@kleineme "|> You can't play a Swiss where you have 1 round less than the participants @odoaker2015 (#26)
Yes, you can"
Then I suggest the following experiment: since you probably come from Germany, I have seen this from your teams, you might have the software Swisschess. If not, you can download an appropriate software. Simulate a tournament with 10 participants and 9 rounds in the swiss system. In each round you let the player with the higher rating (DWZ or Elo, it doesn't matter in principle) win. I can't wait to see if you can make it all nine rounds! I tried this once and after 7 rounds no further pairings could be calculated. You must also use Dutch System, as in Lichess!
And you still haven't shown me the crucial passages in the Handbook of FIDE § C. 04!
Can't wait to read it!
@kleineme And please have a look at this post lichess.org/forum/lichess-feedback/swiss-tournaments-are-on-lichess?page=19#190 Independently of me, someone else did a similar experiment! With similar results!
No problem here with SwissChess v9.24, nine full rounds with FIDE settings. Here's the SWT: http://www.isewase.de/files/test10.swt
Right, that's one of the edge cases. Here it doesn't work out due to the same reason as for Lichess tournament lichess.org/swiss/UFyKy5jN, as explained by TacticalBert in lichess.org/forum/lichess-feedback/swiss-tournaments-are-on-lichess?page=22#211
Here's the SWT: http://www.isewase.de/files/test10_remis.swt and here's the cross table after five rounds: http://www.isewase.de/files/test10_remis-Kreuz-R5.html
Due to the pairing algorithm, during the first five rounds, from the block of players A, C, E, G and I, all have played against the block of players B, D, F, H and J. Now all players have be to paired against opponents from their own block, which is not possible. When we do A-C and E-G, we have no opponent for I.
When we do the same with 8 players, we don't run into this dead lock, so we get 7 complete rounds with all games drawn: http://www.isewase.de/files/test8_remis.swt
@kleineme Ich schreibe mal auf deutsch! Ist gerade einfacher! Deine Spieler haben keine Wertung also keine DWZ oder Elo. Da funktioniert es natürlich! Das Experiment war aber eindeutig! Lasse die Spieler mit der besseren Bewertung gewinnen. Wie hast du die Swiss Datei hochgeladen? Ich kann dir meine schicken!
Und zwar mit jeder Runde! @kleineme
|> And you still haven't shown me the crucial passages in the Handbook of FIDE § C. 04! Can't wait to read it!
Sometimes I have to play a couple of games ;-)
It's not as easy as to point to one single paragraph, you have to read C.04 as a whole.
You have to start with the basic rules in C.04.1, especially d to g. The concrete rules for the Dutch system are laid down in C.04.3. There are three absolute pairing criteria, see C.04.3, C.1-C.3, which are not be violated (and even for those there are exceptions). Additionally there are a couple of so called Quality Criteria, see C.5-C.19, which can be discarded when you else don't come to a pairing.
Wie kann man da swiss dateien hochladen? Das ist ein DWZ Rechner. Ich könnte dir so meine Datei schicken. @kleineme
For our purpose it's irrelevant whether you create the initial ranking list based on Elo or not. I'll send you a PN.
History repeats itself!
Currently, this topic shows a development very similar to:
What happened in the above-mentioned topic?
1- Users asked the Lichess developers to add Swiss tournaments to the platform.
2- The developers and their associates published comments arguing that Swiss tournaments were not feasible in Lichess.
3- A growing number of users (in increasingly overwhelming amounts) insisted that the arguments of the opposers were just unfounded excuses.
4- The developers of Lichess remained silent.
5- Popular pressure continued to grow. The time came when the topic got 11.252 views and 208 comments. The vast majority of these comments were pleas from users who wanted Swiss tournaments in Lichess.
6- After an exhausting debate (and a lot of time wasted), the developers of Lichess finally decided to listen to people's clamors and implemented the Swiss tournaments. Result: many happy users.
At this moment, this topic is in transition to phase 3 and 4. We would save a lot of time and headaches if the Lichess developers decide to jump straight to step 6.
Please, Mr. Thibault Duplessis (@thibault) and the Lichess development team, reconsider your position regarding the Round Robin tournaments. I know that you are very busy with many projects, but the professionalism that you demonstrated by implementing Swiss tournaments so quickly shows that Lichess is up to any challenge.
I trust you and the words you initially posted on lichess.org/swiss:
#49 There's a difference here: I caused a ruckus advocating for Swiss pairings and dare not repeat that. Other developers' time isn't mine to offer, and I severely doubt anyone could find a solution which scales well.
I was convinced that Swiss was a good idea; I am convinced that Round Robin is a terrible idea. Absent very strong evidence (i.e. working tests and working code) it is very, very unlikely my mind will change in the foreseeable future.