Hnefatafl Internet Championship

[Hagbard]

Thank you for giving this subject some thoughts!

1) Very interesting consideration! Also interesting, I think, that by this method you still arrive at the same winner.
The top players, who are likely to win the tournament, are also probable to have many shared opponents, because after the four rounds, broadly speaking all top players have played against each other.
With triple elimination there'll probably be a reduced risk of ties.

2) Last year with timing 4x, the increment was 8 hours per move. This was near the threshold of trouble, because a number of players ended in time problems. 5x gives 10 hours per move.

3) Blitz has its own problems. I assume that it requires much attention from the player. Probably the player has to keep watch over the screen for many hours, so that he can react quickly at all times.


4) Yes! That's the idea.

Round 1 Daldos would be

Code: Select all

-   x   ...
  -   x ...
x   -   ...
  x   - ...
...

(half Robin every second game)

Round 2 Sahkku would be

Code: Select all

- x   x ...
x - x   ...
  x - x ...
x   x - ...
...

(half Robin every second game)

The points of Round 1 and Round 2 are added. No elimination.

The half Robin to avoid the very many simultaneous games of a full Robin. Last year we were 18 players for the race games. That would with full Robin be 34 simultaneous games in Round 1 and again 34 games in Round 2.

[Draganov]

I might be wrong but I think there were games in which the time buffer was reduced and to less than 1 hour and then the time increment was about 2 hours. I think the minimum amount of time increment should be 6 or even 8 hours despite how much time is left in the player's time buffer.

[Hagbard]

Yes, for long games time increment could gradually fall to 2 hours.

[Hagbard]

Tony Higginson ("tonythebook") sent this message:

I'm in the current Sahkku tournament with Nicholai and others. He/I were discussing a points systems for these race tournaments akin to Formula One racing, allocate say 20 points for whomever wins the tournament and 15 for 2nd and then 10-8-6-2 type thing, I'm sure plenty would like that?

Ingenious idea! Formula One racing points!

I googled the Formula One system and found:
https://www.redbull.com/in-en/formula-1 ... stem-guide

This is how the F1 points system works.
Here's an explainer how how the F1 points system works, and how drivers can earn enough points to claim Formula 1 world championship?

How does the scoring system work?Points are awarded based on the driver’s ranking at the end of a Grand Prix.
First: 25 points
Second: 18 points
Third: 15 points
Fourth: 12 points
Fifth: 10 points
Sixth: 8 points
Seventh: 6 points
Eighth: 4 points
Ninth: 2 points
Tenth: 2 points
Since the top 10 drivers receive points, winning isn’t everything. A driver with the most wins throughout the season doesn’t have a guaranteed championship win, as another driver may have posted more consistent results throughout calendar.

What happens if two drivers have the same number of points?
If two drivers finish the last Grand Prix with the same number of points, the world champion will be crowned according to number of wins. If both drivers have the same number of wins, the ranking will be decided by number of second place finishes, then to third places. There will always be a solo champion.
Incidentally, never before in the history of Formula 1 have two drivers had the same number of points at the end of the season.

The Formula One system builds on many races during the year.
Perhaps we could use it in the next Racing Games Championship tournament with four races:

1. Daldos 14
2. Daldos 16
3. Sahkku
4. Sahkku

Four races, because the system works better with more races.
Not Daldos 12, because the short board is a bit too arbitrary.
Two times Sahkku, because Sahkku is more complicated and challenging.

[Hagbard]

Counting the latest six Race Games tournaments 25/2 - 30/3 2024 by the Formula One method:

1. Holidayinnes 105 points
2. Xerxes 88 points
3. Potapych 72 points
3. Hagbard 72 points
5. Nikolai Hristov 59 points
6. Colophonius 54 points
7. Rosk 49 points
8. Jrton80 44 points
9. Jilles 43 points
10. Tonythebook 42 points
11. Brave-frog 25 points

If in a primitive way adjusting for missing tournaments:
1. Jilles 129 points
2. Hagbard 108 points
3. Holidayinnes 105 points
4. Xerxes 88 points
5. Brave-frog 75 points
6. Potapych 72 points
7. Nikolai Hristov 71 points
8. Rosk 59 points
9. Colophonius 54 points
10. Jrton80 44 points
11. Tonythebook 42 points

This corresponds reasonably to the calculated ratings for race games of latest year:
1. 1608 Jilles (105)
2. 1580 Xerxes (664)
3. 1529 Holidayinnes (744)
4. 1512 Nikolai hristov (228)
5. 1492 Potapych (806)
6. 1472 Hagbard (310)
7. 1460 Colophonius (728)
8. 1453 Brave-frog (554)
9, 1418 Jrton80 (730)
10. 1406 Tonythebook (559)
11. 1391 Rosk (706)

[Hagbard]

Tony Higginson ("tonythebook") sent this message:

Are you going to see about setting up a Grand Prix? Maybe advertise it to get at least 10 players? can work out a points system then?

The test running now has 14 players and two races.

For points system, I think we can as well go straight with the F1 method:
First: 25 points
Second: 18 points
Third: 15 points
Fourth: 12 points
Fifth: 10 points
Sixth: 8 points
Seventh: 6 points
Eighth: 4 points
Ninth: 2 points
Tenth: 2 points

Tomorrow morning the second test race can start, Daldos16.


