THE COLORS

The color in the first round

There is no draw for the color on the first board. The idea is the color assigned based on some logic, but mainly to be earlier defined the pairing. The first in the starting rank takes black for the following reasons:

A. Because each round the pairing on scoregroups is based on the starting rank, the highest versus the next has its benefits:

1. Priority in color preference when they have same color history.

2. Priority to choose or be chosen first and usually best opponent that mean better criteria tie.

B. In the first round the two in the middle while are consecutive they play with opponents diametrically opposite capacity. The highest selects the last even number in the ranking and plays in the last chessboard while the next is selected from the first in the ranking and plays in the first board.

When the number of pairs is odd number the colors alternated from the first to the last pair. When the number of pairs is even number the colors are alternated until the first half of the pairs, then the alternation is reversed and the alternation continues until the last pair. Βecause the two in the middle get opponents diametrically opposed to capacity, the idea is to get opposite colors.

21-7-2013

The colors in the remaining rounds

C.04.1 Basic rules for Swiss Systems.

C.04.1.f For each player the difference between the number of black and the number of white games shall not be greater than +2 or less than -2.

C.04.1.g No player shall receive the same colour three times in a row.

The Duch system have exceptions to the rules C.04.1.f  and C.04.1.g in the final round of a tournament:

C Pairing Criteria: C.3 non-topscorers (topscorers are players who have a score of over 50% of the maximum possible score when pairing the final round of the tournament) with the same absolute colour preference (see A6.a) shall not meet (see C.04.1.f and C.04.1.g).

Quality Criteria: To obtain the best possible pairing for a bracket, comply as much as possible with the following criteria, given in descending priority:

C.5 maximize the number of pairs (equivalent to: minimize the number of downfloaters).

C.8 minimize the number of topscorers or topscorers' opponents who get a colour difference higher than +2 or lower than -2.

C.9 minimize the number of topscorers or topscorers' opponents who get the same colour three times in a row.

The Olympiad system have exceptions to the rules C.04.1.f  and C.04.1.g in any round of a tournament:
I. Colour allocation 23.a: If in a score group a complete pairing is only possible without applying articles C.04.1.f and C.04.1.g such a pairing will then be made.

The very normal case on penultimate round two players or teams (say the 1 and 3 or 2 and 4 of the starting rank list) with the same color history and the previous two rounds they have the same color. The normal is to play against each other to claim first place and let's one of them take white for third time and still make a difference (white - black) equal to 3.

The absolute application of the rules for the colors may be cause of creating floater. eg after the second round of the "unlikely" case in a scoregroup if everyone has two victories or defeats and two white or black then all are floaters.

Some participants would prefer and third black in any round not to be up floater in better scoregroup. Also some participants will give to opponents third white in any round in order these opponents not to be down floaters in worse scoregroup.

In any round can be gathered in a scoregroup some participants who have the same absolute color preference example after 4 rounds WBWW, BBWW and while all of them can play each other, they are looking opponents in other scoregroups.

For these reasons, for all the participants the "Basic rules for Swiss Systems" (C.04.1.f and C.04.1.g) can be as follows:

A) one color is not allowed more than two rounds
1) than the other,
2) continuously,

B) All pairs of a scoregroup are created in priority with the limitations:
1) (A1) and (A2),
2) only (A1),
3) none.

Examples and comparison rules for colors here.

3/5/2018

Color Allocation Rules.

For each pairing apply (in descending priority):

1. Grant both color preferences.
2. Grant stronger color preference.
3. Alternate the colors to the most recent round in which the colors were different.
4. Grant the color preference of the highest ranked.

According to these rules after the 8th round in case of the only two players in a scoregroup:

A participant: WBWWBWBW (5W 3B)
B participant: BWBBWBWW (4W 4B)

The paradox is that the A participant will get White for the 9th round because of rule 3.

The quantitative difference in color between the two opponents is not examined while the A participant should get Black.

