dominated. Today the explorer shows something different — the Italian has surged in master practice as engine analysis revealed its long-term strategic potential. The statistics for both are close, but the Italian’s move count is rising every year.

The game below is from Hastings 1895 — one of the most famous tournaments ever played — and starts with 1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5: the Italian Game, Giuoco Piano. Wilhelm Steinitz plays the c3+d4 central break that the explorer would surface as the most statistically popular White plan, while Carl von Bardeleben defends inaccurately and is overwhelmed. Steinitz builds such a dominant position that von Bardeleben eventually walked away from the board rather than face the announced forced mate. The game is given as White wins after 25.Rxh7+, from which Steinitz had a forced checkmate in ten moves.

Step through every move using the board below. The Italian opening phase — moves 1 through 8 — corresponds exactly to the tree paths you would navigate in the master database.

8
7
6
5
4
3
2
a1
b
c
d
e
f
g
h
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
Start

After 8.exd5 Nxd5 9.O-O, the explorer would show you this exact position — “Steinitz vs Bardeleben, Hastings 1895” appearing in the model games panel. That is the value of the feature: the position on your board connects directly to the history of the game, and the history is playable.

How to Build Your Repertoire With the Opening Explorer

The opening explorer is most useful not as a memorisation tool but as a filtering tool. Here is the method that works.

Find your opponent’s last book move. Navigate the tree move by move through your planned repertoire. The statistics at each step tell you how normal each position is. If you reach a position where only 200 master games exist, you are already in territory where you are on your own.

Compare the two or three most-played replies. At each choice point in your repertoire, look at the statistics for your candidate moves. Do not just compare White win percentages — look at draw rates too. A line with 35% White wins and 45% draws may be more practical for a tournament player than a sharp line with 50% White wins and 10% draws, depending on the score you need.

Open three model games from each candidate. Statistics aggregate across thousands of different players and preparation levels. The model games show you how the position was actually played — the plans, the pawn breaks, the typical tactical motifs. Three or four games give you a feel for the structure that a number on a row cannot convey.

Check your own collection against the master tree. In PGN Base, you can filter the explorer by your own imported games. This shows you where your games diverge from master practice. If the master database shows 70% of games continue with a particular move and your games never do, you have identified a gap in your preparation — or an opportunity to find a sideline.

The Sicilian Defense is a useful case study: the master database shows the Najdorf as the most popular choice after 1.e4 c5 2.Nf3 d6, but the statistics for the Taimanov and the Kan are competitive. A player building a Sicilian repertoire for the first time can use the explorer to make an evidence-based choice rather than defaulting to whatever they first read about.

What PGN Base’s Opening Explorer Adds

Most opening explorers show you master data for the global position. PGN Base’s explorer connects that master data to your own game collection.

When you import your games, PGN Base indexes every position in your collection. In the opening explorer, you can toggle to “my games” view and see exactly how your games have continued from any position on the board — alongside the master statistics for the same position. The gap between what you play and what masters play is visible in a single panel.

This matters for improvement because theory is only useful at the moves you actually reach. If you play the London System and your games show that you always reach a particular position after move ten, the master statistics for that exact position tell you more about your preparation than any chapter of opening theory aimed at a general audience. The explorer narrows the focus from “all chess” to “your chess.”

You can also use the model games feature as a study tool. Find a position where you have lost several times, navigate to it in the explorer, and open the master games from that position. You will almost always find that one side had a specific plan or resource that you had missed — something visible in a game but not in a tree of moves.

Import your first game collection into PGN Base and open the explorer on the positions where your results have been weakest. The master database will show you exactly where your preparation ends and where work needs to start.

What Makes the Opening Explorer a Different Kind of Study Tool

A book chapter on an opening tells you what the author thought in the year it was written. An engine line tells you what the engine calculates from a single position. The opening explorer tells you what thousands of the strongest players in the world actually played, across decades of competition, with results attached.

Those are three different kinds of knowledge. Books provide narrative and explanation. Engines provide tactical precision. The explorer provides empirical evidence — and empirical evidence does not go out of date in the same way that specific lines do. A line that scored badly for Black in the master database in 1985 almost certainly still scores badly, because the structural reasons for its weakness do not change when a new opening preparation tool is released.

Kramnik did not use an engine to find the Berlin Defense. He used the master game statistics — the equivalent of what the opening explorer shows you today. He saw that Black’s results in the Berlin endgame were better than in any other Ruy Lopez reply, trusted that signal over theoretical fashion, and prepared accordingly. The world championship was decided, in significant part, by reading the tree.

That same signal is available to any player with an opening explorer and an hour of preparation. The grandmaster advantage in opening preparation has never been smaller. The gap between knowing the tree and not knowing it has never been larger.

For a closer look at how chess engines evaluate the positions you reach in these lines, see how engine evaluations work — the eval bar’s numbers mean something specific, and knowing what they mean changes how you read explorer statistics. For game review after your games, PGN Base’s game review tool connects the opening statistics to your actual results.

Conclusion

An opening explorer is not a shortcut to memorising more theory. It is a tool for understanding which theory matters — for you, against your opponents, in your lines — and why. The master database shows you what has worked across a century of high-level play, filtered down to the exact position on your board.

The Berlin Defense became one of the most played openings at the top level because one player read the statistics carefully and followed the evidence. The statistics have not changed. They are waiting in the explorer, available to anyone who wants to look.