I will accept that that was your intent with it, but the scenario you posed was still something that I claim is unrealistic, as the open game offer would not be around after 5 minutes, nevermind 20. If the times are irrelevant, you could use 30 seconds of waiting as an example, but by choosing long times for your example, it makes the situation appear much worse than it actually would be practically
2 Likes
Why? What is your problem with CG so that you don’t ever want to use it?
1 Like
Not OP, but I think there are valid reasons to not want to use the CG option. Mainly it’s additional mental overhead to both create and accept games. There are so many options you have to be aware of, and you only have a limited time to pick up a game before someone else takes it.
4 Likes
Oh man the life is so short that you can’t tell what you want? Instead of picking, propose.
If we had problems waiting in CG then OK we could rethink the situation but it doesn’t seem to be like this
2 Likes
Creation of CG is also harder than AF. Saved settings helps, but it’s hard to beat one button click.
3 Likes
and there are default settings anyway so that you may just pass through some of them
that can beat the one button click
In the deepness of my heart (lol) there is no valid reason to be that lazy or pushed by time available. That push the button culture emerging from modern scholar programs is a nightmare if you think a bit about it. OGS is progressing on a different path, making it more enjoyable at each step. For example the revamp of the handicap system including small boards. Small boards games are now better integrated into the ranking system and that’s a wonderful specificity of OGS. And that’s not the only new development, I’m quite excited by all these. But now trying to standardize the game through an overwelming automatch, that’s not the way I think to follow. I use to propose Chinese, so I will, same time I’m fine with a NZ or a OGS japanese like, rules. I always found the time settings proposed for normal games in automatches on other servers to be less convenient to me, prefering a real blitz or a longer time. Automatch settings are most about a kind of semi blitz which I was not at my best. Why do we have to be made of the same mold? Save the diversity!
And here i should copy paste all the absurdity I read before (I won’t).
3 Likes
It was proposed in another thread a button to “reserve/claim” a CG offer replacing the button to accept it directly, which would prevent anyone else from accepting for a limited time (like 30 seconds tops). During this time, the person who made the CG offer would see something like “<username> is considering your challenge (time remaining: <countdown>”, and the person who reserved/claimed it would see a popup with all the settings of the game displayed, along with the same countdown, and buttons do either decline or accept. If the timer runs out first, the person who reserved/claimed the challenge is treated as having declined it, and the challenge reopens for other people to reserve/claim. Ditto if the one who reserved/claimed it clicks “decline”. If they click Accept, the game begins
This would remove the perverse incentives to accept CG offers as fast as you can click (and then cancel if you really don’t like the settings, which essentially works like the proposal if you’re accepting, but is far more annoying if you’re proposing offers, as now you have to remake the offer again), as once you click, you get to slow down and evaluate if you’re really okay with the settings without worrying about other players accepting faster
That said, I’ve never had any issues with just making CGs myself, and letting other players deal with the pressure of racing to accept
2 Likes
Great question. Let’s talk about the advantages of AF over CG.
1. Consistency
Players know exactly what to expect from a game mode, reducing confusion and ensuring a uniform experience.
2. Quicker Matchmaking in all scenarios
Standardized settings mean faster matchmaking, as players are grouped into popular game modes quickly without the need for complex filtering.
3. Competitive Integrity
Standardized settings help maintain a level playing field, which is crucial for competitive play and ranking systems.
The randomized matchmaking [also 4.] makes it more difficult for players to collude or manipulate outcomes, preserving the integrity of the competitive environment.
Basically when players face a broader array of opponents, then rankings in general tend to be more accurate and reflective of a player’s true skill level.
4. Random Pairing
AF uses an algorithm to randomly match players, so that players encounter a wide range of opponents. This helps preventing scenarios where the same players consistently play against each other.
It also encourages interactions across a wider segment of the player base, fostering a more dynamic and interconnected community.
