Swiss Pairing Rules - Half point byes

Whether it is legal to give half point byes in a FIDE Rated Swiss tournament has been subject to debate for quite a while. Some pairing programs do allow it (even FIDE approved ones) but until recently the pairing rules seemed to be clear that it wasn't allowed. Under the rules for the Dutch System clause F5 stated

Players who withdraw from the tournament will no longer be paired. Players known in advance not to play in a particular round are nor paired in that round and score 0.


However the last meeting of the FIDE Swiss Pairing Programs Committee, as well as the FIDE Technical Commission had a look at the rules and have now ruled that they are allowable.
In the updated Swiss Pairing Rules clause B1b now states

A player who has received a point or half point without playing, either through a bye or due to an opponent not appearing in time, is a downfloater (see A.4) and shall not receive a bye.


The text of the updates Swiss Pairing Rules is at this link FIDE Swiss Pairing Rules (Dutch System)
(Note: Clause F5 still remains unchanged in the updated rules, but in discussion with SPP Chairman Christian Krause, we agreed that this is overridden by B1b and F5 will be changed at the next meeting).

The discussions concerning this, and other changes to the pairing rules are included in the minutes of the Swiss Pairing Programs Committee.

However the final decision of half point byes was made by the Technical Commission, who's meeting minutes are here.

Nonetheless there may be some time until all FIDE approved pairing programs are updated to reflect the changes in the pairing rules. At the 2012 Queenstown International I observed a number of cases where players who had received a half point bye were then downfloated in the next round, or the round after by Swiss Manager. on the other hand, JaVaFo, which we ran alongside Swiss Manager (under the Vega interface) did treat the half point bye correctly.

Additional Swiss Pairing Rules

If you have looked at the latest set of Dutch Pairing Rules in the FIDE handbook, you might notice a couple of additions. In the "Introductory remarks and Definitions" Section (Section A) a couple of clauses have been added to A7. They are

d. While pairing an odd-number round players having a strong colour preference (players who have had an odd number of games before by any reason) shall be treated like players having an absolute colour preference as long as this does not result in additional downfloaters (GA 2001)
e. While pairing an even-numbered round players having a mild colour preference (players who have had an even number of games before by any reason) shall be treated and counted as if they would have a mild colour preference of that kind (white resp. Black) which reduces the value of x (see A.8) as long as this does not result in additional downfloaters, (GA 2001)

It turns out that these clauses have been part of the rules since 2001, but were only included in the handbook this year (I have no idea why btw). But now that they are there what do they mean?

I asked this question of the Chairman of the FIDE Swiss Pairing Program Committee, Christian Krause. He said that what these rules do is deal with the case where a player hasn't played the same number of rounds as other players in the same score bracket (through byes or forfeits). For example, in pairing an even numbered round a player may have played an even number of games (eg only played 4 of the previous 5 rounds), the players mild colour preference (ie opposite the last colour played if they have had 2 whites and 2 blacks), shall be used to calculate x (the number of pairings in a group that do not satisfy all colour preferences), rather than treating the colour preference as 0 or undefined.

Swiss Pairings Committee

I sat at the back of the FIDE Swiss Pairings Committee as an interested observer. The meeting was pretty quick but there were two significant decisions that came out of it.

The first was a change of wording to correct an error in the Dutch Pairing Rules. Section C.6 is now changed (in part) to read "If now p pairings are obtained in compliance with B1 through to B6 the pairing of this score bracket is considered complete".

This is an important change (in Australia at least), as Arbiters who insisted that pairings should always be top half v bottom half in a score bracket would point to the previous wording to justify their case. Of course this was just an error in the wording, as committee chairman Christian Krause stated at the outset.

The second change was to do with the selection of a player receiving a bye (if necessary). Instead of the bye going to the lowest ranked player (who hasn't already recieved one), the bye will now go to the lowest ranked player who 'equalises' the colours in the bottom score group. What this means that if there is a choice of players in the lowest score group who could receive a bye, choose the lowest player that results in the most number of remaining players being paired with their correct colours.

Apparently Swiss Master already does this, but it is unclear which other programs do it as of now.

There was a long discussion by by Eduard Dubov concerning both his own pairing system, and his proposal for a new system for the Olympiad, and both of these were held over for future testing and discussion.

Pairing half point byes

I had a chat with the chairman of the FIDE Swiss Pairings Committee Christian Krause about how to treat half point byes for pairing purposes. For the moment the FIDE pairing rules are silent on this matter, and a number of pairing programs take a different approach. Sensibly he said that a half point bye should be treated the same as a full point bye ie no opponent, no colour, downfloat. Hopefully this will become part of the published pairing rules in the future, allowing authors of pairing programs to all take the same approach with this issue.

Olympiad Pairing Rules for 2010

During and after the 2008 Dresden Olympiad, there was plenty of discussion about the pairing system used. The major complaints were Match points over Game points, Acceleration, Top v Bottom, and team ordering. At the FIDE Congress in 2009 a lot of these concerns were addressed by the Technical Administration Panel (TAP), and changes were made for 2010. Here is a summary of those changes.

Firstly, acceleration is out. So the pairings start with 1 v (N/2)+1 etc
However Match points remain for pairing ordering and placing.
The team ordering for pairings has been changed from the Olympiad SB system, to a simpler method. The ranking is now Match points, then game points and finally Team rating (as defined at the start of the tournament).
Pairings will be done in the order of top down to just above middle, then bottom up to just below middle, then middle.
Pairings with a pairing group will revert to top v (N/2)+1, rather than top v bottom. The pairings will also try and have top half v bottom half wherever possible. This is more achievable than under the normal swiss rules, as colurs in teams events is less important, and will not be an overriding criteria.
For colours, not team can have a colour imbalance >2 or >2 colours in a row. However teams that both have a +1 (or -1) colour imbalance already can still play each other. Colour allocation will follow the equalisation, then alternation system.

If you want to see a full description of the system then it is published here in the FIDE handbook.