I don't understand. If two players are of equal rank, the winning chance is on average 50%. So how could a player of equal rank "know that their winning chances are 0 - 10%"?

In real life tournaments, you cannot pick your opponents, but on many go servers it is possible to pick your opponents. Because of that, players can avoid playing opponents that have a good recent winning record, or players whom they know are stronger than them in real life. By this they can inflate their server rating relative to the average skill required for that rating.

This is not going to work indefinitely ofcourse, but I guess that players who carefully pick their opponents can succesfully exploit the system to inflate their server rating by 1-2 ranks.

Equal real world ranks in the same system mean about 50% winning chances. Equal server ranks mean that real world ranks are up to ca. +-6 apart. Therefore, some opponents with equal server rank have almost 0% winning chance. This also applies to non-escapers' server dan ranks.

Some people behave very differently on servers than in real life.

Player A is EGF 2k and plays seriously on KGS and is 2d there.
Player B is EGF 3d but only plays blitz on KGS and is 2d there.

In a real-life tournament, B has less than 10% chances of winning against A, but about 50% winning chances on KGS.

The same applies to friendly club games. Winning rates in casual games can be very different than winning rates in tournaments.

People's strength doesn't jump up and down in discrete amounts of 1 stone. You'd expect within 2d there to be near to 1 stone in strength difference between the weakest cohort and the strongest cohort within the rank. In EGF terms think about rating vs rank.

I know, but if two players are 100 rating points apart, the chances of winning of the weakest player are probably much higher than 10%.

When checking statistics from EGD tournament games, winning probabilies near 10% in even games are commonly observed when the rank gap between the players is 3 to 3.5 ranks for dan players and more than 4 ranks for SDK players.

Sorry, misread equal rank as both 2d not equal rating. In a tournament setting the winrate is probably above 10%. Online, not sure, playing environments/player condition are uncontrolled. (Oh definitely wasn't saying it's 90/10 between two 2d's, only that it's not necessarily 50/50. The spread as far as I know in winrate within a rank is not the same at 2d and 10kyu)

I would have thought that even games between kyu players more than 2 ranks apart would be very rare in the EGD stats being usually below the bar?

No offense here, but this is a completely invalid reason. Source is source, you put it out there and smart developers will figure it out and even help clean it up for you (for FREE, typically). You say you have limited resources, yet then push out reasons like that.

Do you think your code base is more complex than the linux kernel, or even many open source games, UI kits, etc? I doubt it... yet many, many people contribute to the linux kernel and various other extremely complex open source projects. I hate to say it but KGS needs a major shake-up or its probably dead, and the mindset expressed in that quote gives little hope that it will actually happen.

I still think that the way to make KGS better today is to improve and promote the HTML5 clients. In particular GoUniverse is underpromoted in my opinion.
Opening the server side source code would have little benefit unless all the work involved goes into improving the HTML5 Client, because the basic design of KGS basically has everything you need, unless you want to turn it into a Casino, or offer the opportunity to play strip 3D-Go.

Second to that is a peculiar reluctance to promote KGS at a basic level

Look at the FFG https://ffg.jeudego.org/php/jouerSurInternet.php
this links to http://elvire.scheibling.free.fr/perso/goweb/kgs.html which is over 10 years out of date

Look at the BGA https://www.britgo.org/gopcres/play#server
It still says that there is only a Java client despite there being several British Admins working on KGS

Okay, people find out how to play online on their own, but isn't it just basic marketing to get your up to date branding out there?

_________________
North Lecale

The BGA page also says

Yes. ShinKGS is over 4 years old, it's a stable client, but here it is described as new.

_________________
North Lecale

Roughly 10% of the EGD games in the range 10k-2d are between players more than 2 ranks apart. That's not very rare, I think.

Ah, from looking at the 2015-2019 data (I figure 2020 data is going to be weird) There's something of a threshold in tournaments here anyway around perhaps 2k where you start being above the bar more often than not and be playing even. Below that it would normally be handicap-1 or handicap-2. Around 90% of even games at mid sdk are G+1 and G+2 but yeah around 10% are G+3 and G+4 which is higher than I thought (I expected it to be less than 5% for mid SDK based on tournaments I've seen). G+1 in the winning rate table on the EGD is all even games up to 1 stone difference?

I think these tables are even games only. I think equal rank is not included in G+1.

I have seen tournaments where MacMahon groups at the lower end are sparsely populated, so it can happen that one needs to play an opponent currently more than 1 MacMahon group away from yours. But in that case, I think handicap is usually determined by the current difference in MacMahon score (minus 1 or 2). AFAIK, only in specific handicap tournaments, handicap is determined by rank difference (minus 1 or 2).

I have analysed raw EGD data quite a bit for the coming EGD update, so if you really want to know more detailed statistics, I could collect those.

The Mc Mahon score is rank+nb of wins, and tournaments are usually played at handicap-1, so on round 4 of a tournament, an 8k with 3 wins can play an even game against a 4k with 0 wins.

Ah, I see where my thinking went awry. We run *small* tournaments so we use strict -1 or -2 rank handicaps below the bar rather than try to use McMahon group based handicaps. I'd completely forgotten that this wouldn't be the norm in tournaments in many European countries with fuller mid-weak kyu fields. My apologies, I really shouldn't have made that error.

Re: EGD stats, the things that have most interested me recently as been how the expected win rate against G+1 evolves with rank and whether are inflection points if plotted out against GoR ratings. I'm curious about the relationship between increases in consistency of results and increases in strength. It's only an idle curiosity though rather than anything worth you putting time into.

You can see the observed winning probabities against G+1 directly on the EGD statistics page.

Around 5d-6d it is about 25% (~190 Elo per 100 GoR)
Around 3d it is about 33% (~120 Elo per 100 GoR)
Around 1d it is about 36% (~100 Elo per 100 GoR)
Around 2k it is about 40% (~70 Elo per 100 GoR)
Around 15k-3k it is about 43% (~50 Elo per 100 GoR)

This table shows that ranks don't have a constant width when expressed as winning probabilities or conversely Elo rating gaps.
As ranks go lower, the Elo gap between consecutive ranks gradually decreases (is this what you mean by "consistency"?).
As ranks go higher, the Elo gap between consecutive ranks increases, more sharply as the rank gets higher and higher.

Below 15k, the winning probabilities go down again a bit, but I think this is an anomaly caused by the disturbance of the rank floor of 20k that the EGD currently uses (soon to be lowered to 30k). I assume that without this disturbance, the winning probability would be about 45% around 30k, increasing to 50% when approaching random play.

A non perfect player can never beat a (hypothetical) perfect player, so the non perfect player would have a winning probabilty of 0 and the perfect player would have an Elo rating of infinity. From handicap games against pros, it looks like AI should be ranked about 12d EGF. A perfect player might have a rank of about 13d.

Also see Elo per stone in Go on this page.

I disagree. A non perfect player has a tiny but nonzero chance of playing a perfect game. In a 300-move game, a random player has about 1/(361 x 359 x 357 x ... x 63) ~ 10^-788 (Edit: 10^-342) chances of playing the whole game perfectly.