As for the real thing where to find the race Grand Prix winner 2024, perhaps starting a week or two after this test?
So we find a race winner now instead of in September?

I'd still propose it run this way:

1. First race Daldos14 starts. Interested players just join the tournament as usual; it is open for 7 days as usual.
2. Second race Daldos16 starts the same way.
3. Third race Sahkku starts the same way.
4. Fourth race Sahkku starts the same way.

5. The points are summed. I understand from the F1 that if a player misses some race, his score from that race is simply 0.

[Hagbard]

We had a full test of the Formula One points system and the race games (hat tip Tony Higginson, "tonythebook").

5 Round Robin rounds:
Sahkku
Daldos16
Daldos14
Daldos12
Sahkku

Sahkku twice, because it's the most complex of the variants.

Personally I've wondered about the Daldos12 game.
However, it's obviously not randomness which makes the 12-board difficult, because the good players win.

All the strong race game players participated.

Results:

F1 points test round 1 Sahkku:
#1 25 points: Holidayinnes (22)
#2 18 points: Tonythebook (16), Hagbard (16), Colophonius (16)
#5 10 points: Rosk (15), Draganov (15)
#7 6 points: Potapych (13)
#8 4 points: Nikolai hristov (12), Cacreal (12), Jilles (12)
> #10 0 points: Xerxes (11), Brave-frog (9), Jrton80 (9), Thorfin (4)

F1 points test round 2 Daldos16:
#1 25 points: Draganov (19)
#2 18 points: Jilles (17)
#3 15 points: Holidayinnes (16)
#4 12 points: Potapych (12), Jrton80 (12)
#6 8 points: Xerxes (11), Rosk (11), Hagbard (11), Brave-frog (11)
#10 2 points: Colophonius (10), Tonythebook (10)
> #10 0 points: Herjan (5)

F1 points test round 3 Daldos14:
#1 25 points: Jilles (15)
#2 18 points: Cacreal (14), Hagbard (14)
#4 12 points: Holidayinnes (13), Draganov (13)
#6 8 points: Potapych (11)
#7 6 points: Xerxes (10), Rosk (10)
#9 2 points: Brave-frog (9), Jrton80 (9)
> #10 0 points: Colophonius (8), Tonythebook (6)

F1 points test round 4 Daldos12:
#1 25 points: Holidayinnes (13)
#2 18 points: Potapych (12)
#3 15 points: Xerxes (11)
#4 12 points: Jilles (10)
#5 10 points: Jrton80 (9)
#6 8 points: Tonythebook (8), Brave-frog (8)
#8 4 points: Rosk (6)
#9 2 points: Hagbard (4)

F1 points test round 5 Sahkku:
#1 25 points: Hagbard (13), Rosk (13)
#3 15 points: Jilles (12)
#4 12 points: Potapych (10)
#5 10 points: Holidayinnes (9)
#6 8 points: Xerxes (8), Tonythebook (8)
#8 4 points: Jrton80 (5)
#9 2 points: Brave-frog (3)

Final result; the parantheses show number of won games.
F1 points test sum Sahkku + Daldos16 + Daldos14 + Daldos12 + Sahkku:
#1 87 points Holidayinnes (73)
#2 74 points Jilles (66)
#3 71 points Hagbard (58)
#4 56 points Potapych (58)
#5 53 points Rosk (55)
#6 47 points Draganov (47)
#7 37 points Xerxes (51)
#8 36 points Tonythebook (48)
#9 28 points Jrton80 (43)
#10 22 points Cacreal (26)
#11 20 points Brave-frog (40)
#11 20 points Colophonius (34)
#13 4 points Nikolai hristov (12)
#14 0 points Herjan (5), Thorfin (4)

The result by this method seems accurate.
Perhaps the number of games can be reduced, fx. in a tournament with 20 players this way:
Round 1 20 players
Round 2 20 players
Round 3 18 best players
Round 4 16 best players
Round 5 12 best players.

Counting the simple way just numbers of wins would in this test tournament give nearly the same result. But the Formula One method of counting makes the results more distinct and adds to the fun, in my opinion -

The tournament took 45 days.

[Hagbard]

December 2023 Draganov wrote in the chat forum:

2023-12-05 22:56:34 Draganov: I also want to suggest to place all of the unknown and new players who completed at least 1 game before the championship but hadn't completed more than 10 games to be placed in a separate group and play against each other. After that the winner of the newcomer's group can proceed to the next round and join the other players who qualified for round 2.

Good idea!
I'd propose a separate Round Robin between
- new players with 1-10 games
- players who were no-shows last year

If there are as many new players as there were 2023, I'd say that the two winners proceed to round 2.
They bring to round 2 their losses from round 1.

[Hagbard]

December 2023 Draganov wrote in the chat forum:

2023-12-05 22:51:53 Draganov: I am happy to win the championship especially on the 25th year of the website. This year, the championship was brutal and exhausting at least for me... I feel really tired after all hard battles.
...
On the other hand, I think the format of the championship should be improved since it takes months to be completed and it is very exhausting for the top players. Maybe this year we will manage to find a working quick time control which will fasten the next world championship.

A thing that perhaps could be tightened:
If you don't move for 2 days, then timeout. (2023 the limit was 3 days.)