Critique Of Tournament Running Software

Running Chess Tournaments
1

Below is data I produced from analyzing the function of the Tournament Running Software I used to run a 3 round Swiss tournament of G/30, d/5. If someone can tell me how to space things out by using HTML or whatever, please let me know. To those of you that the data makes sense, please look at the data and tell me if you think that the software did a good job pairing the players from Round 1 to Round 2 and from Round 2 to Round 3 especially concerning the Color choices because Player Rated 1150 was given White 3 times in a row. Since his 3rd white in a row was the last round, it is supposed to be okay, right?

ROUND 1

Rating Color Points - Rating Color Points / Score

1734 W 0 - 1309 B 0 / 0-1
1150 W 0 - 1629 B 0 / 0-1
1541 W 0 - 1062 B 0 / 0-1
UNR W 0 - 1476 B 0 / 0-1

ROUND 2

Rating Color Points - Rating Color Points / Score

1629 BW 1 - 1309 BB 1 / 1-0
1476 BW 1 - 1062 BB 1 / 1-0
1150 WW 0 - 1734 WB 0 / 0-1
UNR WW 0 - 1541 WB 0 / 0.5-0.5

ROUND 3

Rating Color Points - Rating Color Points / Score

1476 BWW 2 - 1629 BWB 2 / 0-1
1062 BBW 1 - 1734 WBB 1 / 1-0
1309 BBW 1 - 1541 WBB 0.5 / 0-1
1150 WWW 0 - UNR WWB 0.5 / 0-1

Rating Colors Final Points

1629 BWB 3 First
1476 BWW 2 Second Tied
1062 BBW 2 Second Tied
1541 WBB 1.5
UNR WWB 1.5
1734 WBB 1
1309 BBW 1
1150 WWW 0

2

You should always post what pairing options/setting you are using.
Also, if you did anything special for the unrated’s rating (0, UNR, something else)

1 & 2 pretty much straight forward.
3: color is low priority and any switches you would do would change a higher priority – swapping colors on bottom board still gets you a 3 whites so gets you nothing – At least it is not 3 blacks & it also happens on the bottom board (least significant pair)

3

If you opt to set it to not drop/downfloat unrated players from odd score groups then you might have round three with 1309-unr and 1541-1150.

4

Jwiewel,

Should I have manually paired those last 2 pairings during the tournament as you suggested to have a better color application? What would have been proper? color application or more fair rating comparison competition? It is more fair to 1150 that 1309 plays 1541 than if 1150 has to play 1541. What is the hierarchy needing to be observed, color over fairness or vice versa?

5

Chessgary, do you believe I could have manually switched the pairings as Jwiewel suggested?

6

That is why I asked if you did anything special for the unrated. I have a note in all my announcements that TD may assign rating to unrated – I use to work with a director who would make an assignment & then adjust it based upon first half of tournament.

I missed it as an option to avoid WWW, but as it was implied since the 1309 was not eliminated from prize and you might be giving him an easier path just to improve a WWW that does not matter (nobody complains about getting WWW) — e.g. if you knew the unr was strong then make the rating assignment early rather than waiting to fix an issue where you may look like you are helping a player get an easier pair/path to a prize.

It is a bit rare to see a 3W or 3B for a 3 round 8 player with no withdrawals.

7

The 4 matches in Round 1 shows White losing in each game. Do you think that might have had something to do with it?

8

I once ran a 4 round tournament with around 14 players (as I recall) where black won every game in rounds 1 and 2.

9

Round 1 is fairly clear. Top half plays the bottom half, alternate colors for the higher rated players starting with white on top board.

Round 2 also works out fairly clearly, top half of the 1 point group plays the bottom half. The colors are all messed up already because black won all the games round 1, but what can you do. Higher rated player gets white on the top 2 boards as the only tiebreak. Same for the 0 point group except black for due color.

Round 3 is slightly tougher, but let’s start by putting out the score groups.

2:
1629 BW (1150, 1309)
1476 BW (UNR, 1062)

1:
1734 WB (1309, 1150)
1309 BB (1734, 1629)
1062 BB (1541, 1476)

0.5:
1541 WB (1062, UNR)
UNR WW (1476, 1541)

0.0:
1150 WW (1629, 1734)

The 2 score group pairs fairly naturally, with the 1629 getting the due color (black) on tiebreaks.

1476 - 1629

The 1 score group is more annoying, but in score group the 1734 can only play the 1062, which will leave the 1309 to drop down as the odd player.

1062 - 1734

The 1309, now part of the 0.5 score group, should play the highest rated player possible, which is this 1541. 1309 has white priority to avoid triple black.

1309 - 1541

Leaves the UNR to drop down to the 0 group. The UNR wins the color tie due to score (0.5 v 0.0), giving us

1476 - 1629
1062 - 1734
1309 - 1541
1150 - UNR

…which is what your program came up with. I don’t see any immediate good ways to improve the colors without messing up score groups, so I’d roll with it.

Makes getting colors right in rounds 3 and 4 really annoying.

1 Like
10

It’s not quite by the book, but I looked at moving the 1541 up rather than dropping someone down from the 1-point group, but it doesn’t really help pair that group because the colors are just messy.

11

I remember another small events (< 16 players) where the lower-rated player won every game in rounds 1 and 2. Produced some weird pairings and board assignments in later rounds.

12

The big question is how you handle the odd man. One setting some pairing programs have is to keep unrateds in their score group rather than dropping them. It assumes unrateds are simply unrated rather than the weakest in a score group and thus to be dropped down in an odd score group.

If you were using that option then the program would have done those pairings. If you did not then you get what happened.

13

Shouldnt not dropping unrated players be the default based on 29D1?

29D1. Determination. a.In the case of an odd number of players, the lowest-rated player, but not an unrated player, is ordinarily treated as the odd player and is paired with the highest-rated player he or she can play in the next lower group.

This doesn’t really apply here since the unrated player ends up the only one in the score group, making them eligible to drop down.

1 Like
14

If I had manually switched the pairings, would I have had to observe the 80 points and 200 points rules or would they not need to be observed because of the Unrated player is not considered any particular rating level?

15

Ulmont, the 1541 and unrated were the two 0.5-1.5 scores. You either drop the unrated to the 0-2 group while putting the 1541 against the 1-1 or do the reverse by dropping the 1541.

16

Right, but after the odd player (1309) drops to the 0.5 group from 1.0 and plays the highest rated player available (1541), the 0.5 group is only left with unrated players (just the one), who are (is) therefore eligible to drop.

That is, it seems to me that you handle 1.0’s dropdown before you look at who’s left in the 0.5 score group?

1 Like
17

In this case, the 1.0 group has a forced pairing, since the top rated player in the group has already played one of the other two players in that score group, so there’s only one player to drop down. (That’s why I was looking at what happens if you float someone up instead, but it doesn’t really improve things.)

Although I can’t construct a specific example on the fly, sometimes in small events you need to look at lower score groups when dealing with an odd player group to make sure the player you drop down doesn’t create problems in another score group.

18

Agreed, as noted in both 29D1a and 29D2.

Care must be taken in doing this that the odd player can be paired in the next score group, that the remaining members of both affected score groups can be paired with each other, that the odd player has not played all the members of the next lower group, and that the color consequences are acceptable (29E, 29E3)

29D2. Multiple drop downs. It is sometimes necessary to jump over an entire score group to find an appropriate opponent for an odd player.