5. Ease of Use
Simplifies the user interface by limiting the number of options players need to navigate, making the game more accessible to newcomers.
Most newcomers are immediately confused by e.g. the Rules option when they start playing on OGS. I’ve played on Fox for a while now and still have no clue which ruleset they use and I don’t care either to be honest.
So you prefer Chinese rules but you’re fine with NZ rules? Would you mind telling me in detail how exactly Chinese rules and NZ rules differ from each other when playing on OGS. I’m genuinely acutely curious.
Off the top of my head, Chinese Rules use 7.5 komi, NZD rules use 7.0 (that’s the main reason I use NZD rules: to allow draws)
- Chinese has compensation for handicap stones, NZD does not (another reason I prefer NZD, as it preserves the idea that a handicap stone is equivalent to a pass in this respect)
- Chinese forbids suicide, NZD allows it (yet another reason I prefer NZD, as allowing suicide is simpler and has some interesting implications for ko fights)
There are probably some others, but that probably covers the differences which are reflected in the OGS implementation
5 Likes
I think basically all the points you make are arguable or not very relevant.
First of all random pairing and competitive integrity are not things we really have a problem with, afaik. Ranking improvements are being considered and neither of those is up on anyone’s list.
Consistency is a bit of a wash - there is also some variability allowed in AF and as-is they are only visible once you are in the game.
Then most importantly, af is demonstrably slower than CG at the moment. So “faster matchmaking” can not be used as a rationale for AF. This is clearly way more subtle! Though to be fair, I also expect some form of AF to be faster for some/many users than CG. But it is not a given, details matter a lot.
And for ease of use, there are clearly people who prefer CG for their use so it’s obviously again a subtle question.
I can exactly put my finger on it, but it feels like you’re approaching this discussion like a high school debate and trying to make sweeping statements, rather than understanding the nuances and finding the best solution for the actual circumstances.
It feels like your ideas and observations are decent and any localized change you’d suggest I’d probably support. But if they have to be woven into this bigger narrative, then I think you need to write a product design proposal like a senior product manager would, to convince us of this bigger vision.
1 Like
Thanks a lot for the link and the explanation! I very much appreciate it.
I would argue that these differences are negligible and having the option causes more harm than good. Because of the points outlined in my last comment.
What a great conversation starter.
Interesting view point. Generally, I’d argue that OGS neither has random pairing nor competitive integrity per definition because the matchmaking approach chosen by OGS does not provide them.
If this in itself is a problem and how exactly that plays out on OGS is definitely something that is up to debate and might need some data analysis.
For me personally it definitely is a problem. I want a fair and uniform playing field when playing ranked matches.
And that matters because?
Yes, because the OGS AF implementation is bad. That should obviously be fixed as I’ve outlined several times in this very thread. This is kind of the point of this thread.
Yes, because the OGS AF implementation is bad. That should obviously be fixed as I’ve outlined several times in this very thread. This is kind of the point of this thread.
That some people prefer CG does have anything to do with the topic of Ease of Use. There are many people who prefer driving stick over automatic. There is even a meme going around on OGS comparing the OGS matchmaking menu to a control panel inside a plane cockpit. The thing is that the cockpit needs all those buttons (at least I’d hope so) while OGS matchmaking does not (in my opinion).
You are becoming more pleasant with each paragraph. I did not make any “sweeping statements”, I outlined in detail while the AF approach for matchmaking would be better than what OGS is currently doing. And by the way, every other Go server, Chess server or Online Game I’ve ever been on agrees with me on this. And I’ve played a lot of games over the years.
Now that’s what I’d call a plot twist.
1 Like
By “these differences”, do you refer to the 3 differences I listed in my comment as being implemented by OGS, or the differences which probably exist but aren’t (I think) implemented by OGS?
