Voir les contributions

Cette section vous permet de consulter les contributions (messages, sujets et fichiers joints) d'un utilisateur. Vous ne pourrez voir que les contributions des zones auxquelles vous avez accès.


Messages - Lafungo

Pages: [1] 2 3 ... 50
1
Règlement / Rule changes for 2024
« le: 07 juin 2024, à 18:22:26 »
Message from Sjors.

Citer
The forum won't send me a password reset mail, so here's my input for the rule changes discussion:

Question 1: What group stage format should we use?
I'd say let's give round robin at least a try. The thing that sort of irks me about the Swiss system is that you tend to play the same people year-on-year. Which isn't a bad thing in itself but variety is the spice of life as they say.
I feel round robin worked well for JSMKC is that the group was smaller than as it regularly is at SMKC. So, I do feel that proper seeding is key in making it work because I can already see people complaining they're placed in the pool of death. Also, yay name tags!

Question 2: Should BM/MR group stage matches change to 6 rounds instead of 4?
No real opinion here. Mo' play, mo' fun. I also like Moll's idea of making it an odd number to prevent ties. However I think we should really be careful by not throwing away all the time we won with going round robin by extending group stage (especially in BM!). We're not 16 anymore (well, most of us) so keeping the days shorter is also a concern. As for BM, just cycling through the maps as the first map and go from there seems the best way to go about it.

Question 3: What policy should we adopt for missed zooms in TT KO stage?
Completing a run and then wanting a do-over is a no-go in my mind. It is rife for abuse. Especially by more assertive players.
What we do need is a single arbiter that handles the starts. Having two people do it was in my mind mostly the reason for what happened at JSMKC.
It is also important to note that the arbiter giving you the go-ahead is mostly so you don't interrupt the previous player. So, when you get the go ahead you can take a bit of time to check your surroundings, wait a bit if need be, and make sure you are happy before you go. Everything beyond that is on you. If something were to happen during a run that could hinder (multiple) players (e.g. door slamming shut) everybody should be giving the option of keeping their current time as it is on screen or play again.
Also, as someone who was bitten by this myself: nobody but active players should be allowed near a console. I had a higher level player who wasn't yet in LKS practice a few TVs down while I was doing my barrage. That was partly on me, I could have asked them to not do that but there is also the risk of excited friendly (BM, MR, GP) matches disrupting runs.

Question 4: Should a bracket reset impact the point allocation for the players in the final?
Yes. Bracket reset should have influence on the points.
The amounts of points I have no clear opinion on as the whole scoring is voodoo to me so I cannot determine what is fair.

Also, small addition after reading through the forum: those in the know will not be surprised that any measure that disincentivizes trolling in GP between imbalanced players has my automatic upvote.
Higher skilled players humiliating lower skilled players for whatever reason, and everybody going "[shrug emoji] well, that's just how <insert_name> is" is very much at odds with how the community sees itself as being open and welcoming to new players and where there is space for everyone.

2
Règlement / Rule changes for 2024
« le: 04 juin 2024, à 06:53:28 »
Question Is it possible to discuss of the tiebreaker in GP ?
Go ahead. What do you suggest?

3
Règlement / Rule changes for 2024
« le: 04 juin 2024, à 06:52:48 »
Message from Sumner.

Citer
The only one I really want to get an answer in for is option 3 (restarts in TT)
You can not let the person finish then complain, by the time I was entering lap 5 at JSMK I already had looked at Onwas time while my race was still running.
Only option for this I think is the person telling people to start needs to just wait and check the start goes well then moves to the next person, once they move to the next person you can not get a restart

4
Règlement / Rule changes for 2024
« le: 04 juin 2024, à 00:10:28 »
Question Moll: Can it be made clear if match score is important or not, it seems to be different every year and I never have any clue if a 4-0 win is better than a 3-1 or just the same
In case of Swiss: point differential is the second tiebreaker, when two or more players are tied on both WDL and Buchholz scores. Tied Buchholz scores are infrequent, so point differential will only impact a small number of players.
In case of Round Robin: point differential is the first tiebreaker, when two or more players end up with the same WDL score. WDL ties are common, so point differential will impact most players.

