Is your rating accurate ( too low? too high? )

[shapenaji]

So here's the interesting thing, suppose you have a population of players and suppose that they're all getting stronger gradually, at roughly the same rate. Then, playing against each other, their rating change will be a result of their growth rate, (if they are getting stronger faster than their opponents in the sample, then they'll go up a rank, otherwise the rank won't change, even if they are stronger than they were)

In this case, the rank doesn't measure strength, since if you had a population of entirely 4k's, the ones whose level did not change would actually probably drop a rank.

[gaius]

I used to know a dan player who, IMO, did not have the skills of a dan player. But he won a lot of games because he simply refused to lose.

During post mortems or other game reviews with other Dutch 1 dans, I tend to have the feeling that they are stronger than me. Whether that is caused by self-underestimation or genuine weakness I do not know, though frankly, I suspect the latter. Still, fact is that I win about 50% of my tournament games against the very same Dutch 1-dans. Interesting, isn't it?

[Sverre]

There is a self test in case you are not sure about your rating accuracy.

Ask yourself whether you stopped playing rated games or play only if you feel you can show your full strength or avoid players you do not know or rather wash the dishes instead of playing... if this is the case your rating is obviously too high.

Wow, is being overrated the only cause for this kind of behaviour? In that case I have been overrated on kgs for years, there must be something wrong with the rating system.

[entropi]

Ones rating is defined based on the game results, how can it possibly be inaccurate if he plays frequently enough?

How many games must a player win, before he can call himself underrated?
The answer my friend is blowing in the rating system.

[Stable]

Well, my win rate according to the kgs stats site is 60%ish, so logically I must have been consistently under-ranked for most of my go-playing time. (If you are improving you must naturally be under-ranked in any system based on past results.) Since I started out in kgs at 18k that might not be surprising. I put accurate though, as I just hit 1d (woohoo!) on KGS and there is no way I am 2d. I really couldn't judge to within half a stone!

[daniel_the_smith]

I put down "half a stone overranked" assuming that the AGA will probably make me 2d after my last tournament if the ratings ever get updated. Yeah, I won all my games but it was luck, luck, luck...

[Dusk Eagle]

Araban, I guess I really just don't play Go when I am too tired and feel like falling asleep. So that could be why my strength fluctuates less. I also don't know if I have ever played five online games of Go in one day (in real life is a differenty story...) so I guess I don't get fatigued to the same extent as you.

I recently noticed that my win percentage on KGS is 65%. Given that I have used this account ever since I started playing online Go, I would say it's not much of a stretch to say I have been underranked for a significant amount of my time playing Go.

[entropi]

<TwilightZone>
Guys, are you talking about different ranking systems or what? Apparently not. But then I don't get this discussion.

Your rank is nothing more than the numerical definition of your strength based on facts (i.e. your game results taking into account your opponents relative strengths). I don't understand how it can be wrong, it is just a definition.

It is like saying I am more rich than indicated by the money and property I own. How is it possible, how do you then define richness?
Likewise, if you think you are stronger than your rank, then how do you define strength?
</TwilightZone>

But on the other hand, note that win percentage and rank/rating are different things because win percentage does not take into account your opponents strengths and/or handicap stones.

And of course, if what you mean is "I play once in a month on KGS and I am improving fast" or "I haven't played in two years" then your rating is not uptodate. Meaning that there is not enough data to define your current strength. But that's a different story. I suppose people posting in this thread play regularly enough. If not, I then will say "hmm now I see what you mean".

[Stable]

It seems pretty simple to me entropi - your rank is based on your past strength (things you have already done). All you have to do to be stronger than your rank is improve after a game, or for ranking to lag behind increases in strength (KGS when you play many games is well known for this). Given that doing tasks and then sleeping has been shown to make you better at doing those tasks improvment without playing is quite possible too. Or maybe you will have a mini-epiphany about direction of play. Or learn some new joseki. Or do tsumego. Or whatever.

