«We finished level, but I beat him» is the most common argument about second place. Sometimes it settles the matter and sometimes it counts for nothing. Here is how the measure is computed for two players and for a group, why a Swiss so often makes it unusable, and what to do when the mini-table ties as well.
DE = points scored in the games between the tied players
Of all the tie-breaks this is the only one that counts nothing external: not the strength of the opposition, not when the points arrived, not the rating. It asks a single question - what happened when these players sat down opposite each other. That makes it the easiest to explain and the hardest to argue with.
Two players finish on 4 points and met in round three. Whoever won that game is placed higher. A draw means the measure produced nothing, and the organizer moves to the next item on the list.
Here is where tournament halls get it wrong. The player who beat their rival considers the matter closed - but if the regulations list the direct encounter after the coefficients, those are compared first. It is entirely possible for the loser of that game to finish higher. Nothing unfair has happened: the order was published in advance.
When more than two share a place, their mutual games form a separate small table. Everything else in the tournament ceases to exist: only the points scored inside that group count.
Say Anna, Boris and Vera finish on the same score. Among themselves: Anna beat Boris and drew with Vera, Boris beat Vera.
| Player | Games inside the group | Total | Result |
|---|---|---|---|
| Anna | 1 vs Boris, ½ vs Vera | 1½ | first of the three |
| Boris | 0 vs Anna, 1 vs Vera | 1 | second |
| Vera | ½ vs Anna, 0 vs Boris | ½ | third |
The order comes out Anna, Boris, Vera. Note the oddity: Vera drew with the strongest of the three and lost to Boris, and finishes below him even though her result against the group's winner was better.
The most intuitive tie-break has the narrowest field of application. Three situations in which it stays silent:
Routine in a Swiss: over five rounds a player faces five of thirty. Two leaders may easily never have played each other, and then there is nothing to compare.
If three share a place but only two of them met, the mini-table is incomplete: one has a single game, another has two. Comparing those totals is not sound, so the measure is skipped entirely rather than applied partially.
A win by forfeit does not count by default: the measure reflects a result at the board, not a decision by the arbiter. An organizer who wants it counted has to say so in the regulations.
This is why the direct encounter is almost never the only measure in a set of regulations. In a round-robin it always applies - everyone plays everyone, so any group of tied players yields a complete mini-table. Round-robin schedules are in the Berger tables.
There is no settled answer, and the choice is really about tournament philosophy. Putting the direct encounter first is popular because it needs no explanation: players understand it and there is nothing to dispute. FIDE-style regulations tend to place it after the coefficients, on the grounds that the whole tournament path should be weighed before one individual game.
The argument for first place: one game between the two of them is more honest than indirect arithmetic. The argument against: anything could have happened in that game - a one-move blunder, time trouble, a quick draw - and making it outweigh five other rounds is odd.
A practical suggestion for a club event: coefficient first, direct encounter second. The decision then rests on the whole tournament, but players still have a criterion they can follow. Working sequences for different formats are in the tie-break reference.
To apply the direct encounter you need to know who played whom and how it ended. In the Vibe Chess app that is one tap away: every round keeps its pairings and results, so the question «did we even play each other?» is answered in a second rather than by searching through photos of scoresheets. Coefficients are computed automatically alongside.
It looks at the result of the game between the players who share a place. If two finish level and played each other, the winner of that game is placed higher. It is the most intuitive tie-break there is: the argument is settled by what happened at the board.
You build a mini-table from the games between those players only. Each player's points from those games are added up and the order follows the totals. It applies only if all of them actually met each other; otherwise the comparison is unfair and the measure is skipped.
Then the direct encounter simply does not apply and the organizer moves to the next measure in the regulations. This is routine in a Swiss: players meet only a fraction of the field, and two leaders may never sit down opposite each other.
By default no: the measure looks at games played at the board. The reasoning is that a head-to-head result should reflect a sporting outcome rather than an arbiter's decision. If an organizer wants forfeits included, that has to be stated in the regulations.
There is no single answer. Many national practices put it first because it is the easiest to explain. FIDE-style regulations usually place it after the coefficients: first the whole tournament path is assessed, and only then one individual game. Both are legitimate - what matters is announcing the choice before round one.
Because everyone plays everyone: any two players are guaranteed to have met, and any group of tied players forms a complete mini-table. That is why round-robin regulations almost always list it among the first measures.
Move on to the next measure in the regulations. It happens easily: three players may have drawn among themselves, or beaten each other in a cycle. The direct encounter honestly reports «no answer», and the coefficients take over.
Both games between the tied players count and their points are added. A win and a loss give each player one point and settle nothing; a win and a draw place the winner higher.