Life In 19x19

RobertJasiek wrote:Let you be told for the 20th time: It does happen that a 5d becomes the champion.

Let me tell you 21st time that if that is about to happen, we know that before hand and we can use wildcard option.

RobertJasiek wrote:

Robert, it does your cause no good.

Quite contrarily my ability and desire to distinguish between open and hidden personalities does the cause good.

Which wasn't my point. My point was using a condescending tone is unnecessary and damages people's opinion of you.

RobertJasiek wrote:

With respect, you are also not sufficiently strong to be an EC challenger,

What is an "EC challenger"? This expression is used in systems that end in a 2 persons match.

Someone capable of winning the European title.

RobertJasiek wrote:

and this is partially a vote in favour of the current system being good enough not to change.

This is whose opinion? Yours? The commission's?

The OP's

RobertJasiek wrote:

The current system has issues that concern a number of people - that does not make the system "not very good"

If a system is called "very good", then the quality is so good that no change is needed. If a system's basic structure is called "very good" while the system's details are called "worth improving", then change is suggested for the details; the overall system is thus not "very good" but maybe "good". I.e., I want honesty and clarity in the statements instead of dull propaganda.

That is a subjective definition of very good. I have never read an objective definition of "very good" that means "no change required". It is also not enough of a superlative to mean "no improvement is possible".

RobertJasiek wrote:

except perhaps in the eyes of a vocal minority.

Like the AGM? The politicians would be the vocal minority? And who is the majority, how do you assess it except by guessing wildly?

I agree, but to me improving this system is much more important than tiebreaker semantics. Finding a way of collecting individual views from those most relevant to impact the future of the system is important

RobertJasiek wrote:"marketing gag, kyu players in the EC"

Your original comment was EGC, which I saw as the European Go Congress rather than the European Championship. The change makes my original comment not relevant.

RobertJasiek wrote:

[Robert: "It is not over-sized." Again, this is only your subjective opinion, and I don't believe it is based in facts.

Because you do not read the facts. The facts are that 5d can sometimes become European Championship (see 2001), that (currently) almost only the latest ratings are used for seeding the Europeans and that fast enough rating changes for a currently strongly improving 5d into the top 10 or 16 present rated Europeans is pretty hard.

That was a bizarre championship that does not accurately reflect typical trends.

RobertJasiek wrote:

Some stats and data from the last 8 years:

You forge statistics brilliantly. Evaluate 9 years for greater precision!

Pot. Kettle. Black. You've seen the trend of those 8 years, so you choose to cite a championship which I excluded because the data was not ordered on the EGD. Having now taken the extra time to investigate, I find a tournament with only 6 European 6d players, and no 7d players or professionals. It's an extreme end due to non-participation of really strong players, and in this particularly case the 5d may well have found himself in the supergroup even if it was only 16 players (including foreigners). To use this as evidence that the supergroup should be big enough to include 5d players is horrendous statistical manipulation.

RobertJasiek wrote:

* It is very clear to me from this data that 5 dans are out of the running from the beginning, they are simply not title contenders.

See above. You are wrong.

If I extend my comment to state "... not title contenders where 7 dan + players were present", can you still show it to be wrong?

RobertJasiek wrote:

* The number of European 6 and 7 dans combined have never exceeded 13 (and only 3 times out of the 8 did they even exceed 10!)

Thank you for the partial statistics.

And if you include 2001, where the figures of EU 6-7d+ is 6, and no non-EU 7d+, leaving the smallest speculated supergroup size out of the entire dataset.

RobertJasiek wrote:

It is very hard not to conclude that having 24 Europeans in the super group seems like the net is being spread far too wide

Within your tight view inside forged statistics. Forged because deliberately your exclude all contrary information.

Forged statistics? I have looked at the last 8 years and you cherry pick a single carefully chosen year to refute it - your forgery is the greater I'm afraid. Also, your example year of 2001 supports the case also, as the supergroup would naturally have to include 5d players because of the lack of strong participation.

RobertJasiek wrote:

What is the justification for 32 with the expectation of 24 Europeans?

That 5d do have a chance. See 2001.

I didn't ask about 5 dans, I asked about the total number. In your exceptional example, the 5d would be in the supergroup without it being big. The number still requires justification.

Liisa wrote:Tiebreaker is not same as fair tiebreaker. You can also use players horoscope signs as a tiebreaker.

SOS etc. are neither fair nor unfair - they are unpredictable. Playing lightning playoffs is a fair tiebreaker because the players' own performance is measured directly.

Using dynamic amount of rounds is extremely poor tiebreaker

Dynamic rounds are NOT a tiebreaker but enable the first result criterion (NumberOfWinsScore or MMS) to prevail!

and I would rather use horoscopes.

This should a very bad attitude of yours towards the first result criterion.

It seems that you have difficulties of understanding that for the super group size only thing that matters is what is the skill distribution of players.

