Yes, this. To me using automatch means “I don’t really care that much about the game settings, I just wanna play”. If you care about settings go custom.
I only use automatch when I play and I require handicap disabled. I don’t like the idea of choosing a custom game from the list or waiting and hoping someone will eventually choose me. It’s much more comfortable to let the system handle it. 
Polling a sample of games from the API could at least give insight into what time settings usually are used. I don’t think the API shows whether a game was created via automatch though.
Of course there might be other ways the gather some data.
Till now i didn’t give my own opinion so:
Automatch for correspondance is like useless.
Custom games is enough.
Ultrablitz is very marginal i won’t propose it either
So 2 kind a blitz and a i want a bit of time should satisfy most. Like a 50s to 1mn30s/8 to10s and a 8 to 15mn/20 or 30s. Has to be discussed which value exactly.
I won’t dispatch more as this.
We could call them simply quick and slow
1 Like
Game title is “${black} vs ${white}”. The only other feasible way to get the players name in the title would direct challenges. This should be a very rare event.
It will be more difficult to extract handicap preferences from there.
Here is a OGS game dump: Can we get an SGF database dump? . As I understand it, the raw API results are stored in the archive.
Most games will go with the defaults. I have gotten the impression most don’t even know they can set their preferences.
BTW: New Zealand prefered with Fisher timing required, handicap enabled prefered.
Edit: I don’t think you need to identify the automatch games at all if you are interested in player preferences. Custom games will give an impression of what players with strong opinions prefer.
2 Likes
This is why I personally would like the automatch system to be improved, I think many people have a similar position to this.
In fact, I think the custom game pool and the automatch pool should actually be unified, but it needs to be done in a careful way.
Like, when you press automatch, that should at the very least appear in the graph, and if you create a custom game with “reasonable” settings that fit someone else’s automatch settings, you two should be matched.
And there should be an option to prevent it on both sides (an option to avoid custom challenges in automatch settings and an option to prevent automatch picking the game up when creating a custom challenge).
1 Like
These are good ideas, but this thread is about updating the home page. Let’s tackle one mountain at a time haha
4 Likes
Yeah, but that’s what I meant — changing the buttons in the automatch panel is pointless if it doesn’t reflect an actual change in the system, so we’re forced to tackle that first 
Only to some extent. You need to know how many buttons will there be (to some extent). The text on the buttons can be changed at a future time. If the design is resilient enough, even the number of buttons can be changed.
2 Likes
Or we could just keep them the same until someone shows that more buttons is better 
I have several points to make here:
- It’s more about “fast auto-match”, so what are the most popular combinations? For live games, which option is more popular: 10m+30sx3 or 10m+30sx5? That’s my question, we might have to base this decision on statistics of the actual games matched and played.
- Too many number of combinations might result in slower matches made.
- Oberservation from FoxWQ-1: many people cancel games after being matched, I don’t know why, maybe it’s the opposite scenario when too few number of combinations are available and people just cancel those which do not meet their expectations.
- Obeservation from FoxWQ-2: 20m+60sx3 is not a good option, because 60s is too slow, we can do the calculation: 60s x 200 moves = 3 hours 20 mins…I think the appropriate length of a game is about 30 min. to 1 hour, or plus and minus depending on preferences.
Thank you for drawing these up, really glad this is getting discussed again.
I don’t think there should be an automatch option to play without handicap. If you include it you’re doubling the number of pools people will be split between.
For good reason other servers atuomatch options either force handicap, or they have the server population to only search for +/- 1 rank with Komi adjustments (Fox / Tygem). I know this is a controversial issue on OGS, but I feel it’s either handicap only, options that split the pools (bad), or no handicap only (terrible).
Same issue with the time controls. Personally I’m also for having Fischer as the only automatch option, but it doesn’t matter as long as we just pick one. (Let’s also be cool and modern and use AGA rules cause why not).
For the button choices, I’d personally cut it down to only 4, something like this:
- Blitz 1m/10s(3) or Fischer equivalent
- Live 5m/20s(3) " "
- Classical 10m/30s(5) " "
- Custom
4 Likes
Chat on the homepage would be awesome! Would help OGS re-create that awesome feeling KGS had.
Combine that with a thumbnail of the top rated in-progress game in the bottom left and boom: kibitz culture upgrade.
4 Likes
I have to say that I’ve got to be the insane one here. 30 seconds a move being called “Classical” just sounds like absolute madness to me, but then again I have no experience with OTB tournaments, and if there’s no interest in slower games in the community of course I have to accept it.
(Unless you have a good historical argument, don’t call 30s/move Classical though.)
On the other hand, I have a few thought:
-
I think part of the consideration for choosing the automatch settings should also be “Which time settings are good for players, especially newcomers, to learn the game?”, to which my personal guess answer is “an average of almost a minute per move”.
-
One thing many of you don’t seem to realize is that, if the automatch speed buttons are radically different, it’s not going to split the current pool, it’s just going to allow more people to use the automatch buttons for their preferred speed setting.
If someone wants a slow game, like, cough, me, what they do now is create a custom game and cross fingers that whoever accepts their game has read the settings and doesn’t cancel the game or rush it.
With an automatch button, since it’s apparently a niche need I might have to wait quite a while for a match, but then again I was willing to play a slow game, so why should it be a problem.
Whooops, sorry for hijacking the discussion about the landing page and only talking about automatch options again 
That’s different from pro advices for amateur games, which is like 45mn at most and a shorter byo yomi.
Don’t want to start an argument here, but why would pro opinions about amateur games matter? Like, at all?
Edit: Before people point this out, I realized after writing this that part of a pro’s job is probably to teach, including amateurs (I think it’s the case in Japan, don’t know about other countries) – in that case they do have experience with amateurs that can give value to their opinions about what helps amateurs learn.
I do take in consideration their advices on the game but of course that’s not compulsory for anyone. Basically it’s in their interest that go develop and amateur play better.
The automatch button has to adapt to what the majority of players want. I don’t think that many players use 1 minute/move. FoxGo does have an automatch option with 20 minutes main time + 1 minute / byo-yomi period, but in practice, almost no one using that option spends more than 30 seconds of thinking time per move.
2 Likes
Well, at least hey come expecting a slow game. I imagine that option doesn’t often match someone who wants to play at 20 seconds a move with someone who wants to play at 45+ seconds a move.
I feel like you might, between the lines, making the statistical mistake of thinking that “average=typical”.
The distribution of preferred speed settings might be multimodal, with a peak around 15 seconds and a smaller peak around 30-40 seconds – this would pull the average closer to 20-25 seconds, but most players would be unhappy with a setting at the average point (especially when matched with a player with different preferences).
1 Like
It’s possible that many players are interested in 30 or 40 seconds per move, however I think 1 minute is too much for 99% of players. My own preferred time setting for a casual internet game is 20 minutes main time + 30 seconds per byo-yomi period. My opponents rarely play much slower.
Exception: I sometimes play rated FFG/EGF rated games on the internet, and the usual time setting is 1h main time + canadian byo-yomi of 15 stones/5 minutes.
1 Like