07/23/2014

The Accelerated Swiss System

The classic way to pairing for the first round because is considered that from the first to the last participant are evenly balanced, contributing less and less to the emergence of reliable first (and last if interested) positions, as the number of participants increases, or if the rounds are not enough. In a tournament 9 rounds for 20 players in first round, 1 plays with 11 which is very different than a distance in a tournament for 200 players where 1 plays with 101.

In a usual tournament of 9 rounds with 200 players a GM can easily begin with 3 points in 3 games and in the remaining 6 rounds makes draws with the other GM, but if makes 2 wins and 4 draws, then the player touches the top with 7 points. And thus the games of good players are less and less fighting, and all GM are together first in multiple tie. The fighting spirit of players is condition, but if this happens, thus as the tournament become this it is limited at two rounds at least. Thus as usually now becomes in the first round those that are before the middle of initial list they win real point playing with the last ones, while those that are afterwards the middle lose real point playing with first.

The accelerated swiss has been applied in past but if it does not become with right parameters, the players but also the arbitrators cannot reconcile itself with this. If a tournament had 200 players took 2 virtual points for 2 rounds the first 100 players and 0 points the remaining 100 players. Thus the tournament became lightly accelerated and the third round was similar as first round.

If the tournament has x players for the two first rounds the players can take virtual points as follows:
The first x/4 (2 points)
the next x/2 (1 point)
the next x/4 (0 points)
that is to say for 200 players we would have:
1-50 (2 points)
51-150 (1 point)
151-200 (0 points)

This is determined very easily in programs of pairing that support the accelerated system. It is a normal simulation of two first rounds that it is not necessary they become and in the substance the tournament it begins from third round of classic swiss. The idea is become 2 games with more near likely opponents and if somebody makes points it plays also with better player and no thus as it becomes it plays by definition. Thus each tournament it becomes "+2 rounds" important to the end and it improves the tournament of a player that he will makes points. The good player if makes a mistake has rounds to correct it. In tournaments with simple swiss with x participants are required y+2 rounds where y is the smallest integer where 2^y is greater than or equal to x while in accelerated swiss only y rounds required. If the number of players is too much big is not essential their separation in more tournaments because they can be separated in virtual scoregroups.

The virtual scoregroups are created answering in the questions:

a) For how much rounds the players will have the virtual points?

The answer is usually the first 2 rounds.

b) How many scoregroups?

If the players had played already two rounds and the pairing it is not likely to happen the scoregroups would be three with (2) (1) (0) points else five with (2) (1,5) (1) (0,5) (0) points.

c) And the difficult question: Which count of players will have each scoregroup?

The number of players in the scoregroup is the total of players multiply the probability that the points to realised. ex. for 200 players the three scoregroups would be 200 = 200 * (0,5 + 0,5) ² = 200 * (0,25 + 2*0,5*0,5 + 0,25) = 50 + 100 + 50 while the five scoregroups if the draw is considered with probability 20% they would be 200 = 200 * (0,4 + 0,2 + 0,4) ² = 200 * (0,16 + 0,16 + 0,36 + 0,16 + 0,16) = 32 + 32 + 72 + 32 + 32.

Ιn virtual scoregroups of the first round, the count of players is even number. In the pairing of first round in any system with virtual scoregroups, these are not created with odd number of players because the last player of the scoregroup plays with first player of the next scoregroup so that play successive players of initial list. If the scoregroups are 3, then the number of players in the middle group divided by the number of players in the first group is in absolute value the closest result to number 2. If is x the even number of players (the player who takes bye is excluded), then the first and the third scoregroup each has integer(x/4) players +1 if the result is odd number and the middle scoregroup has the remaining players. ex.

100=26+48+26
102=26+50+26
104=26+52+26
106=26+54+26
108=28+52+28
110=28+54+28

29-7-2011

The algorithm in a scoregroup

Usually, during the games, many people wonder how fair a pairing is, specially when it concerns a scoregroup. The answer is not easy to be given but it is confirmed that some of pairs are not "equal".

