Personal rank difference between correspondence- and live-games causes issues getting 'even' games

About a year ago Clossius started a topic about forced handicap in games (Link here).

In this thread I made the following comment:

Even though I have played more live games this year my rank unfortunately does not seem to get much closer to that of my level in correspondence games. Because of sudden drops in the latter because of playing more live games I also started playing on KGS, Fox and Pandanet to practise.

This situation got me thinking that the different ways to play perhaps are to be regarded as different disciplines, each with their own level. Just like in sports for example, a sprinter posesses a different skillset than a marathon runner. It would be silly to expect an even match between the two on any distance. It just seems that my mind lacks the skill to react as fast as is nescessary to keep up with the average settings of a live game (20m + byo yomi) while it does have the ability to perform better (like in correspondence).

Anyway, this situation bothers me. I cannot play more ‘different’ games on OGS because of this and I just like to play here very much! And to me it seems that OGS would like it if more games that suited it’s members could be played. I kind of hoped that with the latest changes made to accomodate players this issue would pop up but it hasn’t. I could be wrong by the way because after post / reaction number 50 or so I lost the interest to keep up.

KGS was supposed to be a go-to platform to play but the activity, or the lack of it, is surprising. Fox is a bit overwhelming and Pandanet has often more bots than opponents at my level. Both are fine but I like playing on OGS best. At first (almost 4 years ago) I felt intimidated but after a couple of months I got comfortable. I still play on DGS by the way, where I first started, even though (or maybe just because) the site is seriously outdated. I will celebrate my 55th birthday this month so that explaines that I suppose :wink:

I always try to check my potential opponent’s profile to see what I can expect and I’m pretty sure my issue is a concern (more or less) to others. I often see quite large differences in rankings between the time- and size of board levels.

Well, now it’s out there. I have often thought about how to bring up this concern of mine (and even whether I even should do that) but here it is.

My suggestion would be to split individual ranks between time- and sizes of board in matches so it would be more easy to find suitable opponents given the circumstances.

Any thoughts?

Kind regards

3 Likes

The simple solution is to use two accounts, one for live and one for correspondence. There are (at least) two types of correspondence players:

A. Those who spend a long time pondering each move, much more than in a live game.
B. Those who play moves in correspondence games when they have time, like in a supermarket, in public transport, etc. and don’t try hard to find the best move.

Players in category A are expected to have a higher rank in correspondence games, while players in category B may have a similar or lower rank in correspondence games than in live games.

The account I’m currently using for live games is 0.7k and the one for correspondence is 2.0d, so that’s a difference of 1.7 stones.

3 Likes

That’s indeed an option but it seems silly to have multiple accounts on one server.

3 Likes

Maybe there could be an option to queue in a particular time category, with the corresponding OGS rank, instead of the general one.

Screenshot_1

Does this happen already? I do not know, I only challenge people directly or via ladders, but since there are many different rankings that are being tracked, they could be used during matchmaking?

To answer your original question, as another person that started on DGS and I still mostly stick to correspondence games, I’d estimate that my strength difference between live and correspondence is 4 to almost 5 stones. (2-3k correspondence on a good “day”, around 7-8k in live games)

1 Like

I think it’s a thing many people do however.

Another option for you is to just use a setting to hide ranks for example and not worry as much if your opponent happens to be a little weaker in correspondence or a little stronger in live, and just play as best as you can :slight_smile:

3 Likes

I guess the overall rank has to do with the number of games played in each category. My total of played correspondence games is twice the amount of played live games.

I always try to play the best I can and I sure don’t mind losing. But I cannot enjoy a game in which I’m supposed to be 3 stones stronger. And trying to explain to an opponent during a live game why I accepted a game ‘beyond’ my rank, which I sort of feel obliged to do, is awkward and distracting. I prefer to play in Zen-mode because I need all my attention focussed on the game.

2 Likes

Yeah you can do zen mode and then hide ranks and then there’s no need to interact with the opponent at all or worry if you’re supposed to be three stones stronger or not :slight_smile:

1 Like

I agree, and that’s why I think the “Overall” rank should be completely eliminated (see lichess for an example of keeping ranks separate), but in the meantime I have some half-dozen accounts each dedicated to a specific type of game. My main “Samraku” account is for Slow Rapid 19x19, Classical 19x19, Correspondence 19x19, and all unrated games except blindfold against randoms, which has its own account (if I arrange with someone to play a blindfold game, it will likely be on my blindfold account if rated, but on my main account if unrated)

3 Likes

That doesn’t change that correspondence games are then easier than live games, and blitz games get even harder for me.

I think that the match making system and auto handicap system should consider the specific time settings foremost and only look at the overall rating if the specific rating has too much uncertainty.

Using different accounts is inconvenient and doesn’t help the rating system get the necessary data.

1 Like

It’s hard to know it’ll really make much difference though? The data was supposed to be showing that overall ranks of players were good predictors of who would win, and separating out all of the categories actually likely means games will be very unbalanced for much longer.

If OP for example primarily plays correspondence and has improved over time, then the live rank might just be outdated from not having been played as much, but also outdated from missing the rank improvement.

It’ll be the same for blitz and rapid (whatever rapid is counted as) etc.

It feels like this would better if uncertainty was increasing over time, which I know was talked about implementing, but I’m not if it has been implemented. Otherwise it really won’t make much difference for outdated ranks in certain time settings.

A person that improved from 10kyu to 1kyu playing correspondence might still get 10kyus in live format if they haven’t played live in a long time but their rank was still fairly certain at that time, or played live on pandanet, kgs, fox etc as the OP mentioned.

1 Like

No, that’s not the case as far as I’m aware of. After playing more live games on OGS my overall rank dropped a bit and that unintentially led to accepting a correspondence-game from an equal opponent who had to give me two stones handicap (I wasn’t paying attention to the settings). Very uncomfortable.

If it ‘just takes time’ I will have to accept that, even if that means that it might take a year or two and in the meantime I will have to chose my opponents carefully.

The main issue here is that I would like it to be relatively easy to find a match in whatever time-setting or board size with a +/- 2k margin.

3 Likes

I get that. But after playing over 500 live games on other platforms in the last 6 months my rank hasn’t improved much yet.

1 Like