A round-robin tie-break with an unusually simple idea behind it: count only what you scored against opponents who themselves finished on 50 percent or better, and ignore everything harvested from the bottom of the table. Here is the formula, how the threshold is set, two worked examples with the arithmetic shown, and the case where Koya and Sonneborn-Berger disagree about who finished second.
Koya = points scored against opponents who finished on 50% or more
The reasoning is straightforward. Points taken off the tail of the field are collected by everybody in roughly equal measure, and no tournament race is decided there. The real difference between the contenders shows up in the games against the solid half - and those are the only games this measure counts.
Unlike Sonneborn-Berger, it does not matter which member of the solid half you beat: a point against the winner and a point against the weakest player still inside the group weigh exactly the same. The measure is blunter, but it is easy to explain to a player: this is what you took off the strong half.
Like every tie-break, Koya is compared only between players on equal points. A high value on its own does not move anyone up the table.
The threshold is derived from the maximum score obtainable in the event, which equals the number of rounds. Half of that is the boundary: everyone who finished on it or above is inside the set of counted opponents.
Your own result is not part of the sum: only the games against other players count. Being in the solid half yourself adds nothing - what matters is how much you took off the rest of that group.
The same club round-robin used to explain the Sonneborn-Berger score, where you can also replay the whole crosstable yourself. The threshold is 2½, which puts four players in the solid half. Three players share second place on 2½ - those are the ones to look at.
| Player | Score | SB | Koya | Solid half |
|---|---|---|---|---|
| Paul Grigoriev | 4 | 9 | 2½ | yes |
| Oleg Tarasov | 2½ | 6½ | 2 | yes |
| Daria Belova | 2½ | 5¼ | 1 | yes |
| Nikita Orlov | 2½ | 4¾ | ½ | yes |
| Maria Savelieva | 2 | 4 | 1 | no |
| Ivan Krylov | 1½ | 4½ | 1½ | no |
½ against Paul + 1 against Daria + ½ against Nikita
0 against Paul + 0 against Oleg + 1 against Nikita
0 against Paul + ½ against Oleg + 0 against Daria
Here Koya ordered the contenders exactly as Sonneborn-Berger did: Oleg first, then Daria, then Nikita. That happens often and it is entirely normal - both measures are asking about play against strong opposition, just in different ways. But agreement is not guaranteed, which is what the next example is about.
A different event of the same size. Roman and Timur share second place on 2½ apiece. Sonneborn-Berger puts Roman higher, Koya puts Timur higher. Here is where the difference comes from.
| Player | Score | SB | Koya | Solid half |
|---|---|---|---|---|
| Alina Egorova | 3½ | 7¾ | 1½ | yes |
| Victor Panov | 3½ | 7½ | 2 | yes |
| Roman Kovalev | 2½ | 6 | 1 | yes |
| Timur Aslanov | 2½ | 5¼ | 1½ | yes |
| Gleb Morozov | 2 | 3½ | 1 | no |
| Yulia Pakhomova | 1 | 3 | 1 | no |
Wins over Alina (3½) and Gleb (2)
Draw with Yulia (1)
Wins over Roman (2½) and Yulia (1)
Draw with Alina (3½)
Roman built his score on a win over the tournament leader and a win over a tail-ender, while losing to the other two members of the solid half. Sonneborn-Berger does not notice that: it sums the scores of the players you beat, and one big scalp outweighs the rest.
Koya looks only at the games inside the solid half, and there the picture reverses: Timur has a point and a half against Roman's one. On top of that Timur beat Roman in their individual game, so the direct encounter favours him too.
The practical conclusion: two of the three measures put Timur higher and one puts Roman higher. Who actually finishes second depends purely on the order the measures are listed in the regulations. Which is exactly why that list has to be published before round one.
The 50 percent boundary is the default, not an article of faith. The FIDE regulations provide a modifier that moves it in half-point steps, up or down. Raising the threshold narrows the counted set to stronger opponents; lowering it widens the set by pulling in the next group down the table.