If the second, I agree that making my favorite rulesets slightly closer to their defined forms would be a negligible improvement, but I would disagree that it would cause more harm than good. I think it would be a nice QoL improvement to bring all the rulesets closer to their actual definitions, but the Rulesets I care about (NZD, and to a lesser extent AGA and Chinese) are already close enough to their actual definitions for me to not think it very important to make them even closer. I would much rather get something like Lasker-Maas if given the choice (or another rigorous Territory ruleset which supports playing out L&D disputes) added as a Ruleset option than bring NZD, AGA, or Chinese rules marginally closer to the already very close target (Japanese Rules I can see an argument being made that it would be worth it to bring them much closer to the J89 rules, using AI to aid in the hypothetical play required by those rules)
If the first, however, I disagree: I do not think that the 3 differences I mentioned between Chinese Rules and NZD Rules (komi, handicap compensation, suicide) are negligible differences. I would choose NZD Rules over Chinese Rules for even one of them, and all 3 is wonderful to have indeed
1 Like
So what if they do? This is typically done by friends who enjoy playing together and may use it primarily as a means of social interaction. Perhaps they used to play IRL and then one moved away, or perhaps they met on OGS and found they have a lot in common. Trying to stop their fun is a dictatorial, killjoy proposal that would make OGS a far less welcoming place.
3 Likes
Haha, yes. I was ironically referring to the other thread where you agreed the rest of us were all like junior product managers 
But a separate somewhat detailed proposal is probably the right thing for you to do if you are serious about promoting a whole set of changes together.
It would be much easier if your ideas could be disentangled into a series of smallish changes that would each improve upon the positive outcomes you enumerated above. Then we could imagine starting with the most obvious ones without necessarily committing to the more contentious ones (eg reducing support for CG).
1 Like
This discussion seems to have been pretty heated so I’m hesitant to add myself in, but I do fundamentally agree with Regenwasser’s point. I don’t have an exact percentage but even playing at peak times (evening in the U.S.) I fail to get an automatch game within a few minutes at least 1/4 of the time. I know this is partially just due to the smaller playerbase compared to other servers, but this seems like all the more reason to pool players into as few matchmaking queues as possible.
When I join the ‘Normal’ automatch queue all I want is a relatively even game that takes about an hour. What I want above all else is to get a game soon after I request one. I don’t care what kind of time control is being used or what the rule set is, it’s all Go. I would rather have one consistent set of options than have to figure out while time is ticking what the rules are at the start of every game. But maybe a quick pre-match summary would be enough to mitigate this issue.
I don’t personally want to babysit the custom game list to catch an appropriate game before someone else takes it. It’s a much better user experience to just join the queue and wait to be paired. I think that making custom games visible to automatch would be a big benefit here.
I don’t understand how any of the proposals in this thread would stop the fun of these hypothetical players. Friends aren’t starting games together by joining the automatch queue and hoping to be randomly paired. Your two players will always be free to set up custom games and play to their hearts’ content. But this is a totally different scenario than playing ranked automatch games. In this case it’s better for developing skills to play a wide variety of players rather than get paired with the same player repeatedly.
2 Likes
Why don’t you make your own custom game?
2 Likes
An interesting stat to understand this proposal could be understanding what the usage of AF vs CG for long term OGS users vs those who try the site a few times and never come back. What percentage of games do each group use each feature? How long were CG listings/searches opened for? What success rates (i.e., games matched and not annulled) for each group?
On the proposal motivation:
I think the main motivation is to make AF faster. The CG crowd seems pretty happy with their experience and seem to not want to change it. So ideally a change wouldn’t mess with the CG experience.
User types:
CG people don’t mind looking at the list and it even seems like some CG offers get accepted so fast some people feel they must race to accept them.
AF people want to press a button and get a game without having to look at a list, graph, or configuration page. The ideal AF experience should be press a button then get a game.