The issue here, is what goes on with the ways of pairing at a scoregroup. This can be more understendable if you read the following examples.

We consider a scoregroup with even number of x players at a round. Using the ELO system (or anyone else considered fair) we put the players in rows. For simplicity reasons, we consider that successor players are distant by one point. (For example 20 degrees ELO). That means that if the first player has 2400 ELO (x at the examples) the next will have 2380 (x-1 at the examples) and the next 2360 (x-2 at the examples) and so on.

We consider that a scoregroup of six players at a certain round, according to the rules, will become pairs in these ways, given by priority:

The first column is the serial number and the second the equivalent way. The players which make pair are given in parenthesis and they are devided by the symbol of deduction. The first number show the bigger player and the pairs are divided by the symbol of addition. The third column shows the different capabillities and their sum up and the forth column is the average and the standard deviation of the differences. The priority of the ways comes clearly out by matching procedure. By the fact that we choose for the pair {(1,middle+1), (2,middle+2) etc} equal distances between the players, fair pairing at a scoregroup is the one in which the different capabillities among the pairs are as much as possible equal. That means that the standard deviation can be, as much as possible smaller. So, which way is the most fair? And, if it is reject (maybe because some of players have been pairs in previous rounds) which one comes next?

We put the ways at serial row according the standard deviation with second criterion the average and the result is:

The above way is different from the first way which follows the rules. The way (15) and (1) come first because their standard deviation is 0. At the way (15), where each player plays with his next, the sum up of the differences gets the less possible, whilst at the way (1) where the players play with the above the sum up gets the highest. Wich way of the two has priority?

The way (15) follows the rule of the system "players whith the same points or almost the same, play together" but "toughens" the games from the first rounds and the winner of the games really diserves to be the winner. The way (1) shows weakness at a scoregroup with several players, usually at the first rounds that the results of the games are almost forseen and after the last round is very possible that the winners have the same points. In addition players who are little lower than the middle of the scoregroup are misjudged (they play with the best of the scoregroup) comparing to the players at the middle of the scoregroup and the ones which are a little higher them (that play with the worre of the scoregroup).

This is an example with 8 players that shows the observations mentioned above.

We put the ways at serial row according the standard deviation with the second criterion the average and the result is:

From the example, after the sorting, it is clear that the ways, where the first players play with the last ones, have passed on to the end.

Of course, a system that needs that kind of calculations, needs computer to operate. It can be used at a scoregroup or ...at all players of the round which get into rows according to any criterion (i.e. pointsX10000+ELO, i.e. 5,5X10000+2400=57400). The calculations are being done by the exact differences of the players at each pair and again we can't say how much time is needed for the absolute way.

A simple way that is possible to be used is the first player that asks for an opponent to choose by rate: middle+1, middle, middle+2, middle-1, middle+3, middle-2 etc down or under the one in the middle, not to goes choose first the middle+1 till the last one then from the middle to the above. Specially when there are limitations about the pairing, the choice from the middle to the last is more likely to be mode.

That simple way that mentioned above sets a little better the priority in ways. For the 6 players from the first example the ways would be:

More criteria can be suggested to define the choice of the opponents. When the first player chooses firstly the middle+1, it's a high handed act because it does not mean that the different capabillities {(1,middle+1), (2,middle+2) etc} are equal. The choice depends on the setting of the ELOs at a scoregroup, even if all the players have the same ELO in a scoregroup however the pairs are mode is always fair.

To the system used today rules are added in order to be fair in some cases but are unfair in other cases. Further more the fact that this system has 50 and maybe more rules and its limitations (scoregroups of the above, the middle and the low, up and downs and colours) makes the use of a computer necessery. That is a complicated problem to be solved and still its solution can't be cheked by arbiters and the players.

So, it is best not to understimate the simple swiss system by which the player who asks for an opponent chooses the one who finds first. Simplicity is always beautiful!

1-12-1993