I'll be sure to make it clear one way or the other during rule announcements at CDM.

5
Règlement / Rule changes for 2024
« le: 03 juin 2024, à 23:38:49 »
Here are my personal responses to the above questions:

Question 1: What group stage format should we use?

Round Robin.
I think that the logistical advantages significantly outweigh the disadvantage of fewer meaningful matches. The use of match cards at JSMKC was a great success, and I think it demonstrated clearly just how efficient Round Robin can be, which also means that we can accommodate larger groups. Another important point is that it will (finally) allow Patrick to participate as a player again without having to worry about scorekeeping.

Question 2: Should BM/MR group stage matches change to 6 rounds instead of 4?

6 rounds.
I think this will help make the outcome of meaningful matches be and feel more skill-based and less luck-based. I don't think that the added time will cause any scheduling issues.

Question 3: What policy should we adopt for missed zooms in TT KO stage?

The option of having the player immediately raise their hand without playing the attempt is the option I like the least. I think that it puts undue pressure on the staff member involved to grant a retry, as the alternative is a forfeit for that track.
As for the other two options, I would prefer the one in which a retry is possible but the staff member involved can still deny it without having to hand the player a forfeit. I think this achieves a better balance in terms of policy. That said, I would be OK with the option of not allowing players to retry at all due to noise interference, because that way everyone knows what's up and there's no room for discussion (although there can still easily be salt and controversy from missed zooms due to noise).

Question 4: Should a bracket reset impact the point allocation for the players in the final?

No.
I think that what should matter in terms of points is whether you got gold or silver, and not the path you took to reach that result. I also think that it would be very tricky to establish some kind of corresponding notion for TT, which would lead to an awkward outcome no matter what.

6
Règlement / Rule changes for 2024
« le: 03 juin 2024, à 23:38:30 »
Once again, I'm opening a topic to discuss potential rule changes for CDM. I would have liked to open it earlier (as was the case last year), but I spent several months in the lead up to JSMKC in regular contact with Kasmo to iron out the ruleset there. As you'll understand, I wanted a break after the event and decided to give myself the month of May as time off before broaching the subject of CDM rules at the beginning of June. I've summarized the points that I want to discuss in a few key questions, and I'm hoping that people will participate promptly.

To avoid issues from previous years, I am setting a firm deadline for responses to be taken into account for the 2024 ruleset. The deadline for feedback/opinions is July 31. Nevertheless, please respond to this topic as early as you can. I will then make adjustments to the 2023 ruleset as needed and post the full 2024 ruleset as soon as possible in August.

Below are the points I would like to discuss for this year, based on my own observations and feedback I received from other participants. They are ordered from most to least important/impactful.

Question 1: What group stage format should we use?
This is the big question for this year, and there are two high-level options: Swiss (as has been used for the last few years) or Round Robin (as was used until 2013, and is used at the NTSC championships). Below are some considerations for both.

Swiss:
The main appeal is that most matches played are between opponents of similar level, which is a huge benefit. The format is also quite easy to extend, in that we can add players up until the last minute with essentially no impact on the running of the tournament.
However, there are some pretty significant drawbacks, mainly with regards to logistics. One major drawback is the time spent between rounds: because the next round of match pairings depends on the results from the previous round, all players must wait for the slowest match to finish. Then, the scorekeeper needs to finalize all results for that round, generate the next round's pairings, and announce them to all players. From my observations over the last few years, this equates to around 10-15 minutes on average of downtime between the end of a typical match and the start of the next round, which adds up to over two hours' worth of accumulated time over the course of a day.
Another important consideration is the scorekeeper. Patrick has been the sole person handling this task for years now, and has had to sacrifice the majority of his participations as a player in order to run these single-group formats (initially Belgian and now Swiss). As long as we continue to run such a format, having someone (presumably still Patrick) sacrifice their week for scorekeeping will remain a necessity.

