It is currently Wed May 22, 2013 9:54 pm

All times are UTC - 8 hours [ DST ]




Post new topic Reply to topic  [ 131 posts ]  Go to page Previous  1, 2, 3, 4, 5 ... 7  Next
Author Message
Offline
 Post subject: Re: KGS ranking revisited
Post #21 Posted: Wed May 09, 2012 4:08 pm 
Lives in sente

Posts: 760
Liked others: 109
Was liked: 116
Rank: KGS 3 kyu
Universal go server handle: speedchase
Then wouldn't you have to wait until it actually trickled down?

Top
 Profile  
 
Offline
 Post subject: Re: KGS ranking revisited
Post #22 Posted: Wed May 09, 2012 4:40 pm 
Lives in gote
User avatar

Posts: 450
Location: Portland, OR USA
Liked others: 257
Was liked: 282
KGS: wms
emeraldemon wrote:
Is there a reason not to anchor the strongest players at 9d and let it trickle down?

Yes, a very strong one. That means that as the strogest player rank changes, everybody shifts by the same amount. When the strongest player stops playing, if the 2nd strogest player is 2 stones weaker, then suddenly everybody jumps two stones.

Our current system sometimes has some changes when anchors are adjusted, but overall it is far more stable than anything based on a single player would be.

Top
 Profile  
 
Offline
 Post subject: Re: KGS ranking revisited
Post #23 Posted: Wed May 09, 2012 5:51 pm 
Lives in gote

Posts: 397
Liked others: 153
Was liked: 128
GD Posts: 846
I'm not really fond of the stochastic walk ratings seem to take on KGS even when you don't play and I certainly don't like the expiration to '?' when you don't play for a while, but overall it may be the most accurate system in current use for players who play there regularly.

Other servers have had resets. IGS did some years ago (and I mean before the 2005 change, somewhere in the summer of 2002 when there was a 3-stone shift. I wasn't playing then so I don't know for sure. I guess right before than an IGS 1d was pretty darn kickass relatively speaking.)

If I wanted to get attached to a number, I'd buy a certificate.

_________________
The person who says "don't......" is just showing he doesn't know Go at all, I think. :-) - Tei Meiko 9p

Top
 Profile  
 
Offline
 Post subject: Re: KGS ranking revisited
Post #24 Posted: Wed May 09, 2012 6:17 pm 
Oza

Posts: 2975
Location: North Carolina
Liked others: 208
Was liked: 434
Rank: AGA -4.99385
GD Posts: 65
KGS: MssrBlanc 2k
DGS: Hyperpape 5k
OGS: Hyperpape 7k
Pippen: my point was that if that's what you want, you can't complain about KGS, since it's just one rating system among many.

Aligning all the ratings systems would require collaboration from around the world, and would also happen to be impossible at the present time.

_________________
When you think your opponent has made a mistake, that's when you should be the most cautious.

Top
 Profile  
 
Offline
 Post subject: Re: KGS ranking revisited
Post #25 Posted: Wed May 09, 2012 8:25 pm 
Oza
User avatar

Posts: 2303
Liked others: 256
Was liked: 533
Rank: kgs 6k
snorri wrote:
I'm not really fond of the stochastic walk ratings seem to take on KGS even when you don't play ...


It's not really stochastic, is it? It assumes that you would be able to beat the same people you were able to beat when you were last playing.

Top
 Profile  
 
Offline
 Post subject: Re: KGS ranking revisited
Post #26 Posted: Wed May 09, 2012 10:41 pm 
Dies with sente

Posts: 113
Location: UK
Liked others: 79
Was liked: 18
KGS: 6kyu
.

Top
 Profile  
 
Offline
 Post subject: Re: KGS ranking revisited
Post #27 Posted: Thu May 10, 2012 1:49 am 
Gosei
User avatar

Posts: 1440
Liked others: 595
Was liked: 577
Rank: KGS 5k
KGS: goatsunday
Kaya handle: daal
jts wrote:
snorri wrote:
I'm not really fond of the stochastic walk ratings seem to take on KGS even when you don't play ...


It's not really stochastic, is it? It assumes that you would be able to beat the same people you were able to beat when you were last playing.


