Hnefatafl Internet Championship

[Hagbard]

This is my bid for an evaluation of the Championship Tournament 2020:

The tournament setup this year was similar to last year.

Changes from last year:

• The champion winner of the year before, if he joins, goes directly to the Final.
• World Championship not only in the Copenhagen strong king variant, but simultaneously also in the Saami Tablut historical king variant.
• 3rd criterion to solve a tie in case of direct comparison and Sonneborn-Berger score equals: win with shortest game length.

The changes worked well.

Like last year, the criterion for distributing the players into Round 1 and Round 2 aimed for 7 players per Round 1 group and 6 players per Round 2 group.
And like last year the timing was graduated with a short timing for Round 1, longer timing for Round 2 and longer yet for Round 3.

What could be improved next year:

• The game timing of the Final made it possible for a game to last 3 months. The timing should be modified, so that for an extremely long game the minimum time before timeout is gradually reduced from the initial 25 hours to 13 hours.
• The tournament Round 1 should start September 1st instead of Sep. 15th.

The winner was found after 130 days (last year 85 days), and the whole tournament was finished after 186 days (last year 87 days).

34 players joined the tournament.
In total 284 games, 15772 moves and 2423 killed pieces.
Longest game was 189 moves.
Shortest game was 10 moves.

Tournament page:
http://aagenielsen.dk/wtfturnering2020.php

Congratulations to World Tafl Federation Champion 2020, Mario Aluizo ("casshern"), Los Angeles, USA, winner of both tournaments!
Many thanks to Adam Bartley for being the umpire of the Tablut tournament, and to the 34 players who played in one or both tournaments!

[Hagbard]

An alternative grouping of the World Champ. tournament.

The latest years a grouping of 6 has been used.
This gives you 5 opponents and 10 games per round. All players have 10 games and many have 20 games.

A grouping of 5 would make it less work to participate. All players would have 8 games and many would have 16 games.

If an extra round is added, then a 32 player tournament would look like this:

1. round.
3 groups of each 6 players (usually one no-show per group in the first round). This is 18 players, and 6 proceeds to next round.
Short timing.

2. round.
3 groups of each 5 players. This is 15 players (6 + 9), and 6 proceeds to next round.
Short timing.

3. round.
2 groups of each 5 players. This is 10 players (6 + 4), and 4 proceeds to next round.
Medium timing.

4. round, the Final.
1 group of 5 players (4 + 1)
Long timing.

[Draganov]

I am writing this article to present you the idea of making a tafl elimination tournament. After I had taken part in the 2020 World Tafl Championship, I found out that the tournament was really exhausting and took 7 months to finish. The final was not so dynamic and interesting since the winner (Casshern) secured his title 2 months before the end of the tournament. If we have an elimination tournament format, then the winner will be decided on the last day of the tournament. It will keep the interest of the viewers until the end.
Here are some format explanations:

1) First round (Group round)
In this round we can split the lowest ranked players in groups. We may have 2, 3, 4 or as many groups as needed. It depends on the number of the players that joined the tournament. The players are going to play 2 games against each other. The best 2 players from every group will qualify for the next round.

2) Second round (Elimination round)
In this round we are going to have the best two players from each group from the first round plus the 8 or 10 of the highest rated players who joined the tournament on this level. The point is to have 16 or 32 players here in this round. It depends on the number of the players who joined the tournament. If we have 16 players in this round, then we are going to have a 1/8 final here.
All the players are split into matches. The highest rated player will face the least-performed player from the second finishers in the group phase. The second best-rated player will face the second least-performed player from round 1 and so on. In every match the two players are going to play 4, 6, 8, 10 games (it needs to be decided). All of the games start at the same time. The player who wins most of the games continue to the next round. In case of a tied players we may decide the winner by additional criteria such as the shortest win achieved by a player.

3) Third round (Elimination round)
In this round we will have 8 players and will split them into 4 matches. They are going to play in the same rules like in the previous round. We can organize the matches by matching the highest rated player with the lowest rated.

4) Fourth round (Elimination round)
This is the semi-final with only 4 players left. We can again match them by their rating.

5) Final round (Elimination round)
In the final we will have only 2 players left. They should play a match of 10 games to decide the winner. In my opinion having 10 very high quality games between the two greatest players in the tournament would be beneficial for the development of the tafl game. We could advertise the final match as a Player 1 VS Player 2 and even make some posters and present them on the Internet to attract new tafl players.

Let me know what do you think about my proposal. We can test this elimination format in another smaller test tournament without having a group phase. Let’s have a test elimination tournament and see if this format will work. As a start we can have matches consisting of 4 games. However, all 4 games must start at the same time.

[Abbas Agraphicus]

Hi, I personally think the elimination tournament format is a good idea, but maybe I'd keep the number of games lower, maybe the regular 2 games til the quarterfinals, 4 games in the semifinals and 6 games in the final.
Bye!

[Draganov]

In my opinion having only 2 games between players in round 2 will not be enough to determine the better player. There is a solid chance that the result after only 2 games will be 1:1. Then we will have two options. The first one will be to resolve the tie by applying some additional criteria such as game length, number of taken pieces or something else. The second option will be to start a new set of 2 games between the tied players in order to resolve the tie. However, it will increase the length of the tournament.
When the players have to play 4 (the minimal reasonable amount), 6, 8 or 10 games against each other, then the chance of a tie is minimalized. With the rule of all games starting at the same time, it doesn't matter how many games the players will play against each other and it won't increase the length of the tournament.
However, I agree that on the lower level of the tournament we had better had less games between the players. It will not be interesting to see match of 10 games between a 1800+ player against a 1400 player. On the other hand, in the final round would be nice to see a long battle between the two best players.
If one day this format becomes the standard one for the World Championship, then the reigning world champion will be better to join the tournament on at least the semi-finals or even we can use such a tournament as a Candidate Tournament and the winner will face the reigning champion in a long match of 10 games in the same manner as in the chess world championships.

