IMO we should use that information only for match-making. I have no problem with a player being matched at a higher rank when he says that he improved much. But the rating should only follow if he shows results consistent with that. So instead of sandbagging people you play at the higher rank, but your rating isn't reset, but improves rapidly if you beat people at that level.

_________________
Sanity is for the weak.

But at the same time, your opponents are judged as if they played a much weaker player. If your rating is around that of a 10 kyu, and you enter as 5 kyu, then anyone who loses against you is effectively treated by the rating system as if they had lost against a 10 kyu.

The cheat your way up argument is nonsense. If we have 20 1k players, and 20 4d players, completely hypothetically, and the 20 1k players all get 4 stones stronger, then their ratings once it settles down are 2d, as are the 20 4d players. The 4d players aren't any weaker, they have just had their rating points sucked away. With GoR being based on ELO, resets are the only way of preventing this from happening. Cheating is a non-argument, it's maintaining the ranks of the dan players. This isn't exclusively a UK problem, it's a European problem. 2100 was 1d back in '95 or whenever EGF started the system, it's now under 2040 for precisely this reason. This means that people of the same ability have lost about 60 points, and the system is being "hacked" to make them still appear with the same rank. Not only this, but if you "reset" as 2100, you're resetting as GoR 2100 instead of your country's (or Europe's) 1d equivalent rating.

The whole thing is rather broken, to the point where over-prescriptiveness is fixing a symptom, badly, instead of addressing the cause. I would frankly prefer the abolishment of formalised rating systems completely, now that enough players have online accounts and a fair idea of "where they fit" in the grand scheme of things. The problem of "localised groups" have their rank isolated from the rest is already an issue (GoR gives the impression it isn't, but GoR 2000 doesn't always equal GoR 2000!), and having a formalised rating system doesn't fix it, just gives some false impression that you have.



If someone says "Wow, you're good for 1k/2k/whatever" that's a compliment. If someone says "You really shouldn't enter with a grade just to win prizes, it's cheating", it's not a compliment, and I don't care if it is a reflection on my potentially under-rated ability or not, it just hurts to have my motive challenged like this, and makes me want to avoid similar situations in the future. I don't need or want people to think I'm strong or weak, I just want to play Go. Ratings and ranks in real life tournament have not helped my experience in any way so far.

Since the uncertainty of your rating is large this won't cost them much. And since you play several of them it's very likely that the estimated ranking for the time when the tournament happened gets updated. So the loss of points will be distributed between them, and if their ranking is more stable than yours they won't be affected much.

_________________
Sanity is for the weak.

Well, I don't want to hijack this thread, but at least this is not true. Every game adds additional points to the system, because there is a variable for that in the equation. Whether this is too small or too large relative to the improvement in the subpopulation / the number of tournament games played remains open to debate. But it does exist (http://senseis.xmp.net/?EuropeanGoDatab ... list#toc14 and http://www.europeangodatabase.eu/EGD/EG ... php#System), although I agree that resets are a useful feature.

That is, while in your example the may all end as 2d in the short term, if they continue to play each other for the next twenty years in many games without any improvement at all, they will all be 4d, 5d or even 6d.



But you're much more likely to win a (main) prize when you start with a higher rank (higher MMS, above the bar whatever) than by sandbagging. I don't get it. Is this about the most wins encouragement prizes, really? People will always have an opinion whether you are strong or weak, you can't help it, even in online play.

I'm aware of the hack. It's quite clearly not enough to compensate for player improvement just looking at the statistics so far. If it was, GoR 2100 would, in general, be the same as 1d. There is the argument that people are unreasonably entering as 1d when they aren't strong enough, thus pulling the mean down, but most of my calculations are based on long established dan players who have been neither improving nor getting worse (like 2110 rated 2 dans who were once 2200 rated 2 dans, with little change in result quality against similar opponents that would imply a decline in ability). So my statement is still well supported, even if the EGF have implemented something that makes it very slightly mitigated.