Yes, it assumes that while you were gone that you've been playing on Tygem. :lol:

_________________
"When you are learning, it is not so important whether you are right or wrong." - Bill Spight

Top
 Profile  
 
Offline
 Post subject: Re: KGS ranking revisited
Post #28 Posted: Thu May 10, 2012 12:45 pm 
Oza
User avatar

Posts: 2303
Liked others: 256
Was liked: 533
Rank: kgs 6k
daal wrote:
jts wrote:
snorri wrote:
I'm not really fond of the stochastic walk ratings seem to take on KGS even when you don't play ...


It's not really stochastic, is it? It assumes that you would be able to beat the same people you were able to beat when you were last playing.


Yes, it assumes that while you were gone that you've been playing on Tygem. :lol:

Well, not necessarily. If your most recent partners decline, you'll decline to. It just assumes that, in the absence of evidence, you can still beat the same people and lose to the same people.

Top
 Profile  
 
Offline
 Post subject: Re: KGS ranking revisited
Post #29 Posted: Thu May 10, 2012 1:34 pm 
Beginner

Posts: 6
Liked others: 1
Was liked: 1
Rank: KGS 7K
GD Posts: 12
KGS: witwit
The amount of importance some people are placing on ranks in this thread is quite silly. The most important thing for a ranking system is to be internally consistent, ie trying to have differences in rank equate to 50/50 games with the appropriate number of handicap stones. Why does it matter if there is a "shift" in rankings and some bits associated with your account on a remote server change from 5k to 7k? All that matters is your actual strength which can never be characterized by a single number but by comparison.

_________________
Games * Rank Graph
7k

Top
 Profile  
 
Offline
 Post subject: Re: KGS ranking revisited
Post #30 Posted: Thu May 10, 2012 1:46 pm 
Oza

Posts: 2630
Liked others: 0
Was liked: 280
witwit wrote:
The most important thing for a ranking system is to be internally consistent


A system with sudden shifts IS inconsistent.

Top
 Profile  
 
Offline
 Post subject: Re: KGS ranking revisited
Post #31 Posted: Thu May 10, 2012 2:20 pm 
Oza

Posts: 2975
Location: North Carolina
Liked others: 208
Was liked: 434
Rank: AGA -4.99385
GD Posts: 65
KGS: MssrBlanc 2k
DGS: Hyperpape 5k
OGS: Hyperpape 7k
Robert, like it or not, he's arguing about a reasonably well-defined concept. The fact that you don't like using the word "consistent" in naming that concept (which could naturally be named "momentary internal consistency") is pretty much irrelevant.

_________________
When you think your opponent has made a mistake, that's when you should be the most cautious.


This post by hyperpape was liked by 2 people: illluck, Phelan
Top
 Profile  
 
Offline
 Post subject: Re: KGS ranking revisited
Post #32 Posted: Thu May 10, 2012 10:34 pm 
Oza

Posts: 2630
Liked others: 0
Was liked: 280
One should not call an inconsistent system consistent.

Top
 Profile  
 
Offline
 Post subject: Re: KGS ranking revisited
Post #33 Posted: Thu May 10, 2012 10:55 pm 
Gosei
User avatar

Posts: 1440
Liked others: 595
Was liked: 577
Rank: KGS 5k
KGS: goatsunday
Kaya handle: daal
RobertJasiek wrote:
witwit wrote:
The most important thing for a ranking system is to be internally consistent


A system with sudden shifts IS inconsistent.


But he said internally consistent. If everyones rank is affected in the same way by the recalibration, the system will be just as good at predicting an even match as it was before.

_________________
"When you are learning, it is not so important whether you are right or wrong." - Bill Spight


This post by daal was liked by: Boidhre
Top
 Profile  
 
Offline
 Post subject: Re: KGS ranking revisited
Post #34 Posted: Fri May 11, 2012 12:04 am 
Oza

Posts: 2630
Liked others: 0
Was liked: 280
A rating system moving too many players upwards is internally inconsistent because a rating system shall be able to distinguish players instead of creating heaps of fake strong subpopulations.

Top
 Profile  
 
Offline
 Post subject: Re: KGS ranking revisited
Post #35 Posted: Fri May 11, 2012 3:13 am 
Lives in gote
User avatar

Posts: 416
Location: Germany
Liked others: 157
Was liked: 44
Rank: KGS 3 kyu
GD Posts: 325
RobertJasiek wrote:
... heaps of fake strong subpopulations.