Round Robin:
The main appeal is that the format is simple and straightforward, both for players and logistically. In addition, with proper preparation it's possible to speed through the group stage with minimal downtime between matches. This was demonstrated at JSMKC, where match cards were given to each player that listed the order of matches they were to play, and they could then chain matches without having to consult any member of staff. This means that the scorekeeper does not need to do anything during the group stage itself, and thus is free to participate as a player as well. Finally, the distribution of players into groups means that everyone gets to play against opponents of all levels, which has been a point of contention that has been voiced by a number of players over the years.
The major downside of this format is the limited number of meaningful matches, i.e. matches between players of similar skill level. This means that a player's ranking may realistically depend on only a small percentage of their matches, with the rest being "automatic" wins or losses. Furthermore, there can be a concern with regards to the balance of the groups, which depends on an accurate assessment of participants' skill prior to the start of competition. This effect becomes more prominent as the number of groups increases. An additional consideration is that in order to prepare the groups properly with match cards, the groups themselves must be determined in advance so that the match cards can be made. This means that there needs to be a cutoff point after which new registrations aren't accepted, which also means that registrations on the morning of a given mode wouldn't be possible.

If we were to go with Round Robin, we would also need to decide on the number of groups. I think that the only sensible approach would be to vary the number of groups between 1 and 4 depending on the number of participants, but the difficulty is in deciding on cutoff points. An example proposal would be something like this:
- 1 group for up to 24 participants, and no barrages
- 2 groups for 25 to 40 participants
- 3 groups for 41 to 47 participants
- 4 groups for 48 to 64 participants
The reason for having such a small (example) window for 3 groups is that an added layer of complexity would be needed for the progression to KO stage. For 1, 2, and 4 groups, a systematic progression can be used where each player is placed in KO stage based only on their rank within their group and which group they were in. For 3 groups, such an approach is not possible, so we would need to decide on a method for establishing the progression. The base solution (as proposed by Narnet last year) would be to order players first by their rank within their group, then by comparing their records to the other players with the same rank in other groups (so first comparing WDL score, then point differential).

Question 2: Should BM/MR group stage matches change to 6 rounds instead of 4?

The main argument for change here is that matches between players of similar skill can feel like little more than a coin flip over only 4 rounds, and it's not uncommon for the outcome of a match to be determined by a single lightning or a couple rogue greens. Given that a significant chunk of time during group stage is spent waiting between matches rather than playing said matches, the time cost of lengthening all these matches should be relatively small. That said, there is still a time cost involved and it's worth considering whether it would be worthwhile. It would also constitute a change for an aspect of the championship that hasn't been touched in over 20 years.

Another thing to consider is how we would decide the courses to play on for each match, in particular for BM. A couple options that come to mind are to either have all matches start on Map 1, or have the starting map be randomized. If randomized, I would suggest that we do it the same way as cup selection in GP, so for example in the first four rounds of matches each map would be selected as the starting map exactly once.

Question 3: What policy should we adopt for missed zooms in TT KO stage?

At JSMKC, we had a couple instances of players being granted a retry after missing their zoomstart due to external noise interference and complaining to staff upon completion of their attempt. The responses from different people regarding these incidents made it clear that we need a concrete policy for handling these situations, as they are not currently covered by the existing ruleset. Here are a few options based on different people's responses:
- in case of noise interfering with a zoomstart, the player should immediately raise their hand to signal a staff member and not attempt to drive the course at all. The staff member then decides whether the player is granted a retry or not. If not, they are given a 9'59"99 for that track.
- in case of noise interfering with a zoomstart, the player should complete their attempt, then signal a staff member without looking at any other player's results. The staff member then decides whether the player is granted a retry or not. If not, the player keeps the time set in their attempt.
- players are not allowed to retry in case of external noise interference.

Question 4: Should a bracket reset impact the point allocation for the players in the final?

The idea behind a potential "yes" answer to this question is fairly straightforward: there is a significant difference between a player winning gold without losing any matches in the KO stage, and a player winning gold despite having lost a match (whether it's from a bracket reset or from earlier in the upper bracket).
Conversely, a potential "no" answer indicates that what really matters is getting gold or silver, and not the path taken to reach that result.

In case of answering "yes", a decision would also be needed for how exactly to change the point allocation. One potential example would be as follows:
- no bracket reset: 2000 points for gold, 1600 points for silver.
- bracket reset: 1900 points for gold, 1700 points for silver.