What evidence is there that the GoR system has made the local, national, and international tournament scene a better place? We still have Japanese 5d players who struggle to beat Dutch 2d players _yet carry on entering as 5d_. We have French SDK players who aren't allowed to self-promote destroying other European players (we had a French "1k" come to a tournament last year who dominated a British 3d and 4d). We have Polish players ejected from their federation for entering as a rank they hadn't "earned". What positive things have come out that offset all of this?



Not in this country. If I enter as 4d, I'll be definitely above the bar and have a chance of winning a bottle of wine. Alternatively, I could enter as 7k, definitely win 3 games, and definitely get a bottle of wine. The only difference is I won't get first pick.

But, frankly, I don't care about prizes either. What I _do_ care about is unpleasant atmospheres caused by accusations that either amount to "cheating for a better rank", or "cheating to win your games (and therefore prizes)".

As an addendum, if we carry on with formalised ratings, I'm wholly in favour of adopting a rating system that can incorporate volatility or cater for unexpected results. There are plenty of systems that do this: Glicko is a simple one not too different in complexity to ELO, WHR looks very promising, particularly as the computation aspects for real life tournaments are a non-issue.

What I really hope for is that people attach less value to ranks and ratings. I want tournaments to be based on social interactions whilst playing a mighty fine hobby against people of similar strengths, with a prize entirely optional (other of the honour of "being the winner").

Ouch. Imho, countries / national federations that discourage or even disallow rank resets simply confuse the system, while they feel good about their "earned" ranks. Polish tournament participation is decreasing for years now - is this related to the rank fixing policies? There must be an immense amount of anger about such a policy. Any comments?

The longer I read such threads the more I believe that rule expertise and rating science are harmful to the promotion of the game.

I suppose my situation was an unnecessary distraction from the main thread, and for that I apologise. I think my point is really your last one Tapir, the rating system as it stands has led to a number of formalised policies that seem to be damaging to Go as a whole, such as the situation in the UK and Poland (and elsewhere), and I think it is making tournaments a much less enjoyable place to play your Go. It's certainly causing unpleasantries over here (this thread started by such an example), even though the person in question is unlikely to be phased by the issue.

In short, I agree with you 100% that this rule/rating pedantry (ok, science is less pejorative ) is harming the promotion of the game.

Addendum: Apologies for ranting on a bit. I suppose I've become a bit of an evangelist on this one these days as I see it doing what I perceive to be so much harm.

I don't think the term science applies, really.

Yes, there is science behind the design of rating systems, but that does not mean the current policies in individual countries are science based.

In the Netherlands, there was a proposal to regulate kyu ranks through the EGF ratings (dan ranks are already regulated differently). We then had someone with a PhD in mathematics actually research whether it was a good idea, and his conclusion was that self-chosen kyu ranks were far more reliable than EGF ratings.

If you provide a link to that research, I will like your previous post

It is in the archives for the AGM of the Dutch Go Association, but it is in Dutch:

http://www.gobond.nl/Documenten/Bestuur ... rating.pdf

The bulk is tables showing:

Given a (0,1,2,3,4,5) kyu difference in rank, what are the winning percentages for certain rating differences?

and the reverse:

Given a (50,100,150,...,500) point rating difference (rounded), what are the winning percentages for certain rank differences?

The numbers show very clearly that rank is very much more predictive for results than rating

Dutch is sufficiently similar to German to extract the main idea. Thank you.

I could write a response about the original posting, which referred to a real case recently. But I'm not sure whether the readers of this thread would actually be interested in the facts, as the thread seems to have moved on.

Please let me know if you are.

Jon Diamond
President
British Go Association

I'd was interested to know if anyone would do this in the absence of a "national policy on ranks/ratings". I wasn't especially interested in the facts of why it was done.