What matters is
- strength distribution of players
- number of rounds
- objective of whether to be sure to have a possible strongest players in

Even the number of rounds is irrelevant! The larger is the skill distribution the less there are rounds for strongest players to play each other and solve mutual arrangement

Rather than the number of rounds being irrelevant, it is also relevant. That there is some interdepence to strength distribution of players does not imply that it would be an identity of cause.

From the pairing perspective it is not very good idea to use folding, but split and slip is better pairing method, at least for the first rounds.

For the first few rounds, there seems to be agreement among experts that fold is not the best.

For latter rounds split and random is also possible.

For later rounds, why should fold not be as good or better?

because your reasoning skill and reading comprehension is just flawed.

LOL

entertainment is unknown and flawed subject for jasiek.

Wrong.

Entertainment for the kyu player about the EC comes from watching EC players, games or results. (Maybe for some also from boasting to have also played in the EC.)

For this entertainment, it would be better of main tournament and EC would be at different times so that kyu players could see more of the EC (top) games. That kyu players have a shorter thinking time helps them only partially.

However, experience tells us that only very few do watch top games live. Between 0 and 10 in room 1. Maybe one or two dozen in a video room. What does attract kyus much more is live commenting; then the rooms are full. However, they are fuller during afternoon commenting!

So although the live entertainment does play a role, it is a small role. On average, less than 10% watch live. Maybe you are an exception as much as I am (I watch rather a lot live). Such exceptions are too rare to play a significant role though.

Liisa wrote:we know that before hand and we can use wildcard option.

A wildcard would very unlikely have identified Kulkov. Instead they would have chosen van Zeijst or whomever. Politicians are weak at predicting playing strengths well!

RobertJasiek wrote:

Liisa wrote:we know that before hand and we can use wildcard option.

A wildcard would very unlikely have identified Kulkov. Instead they would have chosen van Zeijst or whomever. Politicians are weak at predicting playing strengths well!

Kulkov's Gor was 2555 before the tournament. By any standard he would have been included to the super group.

{addendum: Even perhaps in Groningen that was very high level tournament, because 2555 is higher gor than what some 6 dan players has.

It is good idea for you if you just stop thinking that you are 5 dan, but think that your rating is EGF 2404, because that is more proper estimate of your current skill.}

topazg wrote:My point was using a condescending tone is unnecessary and damages people's opinion of you.

If people want their opinion to be appreciated higher by me, then they should have the courage to write under their real name. Maybe some people lacking that courage disagree with me. So be it. I do not raise my opinion on lacking courage just because they would prefer me doing so. I appreciate courage more than lacking courage. If that damages some people's opinion about me, they might reflect to change theirs and acquire greater courage. I do not consider my advertisement for greater openness and courage unnecessary, as you seem to suggest.

I have never read an objective definition of "very good" that means "no change required". It is also not enough of a superlative to mean "no improvement is possible".

The scale of evaluation is wrong when something is called very good but still significant changes are being suggested. A thing should thus rather be called "good with easy potential for very good in case of moderate changes".

That was a bizarre championship that does not accurately reflect typical trends.

We can assume that the trend is, on average, towards 9d. Maybe after another 15 years, a 5d would not have any further chance in any year, not even in case of an EGC in Iceland. It is, however, wrong to confuse future with present time. Currently we do not have that many 7d to be sure that some 7d attends in each year. A congress with only 6d or below is pretty much still possible. Then a 5d does have his very realistic chances. Even an ordinary 5d having an above average tournament for himself.

because the data was not ordered on the EGD.

Never trust only one source of history;)

It's an extreme

Juding from the last 9 years. Expand to the last 30 years and you might change your mind.

in this particularly case the 5d may well have found himself in the supergroup even if it was only 16 players (including foreigners). To use this as evidence that the supergroup should be big enough to include 5d players is horrendous statistical manipulation.

Has anyone done the actual rating statistics for the supergroups in all years?

And if you include 2001, where the figures of EU 6-7d+ is 6, and no non-EU 7d+, leaving the smallest speculated supergroup size out of the entire dataset.

That would be saying: We may as well forget about the 2001 title holder...

the supergroup would naturally have to include 5d players because of the lack of strong participation.

Hear, hear.

The number still requires justification.

In either direction of size. Ok. We are approaching the assessment problem a bit more closely. With more data and evaluation, we might actually get somewhere. Like finding the optimal supergroup / EC size, depending on other parameters.

Liisa wrote:Kulkov's Gor was 2556 before the tournament. By any standard he would have been included to the super group.

Ok, thank you, that is a convincing argument! Now does that mean that, under "normal circumstances of other parameters" (like only a modest number of non-Europeans), a likely good number for the supergroup / Swiss / round-robin size would be 22, 20, 18, 16, 14, 12, 10 Europeans? Just which of them is large enough to safely include the currently strongest European?

And what does a small supergroup in a McMahon actually mean for SOS? Top players would more likely also get players of lower and yet lower MM groups of opponents.