7
Le Championnat Super Mario Kart / Time Trial Anti-Records (ARs)
« le: 21 janvier 2024, à 12:41:44 »
With the site update done (thanks again ScouB!), it's also time for the 2023 AR update!

MC11'13"21flo2332013
DP11'25"23ZTI2010
GV11'28"39Conor2015
BC11'52"75Sumner2023
MC21'42"70Jarmou2017
CI11'04"07Gatchan2013
GV21'22"50Gatchan2012
DP21'44"67ZTI2008
BC22'08"25Drew2011
MC31'49"59Drew2021
KB11'09"59Narnet2023
CI21'16"95Lafungo2016
VL11'02"90KVD2014
BC32'15"71Harold2009
MC41'49"60MJ2021
DP31'38"66MJ2009
KB21'23"12Patrick2021
GV32'00"39BPA2009
VL21'17"92MJ2018
RR1'44"67Jarmou2017

MJ3 ARs
ZTI2 ARs
Jarmou2 ARs
Gatchan2 ARs
Drew2 ARs
flo2331 AR
Conor1 AR
Lafungo1 AR
KVD1 AR
Harold1 AR
BPA1 AR
Patrick1 AR
Narnet1 AR
Sumner1 AR
Geo0 ARs
Zarkov0 ARs

Some notes:
- The KB1 AR demolition derby from last year was extended this year by none other than Narnet, who managed to qualify for Top 16 through both rounds of barrages after finishing #21 in group stage. Narnet's dedication to not learning how to slide has paid off, and contributed to his success in joining the illustrious group of AR holders at Zarkov's expense!
- After getting outdone in his first ever TT Top 16 track in 2022 (by a mere "40), Sumner returned with a vengeance in 2023 and qualified once again for TT Top 16. He then converted into a massive AR on his final of three opportunities, demolishing the previous BC1 AR by over 10 seconds!
- The weakest AR, VL1, remains unperturbed for nearly a decade now, in large part because it never seems to show up in the early stages of Top 16. For reference, in 2023 VL1 was the first track of Q1 barrages, and the current AR would have ranked #2 in that round! Perhaps the Kart Gods do not wish to take KVD's greatest TT accomplishment away from him?
- MJ remains as the uncontested King of ARs. Will anyone be able to take him down?

8
Règlement / Idée de règlement pour 2024
« le: 11 septembre 2023, à 11:29:05 »
Rien à redire sur ton post, je suis évidemment d'accord avec l'idée qui y est émise.

Pour info, je ne prévois actuellement aucun changement significatif du règlement pour l'année prochaine, notamment vu que je n'ai eu aucun retour négatif à ce sujet au CDM 2023 (hormis les problèmes évidents de temps perdu). Je modifierai le règlement officiel pour incorporer les changements de dernière minute, mais on peut parler dès maintenant de potentiels changements pour 2024 si vous avez quelque chose à dire.

9
Règlement / 2023 Ruleset / Règlement 2023
« le: 10 septembre 2023, à 12:58:27 »
En principe Patrick devrait partager le fichier ce week-end (après ajout des points pour/contre en poules de GP pour aider ScouB avec l'encodage sur le site). Ça m'intéresse aussi de l'avoir donc je te le passe via Discord dès que possible.

-----

For posterity, here are the last-minute rule changes that were implemented in 2023 (all based on the earlier posts by Narnet in this topic):
1. no modified pairing for Rounds 11 and 12.
2. pairings done in the standard Swiss way: instead of 1 vs 40, 2 vs 39, etc. for Round 1, the pairings were 1 vs 21, 2 vs 22, etc. This logic was applied in later rounds, grouped by players with the same WDL score.
3. Goal average / point differential was no longer used for pairings at all.
4. The primary tiebreaker for the final group stage rankings was changed from goal average / point differential to truncated Buchholz. This means that each player's tiebreaker score was the sum of their opponents' WDL scores, with the worst one removed.
5. The small modifier applied to players' overall points obtained from each group stage based on their point differential was removed. Thus, the amount of points received for the overall ranking from 2P group stages was based exclusively on WDL scores.