It occurs to be that it is in fact possible to treat a new player's rank as provisional. Tournament results can be resubmitted to the EGD. If such an occurrence was occasionally demanded, and was adequately explained in the submission comments, I don't see any objection coming from the EGD staff. This would strike me as an improvement on current practice. If a player's initial entry rank was found to be truly misleading after, say, 10 to 15 games, it could be harmlessly corrected. There are several players out there who could have benefited from such an approach.

I would be very interested to know the precise facts behind this one Jon.

I would like to know, specifically, why entering as a 2300 GoR player when his performance exceeded 2300 was considered worthy of resetting to 2200.

I wouldn't normally comment in such a specific vein as below, but the original posting and subsequent discussion has been sufficiently detailed to have identified a specific case which was discussed by the BGA Council recently. The facts are:

- the player referred to was known to the EGD rating system, but at a lowish kyu rating due to there being no known games for many years
- in the EGD rating system if the entry grade is two or more grades above the previously entered grade then it does a reset and takes the entered grade as the base point for calculating future ratings, so what the entry grade is makes a significant difference to the exit rating (and other people's ratings)
- the player entered the Maidenhead tournament at an entry grade of 3 dan which the Tournament Director accepted
- the exit rating for this player would have exceeded 2300 only if the entry grade was 3 dan or above
- the BGA Rating Officer noticed this and queried the basis for the entry grade, referring the issue to the BGA Council
- the BGA has a policy statement, at http://www.britgo.org/policy/policies7 relating to rating resets.
This was approved some considerable while ago, to encourage kyu players who are rapidly improving to do so, but to restrict resets at dan levels because of the possible implications for the rating system as a whole
- the maximum dan grade for entry to a tournament for effectively an unknown person in this policy was set to 2 dan (and subject to pre-approval), except for people with grades assigned by foreign Go associations
- the BGA Council considered this issue and has a) submitted this tournament to the EGD using an entry grade of 2 dan and b) written appropriately to the person concerned, who will be entitled to enter his next tournament at 3 dan because of these results

----------------------------------------------------------------------------------

There is a continuing debate about the EGD rating system and whether it is deflationary or not - further statistical analysis is definitely required.

However, it should be noted that UK grades, as per our published list, are pretty well correlated to the European average (UK 2031.7 v 2039.5 for an average shodan). The EGD rating system uses at its basis a theoretical rating for an average shodan of 2100, which is only achieved by Slovakia and the Czech Republic (who originated the system).

It should also be noted that strengths published in the UK rating list are based on the person's rating and the European average correlations, as described in the FAQ, not on the theoretical basis for the EGD system.

Regards
J Diamond
President, British Go Association

Given that policy the decision was very reasonable.

I still think the math behind GoR should be improved so that the grade used to enter the tournament doesn't affect the rating directly. But of course such a decision must work with the current rating system and policies, and not some hypothetical system I'd like to have.

_________________
Sanity is for the weak.

Many thanks Jon, very helpful. Has there been any similar attempts to evaluate with UK 2031 is typically equivalent in playing strength to 2031 from other European countries?

Also, if we have visiting strong players, "the maximum dan grade for entry to a tournament for effectively an unknown person in this policy was set to 2 dan" is kind of worrying - With no-one potentially capable of assessing their strength, could be in principle be forcing Korean 6 dans to enter as 2 dan, even if this is retroactively updated? From what you've said above, there is a precedent set now that effectively prevents other behaviour - the argument of "if they perform at or above their entered grade, we will take no action" sounds particularly appealing to me, but as the Maidenhead 3 dan actually overperformed his 2300 rating and was reset anyway, I presume this is not the case?

PS I was at the council meeting when that policy was formulated. I remember the discussions, and now regret that it has been set in stone in this way. The discussions were to give a way of preventing excessive use of resets to pick up rating points, but it has lead to "by the book" decisions that appear to be lacking in common sense.

@topazg
I think the exception for people with graded by foreign go associations applies to your Korean 6d.

_________________
Sanity is for the weak.