Life In 19x19

Posted: **Sun Jan 12, 2014 6:51 pm**

@Mef: And I won't argue with you that the KGS ranking system is a feasible one. I simply argue that other systems, such as Tygems, are more volitile. And for that reason, I have no fear of getting stuck at a rank and having to play 10s or 100s of games to change it.

On KGS, you might note, I have played maybe less than 10 games in the last few months, because I want to play only when I am confident that I can have a good game. Otherwise, I'll have to win a ton more games to be able to play stronger players. I've played more than 10 in a day on Tygem. There's your difference.

Posted: **Sun Jan 12, 2014 7:05 pm**

Kirby wrote:On KGS, I *am* afraid to play ranked games. That's because, if I screw up, it will be harder in the future to achieve the same rank. If I have a bad day and lose 10 games, then even if I have a good day tomorrow, it takes longer to get my rank back up.

Except that this is not true. As long as you are playing games at a consistent rate, you will see no difference in how your rank moves. Much like when you do well you feel our rank doesn't go up as much as you feel it should...When you have the slump on KGS, your rank will not go go down as much as you feel it should. Any disparity you think exists is from selective perception.

The rank still converges on Tygem, too, for these players, but allows more volitility for those that are interested in challenging themselves the most.

This is also not true. Tygem's system promotes Brownian motion more than it promotes convergence. If you will allow me a thought experiment I can demonstrate why:

Imagine you have a group of 11 friends who are all 10k, and they decide they want to improve so they all sign up as 1d on Tygem. In order to track their progress against themselves, every week they each play everyone else in the group once, then play 5 games against other people of the same rank. If they always win against each other 50% of the time, and lose against everyone else they play, then they will all maintain 33% win rates and they will never demote. Note: this is with no malicious intent by any player as well as every player always trying their best to win every game.

What's more -- this type of example could be used to keep any arbitrarily strong or arbitrarily weak group at any arbitrary rank. You could quite literally maintain a set of random playing bots at 5d. This is all without foul play on the part of the players. This example also allows you to assume that everyone else on the server is properly ranked. Once you introduce inaccuracy and random fluctuations even more crazy things could happen.

This is because tygem assumes that any given player has a 50% chance of winning a rated game at any given time. All wins and losses are treated equally.

Any ranking system that attempts to have predictive power will take that group of 11 and have them spiral down relative to the aggregate until they settle at where they should be.

Posted: **Sun Jan 12, 2014 7:20 pm**

Surely Mef, (if I understand it correctly) the crucial flaw in KGS's system is that it isn't the rank a person had when you beat them but the rank you had now that is counted in the calculation? There's an assumption in the system that someone I beat two months ago I could beat today even if they've improved. Or if they got worse the opposite effect would occur. This strikes me as a rather strange way of calculating someone's rank.

If I misunderstand it please correct me.

Posted: **Sun Jan 12, 2014 7:36 pm**

Boidhre wrote:Surely Mef, (if I understand it correctly) the crucial flaw in KGS's system is that it isn't the rank a person had when you beat them but the rank you had now that is counted in the calculation? There's an assumption in the system that someone I beat two months ago I could beat today even if they've improved. Or if they got worse the opposite effect would occur. This strikes me as a rather strange way of calculating someone's rank.

If I misunderstand it please correct me.

You are correct with one caveat. This is all true, in the absence of new information being available to the system. If both you and your old opponent have continued playing games (against each other, or against other people) in those two intervening months, that single game will only carry about 1/4 of the weight of a new game being played. This is actually a major strength in KGS's algorithm that allows it to provide reasonable ratings with limited data. On other servers you may need to play 20 games before you can have a solid rank. On KGS it is usually accomplished with 2-4 games. There was a study done using KGS data that compared time-dependent algorithms to incremental algorithms (Elo, Glicko, etc) and the former were found to have superior predictive capability.

Posted: **Sun Jan 12, 2014 7:53 pm**

Kirby wrote:
Mef wrote:The "more weight" thing is not exactly accurate.

Let me clarify what I mean by "more weight". On Tygem, I have no fear of playing ranked games. If I get into a slump, I know that if I get my act together, within 20 games, I can move my rank back up.

On KGS, I *am* afraid to play ranked games. That's because, if I screw up, it will be harder in the future to achieve the same rank. If I have a bad day and lose 10 games, then even if I have a good day tomorrow, it takes longer to get my rank back up.

So while some people have different rating system preferences, my opinion is simple: I'm afraid to play a lot on KGS. I'm not afraid to play a lot on Tygem. Therefore, I prefer Tygem's system.

---

This begs the question, why am I afraid to play? Isn't it my "ego" that's making me want to keep my rank up?

To that, I can direct you to the advice that many have given about playing stronger players. People don't often say to play games as a way to improve, but when they do, it's almost universally that you should play strong players. With KGS, if I screw up because I'm having a bad day, I will not have the opportunity to play strong players for a much longer time, given that I play a lot of games.

Now if someone is not interested in improving, or if they want to be the same rank for a long time, then KGS is the system for them - play a lot of games, and the rank will converge.

The rank still converges on Tygem, too, for these players, but allows more volitility for those that are interested in challenging themselves the most.