That sounds scary ...

_________________
Stay out of my territory! (W. White, aka Heisenberg)

Top
 Profile  
 
Offline
 Post subject: Re: KGS ranking revisited
Post #36 Posted: Fri May 11, 2012 4:18 am 
Oza

Posts: 2975
Location: North Carolina
Liked others: 208
Was liked: 434
Rank: AGA -4.99385
GD Posts: 65
KGS: MssrBlanc 2k
DGS: Hyperpape 5k
OGS: Hyperpape 7k
RobertJasiek wrote:
One should not call an inconsistent system consistent.


You're begging the question! As I pointed out, a good descriptive name for the system he wants is "momentary internal consistency".

_________________
When you think your opponent has made a mistake, that's when you should be the most cautious.

Top
 Profile  
 
Offline
 Post subject: Re: KGS ranking revisited
Post #37 Posted: Fri May 11, 2012 9:51 am 
Oza
User avatar

Posts: 2303
Liked others: 256
Was liked: 533
Rank: kgs 6k
hyperpape wrote:
RobertJasiek wrote:
One should not call an inconsistent system consistent.


You're begging the question! As I pointed out, a good descriptive name for the system he wants is "momentary internal consistency".

You could be less obtuse, Robert, if instead you said "One should not call an erratic system consistent." That would make it clear that you are, indeed, making a synthetic claim and would point the way in two directions - first, you could clarify the extent to which the KGS system really is erratic, and second, you could clarify why, if we want a consistent system, we should make it less erratic as well.

The problem with inconsistency, as I see it, is that it leads to situations that violate assumed transitive properties. For example, if F always beats W, W always beats H, and H always beats F, it's difficult to apply any meaningful ranking to the {H, F, W} triad.

However, if a ranking system suddenly shifts, that merely means that it has adapted to new information. If it shifts a lot, that either means it gets lots of new information frequently, or it is very sensitive to what little it does get.

Top
 Profile  
 
Offline
 Post subject: Re: KGS ranking revisited
Post #38 Posted: Fri May 11, 2012 10:50 am 
Oza

Posts: 2630
Liked others: 0
Was liked: 280
jts wrote:
if instead you said "One should not call an erratic system consistent."


This is a more general requirement. Fine.

Quote:
if a ranking system suddenly shifts, that merely means that it has adapted to new information.


So the system outputs for each player: "I have adapted to new information! (I do not tell you which information, nor which adaption.)" ;)

IOW, not each adaption is good.

Top
 Profile  
 
Offline
 Post subject: Re: KGS ranking revisited
Post #39 Posted: Fri May 11, 2012 11:28 am 
Oza
User avatar

Posts: 2303
Liked others: 256
Was liked: 533
Rank: kgs 6k
RobertJasiek wrote:
jts wrote:
if instead you said "One should not call an erratic system consistent."


This is a more general requirement. Fine.

Quote:
if a ranking system suddenly shifts, that merely means that it has adapted to new information.


So the system outputs for each player: "I have adapted to new information! (I do not tell you which information, nor which adaption.)" ;)

IOW, not each adaption is good.

So your objection is not that it's erratic per se, but that wms will not share with you the information that goes into the daily iteration?

Top
 Profile  
 
Offline
 Post subject: Re: KGS ranking revisited
Post #40 Posted: Fri May 11, 2012 11:50 am 
Gosei
User avatar

Posts: 1295
Liked others: 305
Was liked: 192
Rank: KGS 2 kyu
KGS: greendemon
DGS: smaragdaemon
It seems to me that the ideal rating & handicapping system would strive to handicap every match to a 50% win rate. If a player's win rate is much higher or lower than 50%, that player is not being well served by the system. (we do occasionally get threads complaining about this, usually "I've won 10 games in a row and my rank hasn't gone up!")

If this is the metric we want to use, it's very easy to check the error: look at the average win-rate of every player over an appreciable number of games, and find the average distance from 50%. There was a competition a while back looking for improvements to ELO that used basically this metric on historical chess data, I believe.

Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 131 posts ]  Go to page Previous  1, 2, 3, 4, 5 ... 7  Next

All times are UTC - 8 hours [ DST ]


Who is online

Users browsing this forum: No registered users and 1 guest


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
cron
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group