What it is for: the base measure is fairly coarse, and in short events several players routinely end up with identical values. Shifting the threshold resolves that residual tie without moving on to the next measure entirely. If you use a shift, it must be named in the regulations alongside the measure - otherwise players cannot check the arithmetic themselves.
The measure lives in the FIDE Play-Off and Tie-Break Regulations, and it is easy to miss because it sits in the same article group as Sonneborn-Berger rather than under a heading of its own.
The measure appears in article 9.2 as «Koya System for Round Robin» with the abbreviation KS, alongside Sonneborn-Berger. The threshold shift is separated out into article 14.5, which covers modifiers generally.
The regulations were renumbered in the 2023 and 2026 revisions, so older articles, forum posts and even some national federation documents cite article numbers that no longer match the current text. Always check the number against the edition you are actually working from, and cite the edition date in your own regulations.
Buchholz has Bruno Buchholz and the 1932 event in Bitterfeld. Sonneborn-Berger has Gelbfuhs, Sonneborn and Berger, and a documented first use in Liverpool. Koya has nothing: the available sources do not record who proposed it, when, or where the name comes from. In the regulations it simply appears as the Koya System, with no explanation attached.
This is worth saying plainly rather than inventing a plausible story, which is what several sites do. If you find a documented origin, it is genuinely new information.
Count the rounds and halve it. Five rounds means the threshold is 2½.
Go down the final standings and tick everyone on the threshold or above. That is the set of opponents that will count.
Take only your games against the ticked players and sum what you scored: 1 for a win, ½ for a draw, nothing for a loss. That total is your Koya.
For the other round-robin measures, the Sonneborn-Berger calculator does the arithmetic from a filled-in crosstable, the Berger tables give you the round schedule, and the tie-break reference covers the whole set with the order to apply them in.
In the Vibe Chess app you pick the round-robin format, add the players including guests without Telegram, and the schedule is built from the field automatically. Results go in with a button press, the standings order players level on points by the tie-break that suits the format, and the finished event exports as a TRF report.
It is a tie-break for round-robin events. Only the points scored against opponents who themselves finished on 50 percent or better are counted; games against the tail of the field do not enter the sum at all. It answers one question: how did this player do against the solid half of the table?
Take the maximum score obtainable in the event, which is the number of rounds, and halve it. A six-player round-robin runs to five rounds, so the threshold is 2½ points: everyone on that or more is inside the set of counted opponents.
Sonneborn-Berger weights each result by the opponent's score, so beating the winner is worth more than beating a mid-table player. Koya is blunter but more direct: an opponent is either in the solid half or not, and inside that group everyone counts the same. That is exactly why the two measures sometimes order players differently.
That cannot happen in a round-robin: the total number of points is fixed, so somebody always finishes at or above half. What is entirely realistic is several players ending up with a Koya of zero, which simply means they took nothing off the solid half.
The measure is defined for round-robins - it says so in its own name in the regulations. It does not appear in the recommended Swiss sequences, where Buchholz does the job. There is no explicit prohibition in the text, so an organizer could in principle write it into a Swiss regulation, but then the calculation method needs spelling out in detail.
Yes. Games not played at the board are treated as decisive for the purpose of this measure, and a player left without an opponent is treated as having won. Since March 2026 the dummy opponent that stands in for such a round carries a capped score.
Usually after Sonneborn-Berger. The logic is to run the finer measure first, the one that distinguishes between opponents by strength, and to fall back on the blunter question - what did you take off the top half - only when the first one ties.
Yes. The regulations provide a modifier that shifts the boundary up or down in half-point steps, narrowing the counted set to the strongest opponents or widening it to include the next group down. If you use a shift, it has to be stated in the regulations alongside the measure itself.
Nobody seems to know. Unlike Buchholz and Sonneborn-Berger, whose authors and origins are documented, the available sources do not explain where the name came from. In the regulations it simply appears as the Koya System.
Yes, arbiter software and results services can display it if the organizer enables the measure. It is also easy by hand: mark everyone who finished on half or better, then add up your own points against those players.