KGS has succeeded in getting you to treat every single rated game as seriously as possible. :lol:

Posted: **Sun Jan 12, 2014 8:00 pm**

Mef wrote:You are correct with one caveat. This is all true, in the absence of new information being available to the system. If both you and your old opponent have continued playing games (against each other, or against other people) in those two intervening months, that single game will only carry about 1/4 of the weight of a new game being played. This is actually a major strength in KGS's algorithm that allows it to provide reasonable ratings with limited data. On other servers you may need to play 20 games before you can have a solid rank. On KGS it is usually accomplished with 2-4 games. There was a study done using KGS data that compared time-dependent algorithms to incremental algorithms (Elo, Glicko, etc) and the former were found to have superior predictive capability.

Have you got a link to that study? I'd be extremely curious about the caveats attached to that statement (I imagine there are playing thresholds below which the KGS algorithm stops producing sensible predictions and these can cluster if too many people in someone's history as well as themselves start playing irregularly and their ranks become unstable?).

Posted: **Sun Jan 12, 2014 8:08 pm**

Kirby wrote:I simply argue that other systems, such as Tygems, are more volitile.

Well this is certainly true, but that's largely because KGS aims for accuracy over volatility. Admittedly, for many people accuracy in a rank isn't necessarily a priority. It's also true that for some people having ranks introduce an element of randomness may make things more enjoyable. This is one of the reasons why games like poker and backgammon can be so appealing -- even though there is strategy and potentially large skill disparities, the randomness keeps the outcome unknown.

In fact for this same reason I wouldn't be surprised if the poor predictive power of Tygem's system were intentional. Tygem does have a betting feature, and there is a well known history of gambling being associated with baduk in Korea. From a gambler's perspective the worst thing in the world you can have is a public, perfectly accurate rating system, as this allows any bloke off the street to know the outcome in advance.

To summarize -- Do some people find Tygem's ranking system fun? Sure. Does Tygem's ranking system work for people who just want to play? Sure. Does Tygem's system provide significant predictive power in guessing game outcomes? No.

In that vein, I consider it to be a poor metric for measuring improvement for people who use ranks to do that sort of thing. If you just want to play games and have fun? Go to town, I'll be the last one to stop you!

Posted: **Sun Jan 12, 2014 8:16 pm**

I remember I got stuck at about ~6k for a while. I think I just got caught in a 6k mindset. I was playing maybe 6-7 games a day. I started playing over the board games and maybe played on KGS 1 or 2 times per week. The over the board games were against stronger people and it wasn't so much a gradual improvement, but more of a realization of what it takes at a higher level by experience it myself instead of being lazy and playing the same responses all the time. I think over two months I went up to 5k, finally played some F games even with a 2k and won them all so recreated and got re-established at 2k.

Its the culture of KGS combined with the ranking system. Nobody ever plays outside of their rank really, so weak people never get nice games with strong people. Its very selfish honestly.

Posted: **Sun Jan 12, 2014 8:17 pm**

Boidhre wrote:
Mef wrote:You are correct with one caveat. This is all true, in the absence of new information being available to the system. If both you and your old opponent have continued playing games (against each other, or against other people) in those two intervening months, that single game will only carry about 1/4 of the weight of a new game being played. This is actually a major strength in KGS's algorithm that allows it to provide reasonable ratings with limited data. On other servers you may need to play 20 games before you can have a solid rank. On KGS it is usually accomplished with 2-4 games. There was a study done using KGS data that compared time-dependent algorithms to incremental algorithms (Elo, Glicko, etc) and the former were found to have superior predictive capability.

Have you got a link to that study? I'd be extremely curious about the caveats attached to that statement (I imagine there are playing thresholds below which the KGS algorithm stops producing sensible predictions and these can cluster if too many people in someone's history as well as themselves start playing irregularly and their ranks become unstable?).

Whole-History Rating: A Bayesian Rating System for Players of Time-Varying Strength

Description of comparison:

Table of results:

Posted: **Sun Jan 12, 2014 8:22 pm**

Mef wrote:
Kirby wrote:On KGS, I *am* afraid to play ranked games. That's because, if I screw up, it will be harder in the future to achieve the same rank. If I have a bad day and lose 10 games, then even if I have a good day tomorrow, it takes longer to get my rank back up.

Except that this is not true. As long as you are playing games at a consistent rate, you will see no difference in how your rank moves.

Wrong. It *does* take a longer time to get rank back up than with Tygem.

If you like thought experiments, try this one:

Let's say that two casual, non-studying players, Bob and Sally, are the same strength in real life.
Bob has converged to 3d on Tygem after playing 1000 games a day, every day.
Sally has converged to 1d on KGS after playing 1000 games a day, every day.

Now let's say that Bob and Sally both go to Go Congress for a week, and get fired up about studying go. They do lots of go problems, review pro games, and work hard to increase their strength. Let's say, as a result, that they both are one stone stronger in strength after this study. After coming back from Go Congress, let's say that Bob is 4d in Tygem strength and Sally is 2d in KGS strength.

Who needs to play more games to make their rank match their strength?