Unfortunately, a lot of these changes required manual work during the event, which led to some significant delays. It also means that there might have been some mistakes as one point or another. These issues should be fixed by next year, as Patrick will have plenty of time to implement the changes programmatically by the time CDM 2024 rolls around.

One last thing to note regarding this year's implementation: we ran into an issue for Round 12 of BM, as the file got stuck searching for match pairings. We didn't have time to figure out the root cause during CDM (and this pairing couldn't have been done manually in a reasonable time frame), so we improvised on the spot and made Round 12 pairings generate in the same way as the Round 1 pairings. This issue did not occur in either MR or GP, so we had proper Round 12 pairings for those modes.

10
Règlement / 2023 Ruleset / Règlement 2023
« le: 22 août 2023, à 15:01:15 »
Pour commencer, ça fait plaisir d'enfin avoir un retour concret sur le règlement. C'est beaucoup plus facile d'améliorer les failles du système si on ne s'y prend pas seul.  :)

Je sais que j'ai merdé l'année dernière sur l'implémentation des J11 et 12 (je suis et ai été le premier à le reconnaître), mais je ne suis pas forcément d'accord avec tous les arguments que tu présentes. Ceci dit, vu qu'on est à moins d'une semaine du CDM, je ne vais pas entamer un débat sur les détails présentés.

À la place, je vais me concentrer sur l'aspect pratique de la discussion pour le CDM 2023.

D'abord, vis-à-vis du système belge: je mantiens qu'un retour au système belge n'est pas une option. Peu importe sa compléxité ou ces mérites théoriques, il y a des gros problèmes pratiques avec son utilisation. Entre autres, si le système casse personne ne peut le règler, personne peut valider que les résultats produits soient corrects sur place, le fichier est lent et à tendance à crasher, etc. De plus, la seule personne qui est en mesure d'opérer le fichier du système belge (Patrick, qui a passé de nombreuses heures à essayer de comprendre comment marchait ce fichier et a dû régler des bugs de lui-même) ne veut plus rien à voir avec ce fichier et ne s'occuperait plus du tout de cet aspect du championnat si l'on décidait de revenir sur le système belge.

Je pense que les options pour les poules de 2023 sont les suivantes (sans ordre particulier):
- système suisse avec J11/12 corrigés
- système suisse sans appariement modifié en J11/12
- poules "traditionnelles" (Round Robin) avec 2 ou 4 poules (vu le nombre d'inscrits actuellement, avec 2 poules on risque d'avoir des problèmes de planning mais avec 4 poules ça ferait peu de matchs)
- autre modification du système suisse qui puisse être implémentée facilement et en (très) peu de temps

De mon point de vue, s'il y a une préférence pour le système suisse sans appariement modifié, alors OK on fait ça. C'est l'option la plus simple en termes d'orga de toute façon. Mais en tout cas il faudrait décider dès que possible.

11
CDM 2023 / 2023 CDM Predictions
« le: 18 août 2023, à 15:26:45 »
At least we can all agree that Jarmou is getting 8th overall.

12
CDM 2023 / 2023 CDM Predictions
« le: 15 août 2023, à 18:28:04 »
TT
1 Karel
2 Mario86
3 Sami
4 Lafungo

BM
1 Geo
2 Neo
3 Drew
4 Sami

MR
1 Sami
2 Takashi
3 Mario86
4 Lafungo

GP
1 Neo
2 Sami
3 Takashi
4 Geo

Overall
1 Sami
2 Neo
3 Takashi
4 Geo
5 Mario86
6 Karel
7 Lafungo
8 Jarmou
9 Leyla
10 JDR

13
Règlement / 2023 Ruleset / Règlement 2023
« le: 14 août 2023, à 20:36:51 »
Citer
comportant 12 tours ... À partir du 11e tour, les paires sont établies à partir de la 20e place (ou de la place médiane s'il y a moins de 40 joueurs) dans le classement provisoire vers l'extérieur, les joueurs de la partie supérieure du classement provisoire (20e et plus) étant jumelés avec des joueurs de la partie inférieure.
C'est quand même très vicieux de prévoir 2 tours de ce type à la fin! Ce 11ème match est hyper (trop ?) crucial pour les joueurs de milieu de classement!

