EGF Rating System Commission Report 2020

[betterlife]

@Javaness: Sorry, I didn't know about those EGF pair go rules and the significance of the country code in the EGD.
But perhaps it's something to address in a different topic in the EGF subforum, because it's outside the investigations and recommendations of the commission and I know very little about this.

These pair go rules were outdated the day they were made. IMO they are a good example on how to not make rules by defining details which can never work for longer time or would have to be adapted to the reality all the time... all the same in core they do what they should, give some guideline on how to handle the selection process for the international amateur pairgo championship in Tokyo!
The same counts for the super group at the open european championship and actually it's not only there where the EGD ratings are taken into account, it happens like this in a dozen of other occasions, like for example the grand slam qualification.
I agree the EGD itself doesn't say much about it's purpose. There is some info on https://europeangodatabase.eu/EGD/About_EGD.php, it's far from complete. Me personally as manager of the EGD, i would be happy to have a clear goal and purpose, a more detailed description which for example would also help to make decisions on what to do when the calculation aglorithm changes! It should be clear if the EGD wants to preserve old ratings or if it always retrospectivley recalculates everything from the start in 1996.

[gennan]

... a more detailed description which for example would also help to make decisions on what to do when the calculation aglorithm changes! It should be clear if the EGD wants to preserve old ratings or if it always retrospectivley recalculates everything from the start in 1996.

This is one of topics where the commission couldn't reach a concensus for a recommendation.

- Preserving old ratings means that it will take a long time before changes to the algorithm will having a noticable effect (including the change to counter deflation).
- Recalculating all the ratings from the start in 1996 will mean that everybody's rating (and rating history) will change suddenly.

So both rating preservation and rating recalculation have disadvantages and we couldn't reach agreement about which is the lesser evil.
So we'll be delegating this decision to the AGM without a recommendation from the commission.

[mumps]

In general these seem to be acceptable recommendations.

[gennan]

It was mentioned that the formulae cited in the commission report don't help much to judge the impact of those modifications.
So here is a hypothetical rating list with EGD ratings recalculated (from 1996-01-01 until 2020-03-15) by the current formulae: http://goratings.eu/RatingList

That list also reflects the commission recommendations to have the Entry Grade floor and the rating floor at 30k.

Some formulae have been updated after the report was made (work on those is still continueing).

Code: Select all

a = (3300 - rating) / 7 
  and conversely B = -7 * ln(3300 - rating)
con = ((3300 - rating) / 200)^1.6
bonus = ln(1 + exp((2300 - rating) / 80)) / 5

The Grade column shows the Entry Grade of players and the RGrade column shows the Grade computer from the rating.

[Pio2001]

Hi,
Thank you for the information and all the hard work.

Something that would need to be stated is if there is a grade limit. Some say that the highest possible grade is 7 dan. Other say that it is 8 dan. Other say it's 9 dan.
The page about the rating system (https://www.europeangodatabase.eu/EGD/E ... system.php) doesn't tell about any limit.

[gennan]

An upper limit to grades was not discussed in the commission, but the new formulae imply an upper limit of 13d/21p EGF, assumed (roughly) to correspond to perfect play. A theoretical entity of this level would be able to give 5 or 6 stones handicap to a top European player (~8d/4p EGF) and 3 or 4 stones handicap to a World champion (~10d/11p EGF). With this assumption, top AI level is assumed to be about 12d/18p EGF.

[Oberlappen]

I think there should be also thought making a new rating class, for tournaments with top players. The class A to C is not bad for normal tournaments I think (I don´t mention class D cause it´s only online), but in my opinion the system should be completed, wit han class S (for Special) or however someone would call.
We have some tournaments for top players, with way more time per move, then in average tournaments. The typical class A tournament, got 1 hour + canadian Byo Yomi (for example the typical 15 stones in 5 minutes), but there tournaments way more time to think, 90 min. + Byo Yomi or even two hours + Byo Yomi and in some even the Byo Yomi got a quite exceeded time, compared to other tournaments.
When we see top players, we can see that it´s quite a difference there, if they play just 1 hour with Byo Yomi or 90 min + Yomi or even more. In the longer games the stronger player reigns more suprior. In the more average timed tournaments, the outsiders do more often beat the favourites on high Dan levels, then in games with more time.

