Round Robin and knockout competitions have been found wanting in wargames competitions, perhaps the most popular draw has been that known as ‘Swiss Chess”. However, recent discussions with CWS members and observation of the smallish numbers of competitors in many competitions have persuaded me that;
- The “Swiss Chess” system is not well understood;
- That “Swiss Chess” is not always the best system to use, especially where competitors do not outnumber available rounds by many.
Explanation Of The Swiss System:
The basic idea of the Swiss System is to cater for large numbers of competitors in a tournament of few rounds, allowing all to play in every round, yet retaining the knockout idea of providing a dramatic finish by matching the best two performers in the final round. This is achieved by progressively pairing players on the same or nearly the same cumulative scores. So far, so good.
Basic Swiss System Laws:
In Chess Tournaments there are overriding rules that need to be borne in mind when drawing up pairings each round. The following, with parenthetical comments by me, are taken from the Official Rules of Chess (Official US Chess Federation Edition, 1974) pp86 et seq. All pairings are subject to the following basic Swiss-System laws;
- A player must not be paired with any other (Individual) player more than once (A lot of people miss this point).
- Players with equal scores must be paired if it is possible to do so (Bearing in mind rule A above).
- If it is impossible to pair all players with equal scores, every player who is not paired with an opponent whose score is the same as their own must be paired with an opponent whose score is as close to their own as possible.
Each player should be assigned a number signifying relative seeding. For the first round, the top half of the list of competitors are matched against the bottom half by “direct pairing”. In a tournament of twenty, the top seed would play the eleventh, the second the twelfth, and so on. Given an odd number of players, the bottom seed would receive the bye, and should be credited with a Full Point. As a variant, there is much to recommend juggling the first round slightly to avoid matching players from the same club or region. However, in fairness to other players, they must be prepared to face each other in later rounds.
To determine pairings for subsequent rounds, players are grouped according to their points score so far, each group being all players on the same cumulative score. Within each group, players are listed in their original order of seeding. This again allows direct pairing “Top half vs. bottom half” within each group.
Naturally, the points groups will not often divide evenly. In this case the bottom seeded player becomes the “Odd Player” and is paired against the top seed of the next group down. As the draw is constructed “Top down”, this may have a ripple effect through the field, but there is no real problem provided the Basic Rules are observed and top half vs. bottom half direct pairing is maintained as far as possible. The fact that it will not always be possible to pair all players within the group should not cause too many problems either, if those rules are observed.
Where there is an odd number of players, the bye is awarded to the bottom seeded player in the bottom points group, provided that that player has not already received a bye.
Number Of Rounds:
The optimum number of rounds to players is the same as for a Knockout Contest i.e. n rounds for 2 to the power of n players; e.g. say 5 rounds for 2 to the power of 5 (ie.2x2x2x2x2) = 32 players. Naturally, this ideal is seldom achieved, but this does not matter if the number of players is in the right ballpark. Five rounds would accommodate twenty players all right -though 4 would also do- and many U.S. 5 round weekend Chess tournaments fielded upwards of 3 hundred! In this Case organisers would be prepared to see several players with a 100% score.
Problems do arise when there are not many more players than rounds available. If you have 6 rounds available and only 8-10 entries then the top two players will meet in Round 3, after which the mismatches get worse until by Round 6 the top players are playing the bottom. This is totally contrary, in my view of the spirit of the Swiss System. Some other system needs to be devised for this situation.
In my view, scores should be 1 for a win, 1/2 for a draw, and 0 for a loss. Systems such as I got stuck with directing the 1990 NATCON Napoleonics of 1 for the win, 0 for any other result, are in my view unfair. DBM uses an attractive victory points (VP) System, but I don’t really see that it can be readily accommodated in a Swiss System of pairing; it seems much more suited to a Round Robin competition. Otherwise, DBM provides a way of determining win-loss, and the VP’s can go towards breaking ties at the end of the competition.
Much effort even in the Chess world has been devoted to avoiding “grandmaster draws” without success; and it appears that “defending the baseline” or similar Fabian tactics tend to ensure success in Wargames Competitions. To avoid this kind of thing requires some other way of determining an individual result, such as DBM’s system, or by using scenarios, objectives, or counting double for non-light troops in the opponent’s half of the board. One should not tinker with the scoring systems above, but only with the system of awarding points. If one is using a different scoring system, then don’t use a Swiss System for pairing.
Especially where large numbers are involved, or where there is a wide margin for draws -as in chess- tied results are quite likely, if not for 1st, then for minor placings. For Chess tournaments using the Swiss System, the Buchholz (BZ) tie-break method was devised, named after a chap called Buchholz. The basic idea is to determine the “degree of difficulty” of individual results, which is achieved by totalling the final scores of an individual’s opponents. Nothing could be simpler! I have not known it to affect the distribution of prize-money in New Zealand Chess tournaments, but it has determined a Champion. It is quite unusual, though not unknown, for players on the same match score to have the same SB count, though this would be more likely in small tournaments.
For DBM and other game systems using VPs to decide the game result, then their total offers an alternative way of determining the overall placings for Tied results.
Let us follow a Tournament history to show how the Swiss system works. We will use a DBM tournament.
Alan beats Earl, Brenda ties with Fred, Carol beats Gwen, Dora loses to Harry.
After Round 1
At this point several factors emerge:
- The 1 Win and 1 Loss groups are uneven;
- The two players in the 1 Draw group have played each other!
Harry as the bottom seed in the 1 Win group is the “odd” player and must play Brenda, the top of the next group down. In the 1 Draw group, Fred then becomes the “odd” player, and is matched against Dora, the top of the 1 Loss group.
Standings after Round 1
Alan beats Carol, Harry loses to Brenda, Fred loses to Dora, Earl beats Gwen
Comments after Round 2:
After Round 2, clearly Alan and Brenda will play out a “final”. The 1 Win, 1 Loss group is even, but an ordinary “direct pairing” won’t work because Dora and Harry were paired in Round 1. Retaining “top-half vs. bottom-half” gives us Carol-Harry and Dora-Earl, neither of which matches have yet occurred, so we go with that.
Standings after Round 2
Alan loses to Brenda, Carol beats Harry, Dora ties with Earl, Fred ties with Gwen
After Round 3
The BZ Tie-Break scores give a degree-of-difficulty, and you can see that Harry and Fred paid for their first round upsets with very tough competition later. Brenda’s “Swiss Gambit” of a (unintentional?) Draw in Round 1 paid off with an easy Round 2, before facing a tough challenge in Round 3 to win the tournament.
Final Standings & BZ Tie Break Score (where required)
2½ – Brenda
2 – Alan (6)
2 – Carol (3½)
1½ – Earl (4)
1½ – Dora (3½)
1 – Harry (6)
1 – Fred (4½)
½ – Gwen
by Ion Dowman – AATB No.2 September 1996.
This article originally appeared in the Christchurch Wargaming Society’s “Southern Sortie” Magazine in 1993…