Welcome to Smogon! Take a moment to read the Introduction to Smogon for a run-down on everything Smogon, and make sure you take some time to read the global rules.
I checked out the OSDT round earlier and found this post. True to Fc's word, his opponent hasn't logged on once after signing up for the tournament, and is in the 4-2 bracket. I don't really think I need to spend much time explaining why this is a bad thing.
I'd like to propose that inactive users be automatically dropped from long tournaments such as OSDT. Rather than letting someone who doesn't log into Smogon pull a streak of coinflips, we should seek to make sure only active players are making it through our tournaments. If a user has not logged into Smogon for three weeks during a tournament, they should be immediately dropped. Additionally, any user who has only activity losses or coinflips after four rounds should be dropped, with some subjectivity in this one.
For now, I don't think anyone who is actively attempting to play games should be auto-disqualified. It might just be because I can't think of a good point to cut off somebody coasting on activity wins, but I think if you are trying to schedule and getting a series of opponents who don't respond to scheduling or don't make the agreed-upon time, you shouldn't be punished.
As someone who did a lot of work on last year's OSDT and Masters and this year's OSDT (so, not speaking as a TD)
I'm in favor of this. I'm sure awyp can attest for last year how frustrated I was on multiple occasions with people who decided that contacting their opponent was not worth their time, sending their games to coinflips despite being an active individual matched up with someone inactive.
This is really only a problem with Swiss, and it's the biggest problem that swiss has. Can't stop two inactives facing each other, but in standard SE/DE tours, those folks get washed out by R3 at worst. In Swiss, there always seems to be one or two inactive accounts who somehow go positive off of coinflips from people who don't contact their opponents once losing once or twice. I'd love to have been able, as host, to just manually decide to grant a win to someone who actually logged on, even if they clearly just quiet-quit the tour.
Legit wondering, but is there a reason why we can't give a double loss in the case of a pair of no shows? Swiss should be able to handle that case perfectly fine, and that would auto drop these people with the current procedures in OSDT. We also already have cases where people play up or down, so uneven brackets shouldnt be a huge issue. I get coin flips for cases where both people tried and both people failed, but if no one tried, then double losses should handle these automatically.
That being said, regardless of the method, the proposal suggested by Arcticblast seems pretty common sense, and I don't see any reason anyone would be against it. I suppose someone could have been waiting for an Shakespearean run where someone who doesnt play the game makes it to finals only to lose to their fatal flaw of nonresponse, but any complaints of that nature would seem unlikely and moderately frivolous.
I find this thread very necessary and I fully support the suggestion made in the OP.
Last year there was a case that happened to me in Round 5 of BW Cup where I forgot to message my opponent to schedule until sunday. My opponent, however, hadn't been online for over a week and had lost in multiple tournaments to activity over the course of said week. Not only that, but they stated in one of their VMs a bit prior to that that they were dropping out from tournaments because something in real life had come up.
Despite having a message from the user themself in their wall stating that they were not logging in due to real life situations and multiple activity losses over the week prior, my game was coinflipped and I ended up losing said coinflip. The exact reasoning for this decision was, and I quote: "I've discussed this situation with the Tournament Directors, and the decision to coinflip the game will stand. In short, attempting to schedule nine hours before the deadline is not sufficient for an activity claim. Without any show of agreement on an extension, or a claim at an activity win without any post requesting it that doesn't go against precedent, the only option left is a coin flip."
I started a PM with Merritt and a fairy after that, and here's the response I got:
Despite the fact that I did not create a story about what was happening in my opponent's life (my opponent themself made it clear they were unable to play for the foreseeable future), what ended up baffling me the most is that the decision at hand opted to prioritize a completely tunnel visioned approach to the coinflip rule instead of properly addressing the situation.
The end result to all this? This decision directly influenced Classic playoffs, as my opponent went on to face Garay Oak in Round 6 and they didn't even reply to his scheduling VM. Garay went on to qualify for playoffs and I went on to not qualify based on this series. And just for the record, this person only came back to Smogon in january 2024, so over 6 months after stating they had real life stuff to deal with and wouldn't be able to play.
I'm not here to bash on the TDs or whoever was involved with this decision at the time, but I'd like to take this opportunity to reiterate what I stood by in my PMs with Merrit and a fairy when all of this happened - acitivity rules (and activity calls in general) need to be examined on a case by case basis, with a proper understanding that we should prioritize the competitiveness of our tournaments when applying them. These rules should be examined through teleological interpretation, to avoid decisions that end up contradicting what the rules themselves seek to defend.
For now, I don't think anyone who is actively attempting to play games should be auto-disqualified. It might just be because I can't think of a good point to cut off somebody coasting on activity wins, but I think if you are trying to schedule and getting a series of opponents who don't respond to scheduling or don't make the agreed-upon time, you shouldn't be punished.
I agree with the spirit of this proposal so I'd like to elaborate upon it:
How to define "inactive player"
Straightforwardly, I think login times shouldn't be used to judge activity at all. First, how do we know a user who logged in mid-week actually checked their messages, and didn't just misclick on their browser history? But aside from that, for the other players there is no practical difference between someone who never logs on, and someone who logs on to participate in other parts of Smogon but never bothers with the tournament in question. There shouldn't be any more expectation for these players to suddenly want to play in later rounds. Player activity should only be based on visible attempts to play their matches. Particularly this means I disagree with a fairy 's opinion that a player who logged on should be granted the activity win over their opponent who didn't, if it's clear they won't play anymore.