Mef wrote:
This is also not true. Tygem's system promotes Brownian motion more than it promotes convergence. ...

Logical fallacy: Strawman argument.
Your example does not address the point I was making. Let's go back to what I said:

Kirby wrote:The rank still converges on Tygem, too, for these players, but allows more volitility for those that are interested in challenging themselves the most.

My statement is about volatility of Tygem's ranks. Your statement is about the ability that Tygem gives for users to set their own initial rank. Personally, I would also prefer if Tygem did not provide this functionality, and determined rank on its own.

But this has nothing to do with how volitile Tygem's ranking system is.

Mef wrote:What's more -- this type of example could be used to keep any arbitrarily strong or arbitrarily weak group at any arbitrary rank. You could quite literally maintain a set of random playing bots at 5d.

I don't believe this is true. As I understand, you can only set initial rank up to 3d. Do you want to talk about rank setting? KGS provides the ability for users to set their initial rank as high as 1k when using automatch. Granted, this is not as direct as Tygem's methodology, but the rank setting Tygem allows is not what I'm most a fan of. I'm a fan of volitility, which allows you to move up in rank, even if you've played a lot of games.

Compared to Tygem, based on the "thought experiment" above, people are punished for playing more games on KGS more than they are on Tygem.

Posted: **Sun Jan 12, 2014 8:26 pm**

Mef wrote:Does Tygem's system provide significant predictive power in guessing game outcomes? No.

Your statement is accurate for some cases, but not for others. For example, in the thought experiment I provided in the last post, Tygem is a more accurate indicator of the game outcome than KGS, because it will identify the growth in strength faster than KGS does.

Posted: **Sun Jan 12, 2014 8:27 pm**

yoyoma wrote:KGS has succeeded in getting you to treat every single rated game as seriously as possible.

Agreed on this point. Although, before using Tygem much, it also succeeded in having me not play many rated games.

Posted: **Sun Jan 12, 2014 8:39 pm**

Kirby wrote:My statement is about volatility of Tygem's ranks. Your statement is about the ability that Tygem gives for users to set their own initial rank. Personally, I would also prefer if Tygem did not provide this functionality, and determined rank on its own.
...

I don't believe this is true. As I understand, you can only set initial rank up to 3d. Do you want to talk about rank setting? KGS provides the ability for users to set their initial rank as high as 1k when using automatch. Granted, this is not as direct as Tygem's methodology, but the rank setting Tygem allows is not what I'm most a fan of. I'm a fan of volitility, which allows you to move up in rank, even if you've played a lot of games.

Compared to Tygem, based on the "thought experiment" above, people are punished for playing more games on KGS more than they are on Tygem.

My statement was not in reference to the fact that you can set your own rank. My statement was in reference to the fact that Tygem ranks do not necessarily converge even assuming that all players maintain consistent skill levels and play all games honestly. Being able to set your own rank should not be an issue for a rating system, as it should move you to the correct rank over time. Tygem does necessarily not do this. If you maintain a win rate of 33% on Tygem, you do not demote. It is quite easy to have a minority group that indefinitely self sustains a rank in spite of having both a significant number of games played with the greater group and no misconduct during any games. This effect would never be observed on KGS/IGS/DGS/OGS/etc.

On KGS you cannot set your rank to anything, your rank is always determined by the maximum likelihood of your future game results based upon your past performance.

Posted: **Sun Jan 12, 2014 8:45 pm**

Mef wrote: Tygem does necessarily not do this. If you maintain a win rate of 33% on Tygem, you do not demote.

If you are able to win against a 5d for 33% of your games after 10,000 games, your strength must be close to 5d. It would be highly unlikely for a kyu player to be able to achieve this.

In any case, this is unrelated to the point I am getting at - Tygem predicts changes in strength more easily than KGS does when users have played many games. A direct example of this can be seen in the "thought experiment" I provided.

Posted: **Sun Jan 12, 2014 8:46 pm**

Kirby wrote:Let's say that two casual, non-studying players, Bob and Sally, are the same strength in real life.
Bob has converged to 3d on Tygem after playing 1000 games a day, every day.
Sally has converged to 1d on KGS after playing 1000 games a day, every day.

As an aside -- If we actually performed this thought experiment, Bob's rank would never be stable. You would expect his rank to jump +-2 stones everyday simply from random noise.

Life In 19x19

Does KGS rank system encourage users to make new accounts?

Re: Does KGS rank system encourage users to make new account

Re: Does KGS rank system encourage users to make new account

Re: Does KGS rank system encourage users to make new account

Re: Does KGS rank system encourage users to make new account

Re: Does KGS rank system encourage users to make new account

Re: Does KGS rank system encourage users to make new account

Re: Does KGS rank system encourage users to make new account

Re: Does KGS rank system encourage users to make new account

Re: Does KGS rank system encourage users to make new account

Re: Does KGS rank system encourage users to make new account

Re: Does KGS rank system encourage users to make new account

Re: Does KGS rank system encourage users to make new account

Re: Does KGS rank system encourage users to make new account

Re: Does KGS rank system encourage users to make new account

Re: Does KGS rank system encourage users to make new account