So it does matter for really strong players, if they have jsut the normal class A tournament time or the much bigger time from a Championship. Here really shines the strength of the good players and the outsiders has to do more, to get the win, cause the favourite here can mobilize more ressources with more time. I think an class S rating should be establieshed with tournaments who have at least 90 min and + can. Byo Yomi 15 stones in 5 min. (or the fitting pendant in the Fischer Time). Class S should then count 125% on the rating, instead of the 100% from class A tournament I suppose.
In the end, I think this would be a good solution. And it would be nearly exclusive for top players (only exception is the EGC, where more players got so much time).

[betterlife]

An upper limit to grades was not discussed in the commission, but the new formulae imply an upper limit of 13d/21p EGF, assumed (roughly) to correspond to perfect play. A theoretical entity of this level would be able to give 5 or 6 stones handicap to a top European player (~8d/4p EGF) and 3 or 4 stones handicap to a World champion (~10d/11p EGF). With this assumption, top AI level is assumed to be about 12d/18p EGF.

Nice so we should extend to 12d if the robots are allowed to join!! Currently the strongest amateur player rank is 8d. The system can handle 9d but it's accutally not correctly implemented so atm an 9d would be downgraded to 8d by an egd rating manager.
If a player is entered as new then he gets these initial ratings:

Code: Select all

5d	2500
6d	2600
7d	2700
8d	2800
9d	2670 <= this seems wrong
1p	2700
2p	2730
3p	2760
4p	2790
5p	2820
6p	2850
7p	2880
8p	2910
9p	2940

[gennan]

It does seem reasonable to me to have 8d as the maximum Entry Grade for human amateur newcomers to the rating system.

Perhaps it should even be 7d? Most of the current players with ratings around 2800 have started at an Entry Grade of 7d and worked their way up by winning a lot of games against other top players. 8d is not an amateur grade that is easily aquired in China or Korea (I think you need to be something like top 5 amateur nation wide, which is quite formidable). Only in Japan it seems to occur more frequently, but Japanese dan grades are significantly weaker than EGF dan grades. 8d in Japan could be anything from 5d EGF to 7d EGF.

So I think it should not even be possible to accept 9d amateur as a first Entry Grade. It should only apply to human amateurs that can consistently beat quite strong pros (around 8p or somewhere between #200 and #400 on https://www.goratings.org/en/?). I don't know if such amateurs really exist. Does any of you know a name of such a player, who is still alive and actively playing in tournaments? I think he would be pretty famous (like Nie Wei Ping around 1981). In the exceedingly rare case where such an amateur player participates in an EGF tournament, it would justify manual intervention by the EGD manager to set his first Entry Grade to 9d amateur.

For pros it is a different matter. I think they should just declare their official pro Grade and the rating system should initialize their rating to the EGF rating corresponding to that pro Grade, like @betterlife's list (which is already the case, I think).

[Javaness2]

At the minute, professional players have their rating reset when they are promoted. Resets seem very desirable for improving players, but I don't immediately see why they are wanted for professional players. Is there any plan to change this part of the system?

[gennan]

This was not discussed in the commission, but pro grades are not earned lightly, so personally I feel that resetting seems quite harmless to do in the few cases where the EGF rating is not quite as high as the corresponding pro grade rating. In fact, resetting for professional grade promotions may even be considered a good opportunity to calibrate top EGF ratings.

[jlt]

Many tournaments are being canceled due to Covid-19. This will worsen the deflation phenomenon. I am afraid that if the situation continues like this for some time, ranks will become meaningless.

[gennan]

How does tournament cancelling worsen deflation?

[jlt]

If many people improve but don't play in tournaments, once they come back to competing they will become sandbaggers.

[gennan]

It has always been the case that (on average) players improve between tournaments and this would inflate ratings if nothing is done to compensate for improvement. The EGF rating system tries to compensate by allowing rating resets (when promoting 2 ranks or more) and by giving a small rating bonus for each tournament game played. The correct size of the bonus is a bit of a guess, because the actual average improvement is hard to measure and it probably varies over time and over different regions. With statistical analysis, we found that the current bonus is probably too small, so it will be increased when the EGF rating system will be updated. But it's still a fairly crude mechanism.

During the corona crisis, I don't know if tournament players (on average) are improving at the normal rate (whatever that is) or better. Some tournament players will study and play a lot online and improve, but many others will improve less or even decline, because they have less motivation to study when clubs and tournaments are closed, they don't like to play online and/or they don't learn as much from playing online.

But if some players know that they are really improving a lot during the corona crisis (because they have great success playing online), they can just apply a rating reset in the 1st tournament they play afterwards and that will fix the problem. Unless ofcourse self-promotion is prohibited by the national go association of the player. Maybe such go associations might create a temporary policy to consider promotion requests with some online evidence given by the applicant?