It's like saying "I am richer than my last bank statement shows" - you may well be if you earned some money, got interest or received a gift.

[entropi]

It seems pretty simple to me entropi - your rank is based on your past strength (things you have already done). All you have to do to be stronger than your rank is improve after a game, or for ranking to lag behind increases in strength (KGS when you play many games is well known for this). Given that doing tasks and then sleeping has been shown to make you better at doing those tasks improvment without playing is quite possible too. Or maybe you will have a mini-epiphany about direction of play. Or learn some new joseki. Or do tsumego. Or whatever.

It's like saying "I am richer than my last bank statement shows" - you may well be if you earned some money, got interest or received a gift.

Yes, of course I understand that. But then how can you measure this increase? How can you for example say "I got 1/2 stone stronger since my last game" if you have no means to measure it. I mean it is not defined, "strength" is an abstract concept.

For example your KGS rating still changes even if you don't play (which is a particularity of kgs rating system). This of course does not mean that your "strength" (in terms of knowledge) also changes. But since you cannot measure your strength without playing (i.e. you cannot give it a name, a number, etc) you simply have to accept that your rating is the best approximation to (or let's say the most precise definition available of) your current strength.

Otherwise, since of course you normally (if not always) get stronger after each game, your current strength will never be defined.

(Edit: Again, while arguing all these, I assume a reasonably frequent playing under the same rating system. Otherwise your rating is obviously not an approximation or definition of your current strength.)

[Stable]

Ah OK, I understand your objection now. Probably it can't be measured, but that doesn't stop people from estimating! Although again, if for some reason you were winning many games easily without your rank increasing that would suggest you were under-ranked I suppose.

[HermanHiddema]

There is a book on inuition and perception called 'The Invisible Gorilla', which is currently in the top 1000 books on Amazon. Among other interesting things, it refers to a poll of chess players in which they were asked if their ratings were accurate. 21% said they were; 4% said that their rating was too high, and 75% belived that their rating was too low.
Three out of every four players belived that their true strength was more than their rating.

The players were not unfamiliar with the game. They had an average experience of 20 years. So they knew what strenght was.
Nor were they talking about a delay in games being reported, which can give a rapidly improving player an artificially low rating. A year later the vast majority of players were still at or close to their same rating.

I think part of the reason for this is because you tend to remember your wins against stronger players, but forget your losses against weaker players. The first is a memory to relish, something to fondly recall. The second is something best forgotten, something you don't like being reminded of.

I actually keep a record of my win/loss ratio (in tournaments) against players around my strength to make sure that I'm not talking nonsense when I'm asserting that I'm stronger/weaker than my current rank.

[Exologist]

Definitely under-ranked. I tend to still win more than half the time playing as a 2k at the local club. I just seemed to make a ton of stupid mistakes too many times in a row playing online recently. -.-

[DrStraw]

Definitely under-ranked. I tend to still win more than half the time playing as a 2k at the local club. I just seemed to make a ton of stupid mistakes too many times in a row playing online recently. -.-

Just because you are winning over half your games does not mean you are under-ranked. It just means you are steadily improving (assumin the handicap is correct).

[entropi]

Just because you are winning over half your games does not mean you are under-ranked. It just means you are steadily improving (assumin the handicap is correct).

That's exactly what I was trying to say. "Underranked" and "currently improving" are different things. Otherwise almost everybody would be underranked, which would make the ranking system void.

It can only be that your rank follows your strength with a lag when you are improving fast. But calling this "I am underranked by that many stones" (which is not measuable anyway),... I don't know how much sense it makes.
To my understanding, one cannot possibly be underranked (or overranked) if he regularly plays serious games.

One way of becoming underranked is : stop playing under that rating system (e.g. on KGS) and switch to another one(e.g. IGS). After one year come back to kgs, then you will (hopefully) be underranked. Hmm another way can be playing 100 games in a row while drinking whisky-vodka mixed with tequila-cognac. Then you will probably be really underranked