Imaginons un 18vs22 qui est un match serré. 18 gagne/22 perd => au 12ème tour 22 à perdu des places (disons 24ème) et ne pourra pas se refaire comme le permet le système Suisse car devra affronter en 12ème match un joueur encore mieux classé exemple 16 vs 24...
A l'inverse si 22 avait gagné contre 18, il aurait grimpé, 19° par exemple pour avoir en 12ème match 19vs21 ... ou avec un peu de chance un 19vs24 par exemple si déjà joué les autres joueurs

En soit je vois dans cette idée un dernier réajustement pour les dernières places qualificatives mais en faire 2 tours me semble inutile. J'aurais continué pour un 11ème tour de système Suisse et un 12ème et dernier tour avec les appairage depuis la médiane
Il est certainement possible que les détails puissent être ajustés pour paufiner le système. C'est pas pour rien que j'essaie de rendre le règlement le plus transparent possible et que j'attends des retours depuis plus de deux ans.
Ceci dit, je pense qu'il est trop tard pour discuter de changements supplémentaires pour 2023 (à moins qu'il y ait vraiment un gros problème qui a été raté jusqu'à maintenant). On pourra parler de comment améliorer le système pour 2024 si besoin.  :)

D'accord ; et une dernière question, histoire d'être sûr : avec ce critère, à la toute première journée, si on a 42 joueurs au départ (comme en GP l'an dernier), on a 1 vs 42, 2 vs 41, 3 vs 40 etc ? (comme l'an dernier).
Oui, c'est ça.

Et à la 2eme, si on suppose que seuls les joueurs du top 8 (têtes de série de 1 à 8 ) ont gagné 4-0 au 1er tour (en MR ou en BM), on aurait : 1 vs 8, 2 vs 7, 3 vs 6 et 4 vs 5 ? (alors que l'an dernier, on aurait eu 1 vs 2, 3 vs 4, 5 vs 6 et 7 vs 8 ?).
Oui, c'est ça, sauf que c'était déjà le cas l'année dernière (et en 2021).

14
Règlement / 2023 Ruleset / Règlement 2023
« le: 11 août 2023, à 12:45:10 »
Je pense que tu mélanges un peu les pinceaux.

Dans le message auquel tu réponds, je parle des tirages en poules. De part la nature du système suisse (ainsi que dans le système belge utilisé avant), les top players vont forcément s'affronter tôt ou tard puisqu'ils vont gagner la plupart de leurs matchs et se retrouver en haut du classement provisoire (qui est utilisé pour établir les tirages).

Dans ton message, tu as l'air de parler des phases éliminatoires et le fait que l'arbre arrange les matchs de telle sorte que les mieux classés en poules ne s'affrontent pas jusqu'à plus tard. Cette logique est utilisée dans une grande majorité de compétitions à phase éliminatoire, et pour bonne raison: c'est la façon la plus juste d'établir les tirages.

15
Règlement / 2023 Ruleset / Règlement 2023
« le: 11 août 2023, à 00:18:09 »
Pour le premier point, c'est uniquement en cas d'égalité de points VND et de goal average général que la liste de niveau est utilisée pour départager lors des tirages. D'ailleurs, il n'y a aucun changement cette année: il s'agit du même système depuis 2021.

Pour le deuxième point, l'idée derrière reste la même que l'année dernière, mais en corrigeant l'erreur que j'avais commise. Dans l'implémentation utilisée l'année dernière, l'appariement commence en essayant 20 vs 21, mais si ce match a déjà joué on passe à essayer 20 vs 19, 20 vs 22, 20 vs 18, etc. Cette implémentation fait qu'il est possible d'avoir des matchs qui ne suivent pas du tout la logique voulue conceptuellement, comme par exemple le Geo vs Jarmou qu'on avait eu en journée 12 de GP. Cette année, c'est corrigé et les appariements essayés seraient 20 vs 21, puis 20 vs 22, 20 vs 23, etc.
À 42, et en supposant aucun match déjà joué, on aurait 20 vs 21, 19 vs 22, ..., 1 vs 40, et 41 vs 42.

Pages: [1] 2 3 ... 50