JaVaFo: the pairing engine behind the programs

A pairing engine: a piece of software that takes the state of a tournament - who has played whom, with which colours, on how many points - and returns the pairings for the next round according to the FIDE Swiss rules.

• It implements the Dutch system. The default Swiss pairing rules used at the overwhelming majority of open events.
• It is certified. Which means its output has been checked against the rules by FIDE's pairings body - the reason organizers can rely on it rather than argue about it.
• It has no interface worth speaking of. Data goes in as a file, pairings come out as a file. Everything a human sees is provided by the program around it.
• It is written in Java, hence the name, and it runs anywhere Java runs.

Inside other software, essentially always. An organizer encounters JaVaFo without ever seeing it:

• In tournament programs. Desktop software calls a pairing engine rather than implementing the rules itself - the rules are long and the certification is the point.
• In online platforms. Swiss events on chess servers need exactly the same logic.
• In federation systems. Wherever pairings must be defensible after the fact.
• As a checking tool. Arbiters occasionally run an engine independently to verify a disputed pairing.

The data format they exchange is TRF, the FIDE standard tournament file - which is why tournament data moves between programs at all.

These are the two engines an organizer is likely to hear about, and they solve the same problem in different ways.

Because both implement the same published rules, their pairings agree in the overwhelming majority of positions. Where they differ, it is in situations the rules leave open to more than one legal solution - and either answer is correct.

Being open source, bbpPairings is the usual choice for anything self-hosted, which is why it is the engine behind the pairings in our own tools.

You will never install JaVaFo. What is worth taking away is the layer it represents.

• Pairings are not a feature of a program. They come from a certified engine, which is why two different programs produce the same pairs from the same data.
• Hand-made pairings are a different thing entirely. The Dutch rules are long, and any simplified version - «pair the top half against the bottom half and avoid rematches» - drifts from them within two rounds - how a Swiss is structured.
• A disputed pairing can be checked. If a player insists the pairing is wrong, an engine can be run against the same data - the argument has an answer.
• Tournament data is portable. Because of the shared TRF format, an event can move between tools instead of being locked into one.

The practical consequence for a club: use something that calls a real engine rather than pairing by hand, whatever the size of the event - running an event in the app.