[Hagbard]

Perhaps a format test could be run of:

Welsh Tawlbwrdd Double-Elimination ?

Short timing as it is a test run.

[Draganov]

Yes, it sounds good. I appreciate that you are also going to organize a Fetlar double-elimination tournament. Let's test this elimination format. However, I think that for the final we need to have 4 games between the two finalists starting at the same time.
Hagbard, what do you think if we have 4 games between players in each round during the test tournaments? I think the length of the tournament will be the same if we have 2 or 4 games between opponents in a single match. However, in my opinion with only 2 games, there is much more risk the winner to be determined on the additional criteria which is not good.

[Hagbard]

The Swiss tournament system works like the Double Elimination, only without the elimination.
A principle in the Swiss system (Monrad) is that players all the time as far as possible play against their equals.

If one takes the Swiss principles and adds the double elimination, then an example of 8 players (#1 - #8) would be:

Round 1
-------
(the top 4 players are rearranged to delay the meeting of #1 and #2)
#1 #4
#2 #3
---
#5 #6
#7 #8

Round 2
-------
#1 #2
#5 #7
---
#3 #4
#6 #8

Round 3
-------
#1 #5
---
(rearranged so that #2 and #3 does not meet again)
#2 #7
#3 #6

Round 4
-------
#1 #3
#2 #5

Round 5
-------
#1 #2 meet again
If #1 wins, then #2 is out.
But if #2 wins, then both players lost once, and a game more is needed to determine a winner.

When all games as far as possible are against equals, then it is as much as possible uncertain who will win these games.

[Draganov]

The Swiss system sounds really good. What really blew my mind was the fact that UEFA Champions League will use a Swiss system for the 2024/2025 season. I think the Swiss system is nice and will become more popular in the future since it is well-balanced and allows conducting tournaments with a lot of players. It is not affected by the number of the participants. We can have a Swiss tournament with 100 or 1000 players and its length will be exactly the number of rounds that were decided before the start of the tournament.
If we are going to implement a Swiss system to our world hnefatafl championship, then my proposal would be to use it in order to determine the semi-finalists. For example if the world champion joins the tournament, then he/she will join on the semi-finals and will not take part in the rounds before. The semi-finals using an elimination match system will work best in my opinion. It will face 4 of the best players and will be a tafl masacre.
Prior the semi-finals we can use a Swiss system to determine the three semi-finalists except the world champion who will join on the semi-final directly. We can have 3 rounds for the tournament:

1) First round (Swiss system) -> one group of all of the participants who registered for the tournament except the world champion.
In this round we can have as many rounds as we decide. However, it should be a reasonable amount since it will not be good to determine the three semi-finalists by only 2-3 rounds. Here we need to have at least 5 rounds, consisting of two games between players (one with whites and one game with blacks). In order to reduce the length for this huge Swiss system round we can have a standard time buffer of 2 days or 4 days if you prefer to have a bigger time buffer. The best three players (or 4 players if the world champion didn't join) will continue to the next round.

2) Second round (Elimination system) -> semi-finals with the 4 best players.
In this round we can have the world champion joined and the match distribution could look like this:
World Champion vs the Third player from the group phase
the Best player from the group phase vs the Second best player from the group phase
If the world champion didn't join the tournament, then we could have a semi-final with the best four players from the first round and the distribution could be:
Player 1 vs Player 4
Player 2 vs Player 3

3) Final round -> a match between the two winners from the semi-finals
We can also have a match for the third place between the two players who lost their semi-final matches.

[Hagbard]

World Champ. tourn. of Copenhagen Hnefatafl last year had 26 participants.
Had a Double Elimination (Swiss style) been used, the tourn. could've run along these lines:
Round 1 and round 2 can run simultaneously.

Round 1
Distance -1
1 Bye
2 3
4 5
6 7
8 9
10 11
12 13
14 15
16 17
18 19
20 21
22 23
24 25
26 Bye

Round 2
Distance +1
1 2
3 4
5 6
7 8
9 10
11 12
13 14
15 16
17 18
19 20
21 22
23 24
25 26

Round 3
Distance 2
1 3 (out)
2 4 (out)
5 7 (out)
6 8 (out)
9 11 (out)
10 12 (out)
13 15 (out)
14 16 (out)
17 19 (out)
18 20 (out)
21 24 (out)
22 25 (out)
23 26 (out)

Round 4
Distance 3
1 Bye
2 5 (out)
6 9 (out)
10 13 (out)
14 17 (out)
18 22 (out)
21 23 (out)

Round 5
Distance 4
1 Bye
2 6 (out)
10 14 (out)
18 21 (out)

Round 6
1 10 (out)
2 18 (out)

Round 7
1 2

Since each player only has two games at a time, perhaps short timing could be used all the way, and the tournament finished in perhaps 3-4 months, starting Sep. 1st.
If the coming Double Elimination tournament test works all right, perhaps this format could be used for the World Champ. tournaments later this year.