Recently, I’ve made the jump to correspondence tournaments and have been enjoying them greatly so far. However, I’ve noticed a fairly large large discrepancy between my opponents’ correspondence and live ranks. These differences on the order of 4-5 ranks, and I find this a bit hard to believe.
While there certainly are strength differences given different time constraints, performance under the different modes must be highly positively correlated. Thus a 4-5 rank difference strikes me as very unlikely.
I’m curious if this is a high-kyu phenomena, or a symptom of the inherent latency in updated correspondence ranks compared to live ranks, or something of the sort. Admittedly its frustrating at times when you’re expecting an even game but get comprehensively defeated, only to see that your opponent has a live/overall ranking that is much higher than displayed.
the whole ‘overall’ vs ‘live’ vs ‘correspondence’ ranking differences primarily just leads to confusion. it depends on the number of games played with a certain time setting, and a bunch of other stuff.
part of it is that the various communities (9x9, 13x13, 19x19 vs blitz, live, correspondence) are somewhat independent and tend to drift.
part of it is that in a correspondence game, many people lean very heavily on the move planning facilities and so their reading greatly excels what they might be able to do on their own.
the truth is, that in this server, if you look for an even game you might be trounced, or might be at a demonstrably higher level than your opponent.
i recently played a corr game giving 4 stones against a 10k. turned out that his live rank was 1d. it wasn’t a slaughter, but every move i fell further behind. in the end he chided me for using handicap in settings.
meh. other servers have done a better job with rankings and setting expectations. this one doesn’t. as a result i sometimes get schooled and i’m ok with that. in a correspondence game it matters a lot less since i’m dedicating a few minutes every day over a week or two instead of an hour right now.
There is a split between the Live/Blitz players and the Correspondence players. I play almost exclusively Correspondence. I more comfortable with that as I rarely have a hour or two to dedicate to a game, or hours to a tournament. As such, the rating for the Live/Blitz is based on fewer games played less recently.
Having the time to think about a move and do the deep reading (once one knows enough to benefit from it) makes a large difference in the effective playing strength. I resent the “on their own”. I’m playing by the same rules as my opponent and all the effort is my own.
At the end of the day, a rating is only an estimate of current playing strength. Some people do not play consistently from game to game. Some people are improving and the rating is lagging behind their “true” strength.
This is one of the very few that allow both live and correspondence play. That is going to be a challenge . I’m sure the Devs would welcome any suggestions on how to improve the process.
ok, sorry, i’m old. maybe ‘in your head’. for me ‘go’ means constructing a move in your own mind, and playing exactly and only that move. not playing out variations on the board. that said, i really don’t mind if you choose to play that way, i’m just old school.
[quote=“yuri, post:4, topic:7918”]
i’m old. [/quote]
me too
Shall we check who is older? I’ll be 59 next month …
[quote] […] for me ‘go’ means constructing a move in your own mind, and playing exactly and only that move. not playing out variations on the board.[/quote]Yes, in most any live game I’d also see it that way. But correspondence? And when the variations we both try are still from our own minds? I see this as a good training for “thinking out of the box”, to discover moves we’d not dare to play in a live game (and hopefully remember them) …
My case is the opposite. I’m currently rated borderline 6k Correspondence and Overall, 11k Live and Blitz.
Although 5k disparity might be a bit exaggerated, I honestly believe I’m at least 3-4k worth of rating weaker in Live and especially Blitz as I’m older, am very slow in reading, and extremely weak against time pressure.
Elo works best when:
No new members come into the system, and no one goes out.
There is only one kind of game within the system.
And when these 2 conditions can’t be met, the solution needs to be very carefully designed and adjusted for that particular environment. A good solution for one environment (e.g. EGF) may well be not as good for another.
As yuri implies, the stem of the frustration is expectations management. The displayed ranking sets some expectations on how the game is going to play out, and the frustration arises when we realize we are wrong (either through play or checking their profile). It may help if we can see all of the ranks of the opponent on the game screen (without visiting the profile). It’d give a much better idea of what to expect.
Secondly, as Tokumoto points out, ELO isn’t necessarily designed for this system. I feel like there should be some base correlation between live and correspondence rankings, with correspondence ranks typically being higher than that of live ranks. Perhaps an alternative is to provide a margin of error on a pesons rank. This would be particularly useful with newer players who have less games played.
I’ll be 59 next month …