Scoring in Sudoku Tournaments

Sorry. I couldn't find a corresponding topic in the forum right now (from what I remember, it should be somewhere) and so I'm making a new topic. The problem is this: Generally (by default) and also at the online Olympiad of Thinking Sports for schools held last week, tournaments are played in the Swiss system and the number of rounds is 11. This is the case for all games (except sudoku). Because sudoku tournaments are usually not in a duel format. As a result, a problem (discrepancy or inequality) arises when calculating the rating. *) If, for example, in a chess, Russian checkers and gomoku tournament it is possible to earn a maximum of 11*42=462 points (if you have a much lower rating than everyone else and win all the games), but on average the tournament winner earns maybe around 200 points... then in a sudoku tournament (if the number of participants is large) it is possible to earn many times more points (because you earn points from all the participants you beat). For example, if we bring the Olympiad that took place to the level of an elementary school, the best players earned over 1600 points (81 opponents times about 20 points against each other...) and as a result, the current sudoku ranking has also been turned upside down (there are quite a few players who have played their way from a 1400 rating to a 2900 rating in one tournament...) Since previous sudoku tournaments have had a maximum of 20 participants, there have been no major problems so far, but this Olympiad showed in my opinion that something should be changed regarding the points earned from sudoku tournaments. As one idea, I would suggest the following: *) Inequality with other (duel-based) tournaments occurs when the number of participants is greater than 12. (In the case of 12 players, the winner of a sudoku tournament earns points against 11 opponents and this is basically the same as winning all 11 rounds in some other tournament). Everything is fine even if there are fewer than 12 participants.. Maybe in these cases nothing needs to be changed. However, if there are more than 12 participants, then a so-called bonus is earned for each "additional participant"... and this also creates a discrepancy. *) So in my opinion, the number of participants should be taken into account when calculating points for sudoku tournaments. If the number of participants is greater than 12, then the points are calculated similarly to the previous one (all participants and their ranking are compared - i.e. you still earn or lose points against each opponent), but the earned or lost points are divided as follows: new_points_earned=old_points_earned/multiplier, where multiplier=1 (number of participants-12)/11

For example, in a sudoku tournament held during the Olympiad (82 participants), instead of 1600, 217 points would have been earned, and instead of -1600, -217 points... And the current sudoku ranking would be much more adequate. There are certainly other ways to solve this problem, but this is just one option that I proposed. Because the current system, where you earn, for example, 2000 points in one tournament in a tournament with a hundred participants, is not quite normal. In this case, it should also be possible to earn 2000 points in a chess tournament with 100 participants, but this is not possible, because hardly anyone bothers to play a 100-round chess tournament.

my old rating 1302 after the tournament 2371

This is a perfectly appropriate remark, but in defense of the creators of the game, I can still point out the reason why we made such a system: When we first started tournaments, we had a system where basically all the points were reduced to 10 opponents, i.e. the points from each opponent were summed up, then divided by the number of opponents, and then multiplied by 10. This literally threw the system out of balance, because in each tournament the sum of minuses and pluses should equal 0. Unfortunately, when users with a rating difference greater than 750 points participated in tournaments in the past, at some point the balance was lost and the average of all points shifted in some direction (in my opinion, the average of all users may have been below 1500 points). However, in principle, such a system is wrong, because balance is important. I don't think I need to explain why here. This also led to the latest rating reform, where Sudoku ratings were reset - things were out of balance and a new approach was needed. The result is the current system, which really works in tournaments with up to 30 participants or in tournaments where users are already nicely distributed in the rating range of 500-3500 (then you don't get so many points from each opponent). Maybe the issue is that the formula still needs to be thought out and made normal. Fortunately, there is no need to reset the ratings this time, because everything should be balanced.

Agreed. Balance is important, of course. And in every tournament, the sum of pluses and minuses must equal 0.. But I don't understand how dividing by a constant (i.e. dividing each participant's points by a constant) can upset the balance of equaling 0? Maybe I'm misunderstanding some method of calculating points that is used?

We probably calculated the average difference in ratings, that's why it worked. Starting from the larger differences, the point change was 0, but the average of the ratings shifted. In fact, you have to calculate the points you get from each one and then find the average as a result. And at the end, you have to somehow divide the points that are left over (if the points don't go exactly to zero because of the decimal point - but it's possible that in theory they will automatically go to zero if you do all the calculations).

Yes. Due to rounding of points, there may be an error (but in the long term I think there should be no shift) in some tournaments you earn a few more points than you lose and in other tournaments it's the other way around.. but yes, you still have to think it through :)