When and how to disqualify a player
If a player isn't intending to actually play the tour, it should be apparent very quickly, which is why I think just 2 rounds is enough. First round of absolutely no scheduling attempt is a warning (and also a chance to show evidence of PM scheduling). Starting from the second round of absolutely no scheduling attempt, that player is considered "serially inactive" no longer eligible for coinflip wins until they become active again. If 2 serially inactive players face each other there will still have to be a coinflip as normal, but this could be solved by manually changing matchups so this doesn't happen.
I think simply barring from coinflip wins is a sufficient way to drop someone from the tour. Players who explicitly say they are dropping out are already handled in basically the same way. I also like big pichu 's proposal for a double loss but it currently isn't compatible with our official Swisses sticking to power-of-2 brackets. That's not to say that can't be changed but it would be out of the scope of this thread.
And now the biggest question... Does it actually work?
It's intuitive that dropping completely inactive players will get more matches played in later rounds, but how much so? To answer this, we'll need to do some modelling.
For these simulations, I used 480 players, i.e. a 512-man bracket with a handful of byes. Byes were added to account for inactive players who might advance round 1 on a bye. 10%, or 48 of these players will never attempt to schedule any matches. In my experience hosting this is a fairly realistic rate for decently large Smogon tournaments.
In addition, each active player has a 10% chance per round to forget to do any scheduling. This is an intentional overestimate; forgetting to schedule just 1 round is significantly rarer than never scheduling any rounds, but I wish to examine the effects of the proposed policy on players who may just forget sometimes.
As long as both players attempt to schedule, the match is always considered played. This simulation makes no distinction between Alice actually beating Bob, and Bob not showing up to a scheduled time. I am only interested in matches involving players who are completely inactive.
All players are identical in skill and have a 50% chance to beat anyone else. This simulation does not account for a potential correlation between skill and chance of forgetting to schedule.
The 10% chance of forgetting to schedule is independent with any other day. In reality it is more likely that if a player does not schedule 1 week there is a much greater chance they have something going on and forget to schedule the next, or they silently quit.
In Swiss tournaments there is no guarantee players won't rematch in later rounds. This seemed too hard to implement and an irrelevant factor.
Both single-elimination and Swiss were tested. (Swiss assumes players get eliminated after 3 losses.) The current system was tested as a control case, as well as my proposed system. A few more experimental cases tweaking the number of rounds for the rule to take effect were also tried.
Here is the code for my model. For each test case, 100 tournaments were run, and the I measured the furthest round any inactive player advanced.
Elimination tournaments:
In 62 tournaments, no inactive player made it past round 2.
In 32 tournaments, an inactive player made it to round 3.
In 6 tournaments, an inactive player made it to round 4.
Swiss tournaments:
In 57 tournaments, no inactive player made it past round 5 (2-2 record).
In 41 tournaments, an inactive player made it to round 6 (3-2 record).
In 2 tournamnets, an inactive player made it to round 7 (4-2 record).
Elimination tournaments:
In 81 tournaments, no inactive player made it past round 2.
In 19 tournaments, an inactive player made it to round 3.
Swiss tournaments:
In 4 tournaments, no inactive player made it past round 4 (1-2 record).
In 92 tournaments, an inactive player made it to round 5 (2-2 record).
In 4 tournaments, an inactive player made it to round 6 (3-2 record).
Elimination tournaments:
In 59 tournaments, no inactive player made it past round 2.
In 41 tournaments, an inactive player made it to round 3.
Swiss tournaments:
In 1 tournament, no inactive player made it past round 4.
In 96 tournaments, an inactive player made it to round 5.
In 3 tournamenets, an inactive player made it to round 6.
Elimination tournaments:
In 1 tournament, all inactive players were eliminated after round 1.
In 64 tournaments, an inactive player made it to round 2.
In 34 tournaments, an inactive player made it to round 3.
In 1 tournament, an inactive player made it to round 4.
Swiss tournaments:
In 88 tournaments, no inactive player made it past round 5.
In 12 tournaments, an inactive player made it to round 6.
Conclusion
A limit of 2 inactive games gives a significant and huge improvement over the current rule in Swiss. Elimination was also significantly improved. Limits of 3 and 4 also give significant improvements, but I think 2 is the best, not only because it keeps the most inactive players from advancing, but also because it's the easiest to keep track of. After 1 round you can list the players who were totally inactive while also pinging them, then if they are inactive again you can put their name in italics for future rounds so everyone can see they are being dropped.
So, in summary, my proposal is: If a player makes absolutely no scheduling attempt in 2 consecutive rounds, they are considered "dropped" and can no longer win any coin flips until they become active again.
Is this too hard to implement?
The last thing I want to touch on is whether this is practical for the typical host. It adds another step to checking activity and yet one more parameter hosts must track between rounds. In addition, this means the state of a round is dependent on more than just the round before. This is the main drawback to the proposal, but I think the gain in tournament quality is worth it for Swiss tournaments. It is probably unnecessary for elimination tournaments, because as was already said, most inactive players will be eliminated naturally.