I think for beginners this is an ideal place for them to approach go as they won’t understand all the aspects of go rules. A smaller set of rules could make learning go easier and more pleasant hopefully retaining more visitors to OGS. Additionally for experienced go players visiting OGS, it seems other servers generally provide this one-button experience as a primary way to find a match fast and expect the OGS version to work the same. When confronted with long waits, many may simply conclude nobody plays here. Thus the AF is worth improving.
Approach?
So what if we can develop a CG listing linked to an AF search for the AF matching engine? When a user starts an AF search it proceeds as current, if not matched within xx seconds AF creates a linked CG entry with standard settings based on the AF search settings. This linked CG will appear on the appropriate CG list based on time control, board size etc. It will disappear if the AF search is cancelled.
This could benefit AF users by making their games for visible and hopefully have faster matching. It could benefit CG users by providing more games for them to accept (or entertain with games blipping around the UI while they wait).
I think the above matches the spirit of improving AF while preserving the benefits of both pools. AF people press a button, CG people get some more game proposals. Win win?
Further more complicated versions?
UI enhancements to visually differentiate AF posted listings from custom ones. Simplest thing could be the game name.
More advanced features could include creating multiple CG entries with varying game configurations if the AF user has few preferences. If AF wait time is longer, post even more diverse rules within the AF search preferences.
A button for a CG host to start a linked AF search with settings auto defaulted to the closest match to their CG. Similar benefit would be the CG gets auto deleted if AF is able to make a match.
2 Likes
You don’t understand because you either paid no attention to the quote I was responding to, or you deliberately omitted it from your response to me.
Here is the quote again, with the preceding sentence added for greater context:
What’s hard to understand? The author tells you he wants to prevent scenarios where players (who are often friends) regularly play each other. The whole discussion is about either forcing people into the AF by eliminating CG, or at least severely straight-jacketing the available specifications for CG.
2 Likes
I don’t speak for the CG crowd and would in fact prefer a better AF. But my main sentiment is exactly what you articulate here - let’s make AF work better than CG if we have ideas for how to do that, without predicating those improvements on shunting CG in some way and forcing people to use AF.
3 Likes
No hard feelings.
I appreciate and understand your general point. A huge feature proposal like this should be accompanied by a backlog of User Stories (or Issues as some might say).
PS: I would actually have to make several stories out of this one idea cause there are so many potentials that the story would’ve become way too big.
User Story: Simplify the AF menu (Part I)
Background
Currently the automatch finder clusters computer mode into the AF even though they do not fit together thematically.
Furthermore, it includes a wide variety of settings for the match. This causes the AF to not be plug and play but instead cofusing for users, especially new users, as well as fragmentational.
Those settings are currently split up into three different sections:
- A large button for the time setting
- Smaller buttons on the top for the board size
- A gear icon on the top leading to another dialog window with dropdowns
Finally, the largest most prominent icons give a very lacking indication of what the actual time setting is.
Motivation
- Move computer mode to the section below to increase the focus on the actual AF options
- Create single button options that do require nothing more then the click of the button to achieve a plug and play user experience as well as a what you see is what you get experience
- Reduce options that are practically irrelevant to the game of go to keep the available options overseeable, understandable and to avoid fragmentation
- Fix the naming of the buttons to be comprehensible so that users actually understand the game mode before clicking on it
Wireframe
Changes
- Settings are merged into one overview section without any additional dialogues or gear icons
- Buttons for the quick pairing allow for an exact understanding of the game mode
- The ruleset setting is removed. OGS should find agree on one mode for now. This feature proposal is indifferent to the choice.
- The handicap setting is removed. OGS should find agree on one mode for now. This feature proposal is indifferent to the choice.
- The time control setting is removed. Byo-Yomi should be used as is standard in the world of Go
- Rengo gets its own tile instead of being hidden in the OGS menu cockpit
- “Computer” mode is moved next to “Friend” mode where it thematically fits. “Friend” mode is the private setting that was previously hidden in the OGS menu cockpit
- “Automatch finder” is renamed to “Quick pairing” as this sounds better to the author of this story
4 Likes