You could calm me quite a bit if only you would abandon usage of numerical final result tiebreakers. The best supergroup size does almost nothing (positive or negative) for a tie situation after the last round.

RobertJasiek wrote:

topazg wrote:the supergroup would naturally have to include 5d players because of the lack of strong participation.

Hear, hear.

On the offchance you may have misunderstood, I mean for that tournament, not as a general rule.

RobertJasiek wrote:

Liisa wrote:Kulkov's Gor was 2556 before the tournament. By any standard he would have been included to the super group.

Ok, thank you, that is a convincing argument!

I added this afterwards, but your reply skill is too fast.

{addendum: Even perhaps in Groningen that was very high level tournament. That is because 2555 is higher gor than what some 6 dan players has.

It is good idea for you if you just stop thinking that you are 5 dan, but think that your rating is EGF 2404, because that is more proper estimate of your current skill.}

Now does that mean that, under "normal circumstances of other parameters" (like only a modest number of non-Europeans), a likely good number for the supergroup / Swiss / round-robin size would be 22, 20, 18, 16, 14, 12, 10 Europeans? Just which of them is large enough to safely include the currently strongest European?

Like I previously stated that group size itself is irrelevant, but skill distribution within group should not be too large. If you cannot understand this reasoning, then it is bad for you. If you want grude estimates, then I would say that if GoR is over 2550 then it is certain ticket for any super group. Even if that would mean 64 player super group. Usually 2550+ player is strong enough to challenge the top Europeans.

And what does a small supergroup in a McMahon actually mean for SOS? Top players would more likely also get players of lower and yet lower MM groups of opponents.

This is the reason why we need middle group below super group. And also we could prefer second game for already once played pairs if there is no other opponent available in same McMahon-group. (I am not sure how well this will work in practice but 10 rounds might be enough for this)

You could calm me quite a bit if only you would abandon usage of numerical final result tiebreakers. The best supergroup size does almost nothing (positive or negative) for a tie situation after the last round.

Solving tie is not important by itself. Only thing that matters is that we can have as much as possible relevant data. If that is still tie then it is a tie and there is nothing that we can do. We can only solve ties by increasing data (i.e. increasing available relevant rounds). Creating artificial tiebreakers like reducing the amount of rounds dynamically, creates just superficial order for players, but that is false order, if it was gained by reducing available data.

Liisa wrote:It is good idea for you if you just stop thinking that you are 5 dan, but think that your rating is EGF 2404, because that is more proper estimate of your current skill.

Both kinds of values (increasing ranks and dynamic ratings) have their good purposes in principle. I do not believe in the accuracy of any existing Go rating system thus far though. They all make bad assumptions, have design flaws etc.

Skill (principle ability) is something different from strength ("current" performance). I think what you mean is current strength.

Current is somewhat unclear though because a rating system tends to evaluate lots of older games, too.

Ideally what a rating system does is to predict a current probability of (about) 50% against equally rated players.

Like I previously stated that group size itself is irrelevant, but skill distribution within group should not be too large. If you cannot understand this reasoning,

I disagree. The group size in itself is relevant, too. E.g., suppose you had a group size 2000 (all players very close in strength). It would be too big! 10 rounds McMahon do not produce a particularly meaningful result.

I agree that skill distribution within a group should not be too large - but what is too large...

This is the reason why we need middle group below super group.

I presume that a middle group slightly relaxes the problem. It does not solve it at all. The problem is solved if it is dissolved: Do not use tiebreakers for the final results!

And also we could prefer second game for already once played pairs

In McMahon / Swiss, they are very much disliked by most.

Creating artificial tiebreakers like reducing the amount of rounds dynamically,

I think nobody suggests such. What is suggested is INCREASING round numbers dynamically BEYOND a certain minimum.

(Why would reducing the amount of rounds dynamically be a "tiebreaker"?! It would be a heavy impact on the first result criterion, but that does not make it a tiebreaker.)

Oh dear, an argument between RJ and Liisa. This it not going to end anytime soon.

This is just another of too many cases of Robert trolling. The condescension toward Liisa qualifies as a TOS violation in my opinion. It's one thing for Robert to start his own thread in the Rules Forum and pontificate scholastically for eternity, it's quite another when he hijacks a thread and begins by insulting the OP. Topazg enumerated the offenses beautifully, so I won't repeat them. I see absolutely no reason why this type of behavior should be allowed or given free reign on L19.

Robert, knock it off!

deja, trolling is flooding without factual discussion, which is not the case for me. Criticising somebody for hiding his identity must be possible; asking for someone's name is not insulting him. If everybody regardless of his rank had the same right to the EC, then we would be having an all participants supergroup, which would defy the tournament's purpose of being a championship. Therefore, when somebody claims to have a personal interest in playing in the EC, it is essential to know his playing strength.

For your reference, "It might be a troll, liar without taking responsibility, multiple namer or person paid for upholding an opinion." does NOT refer to the OP but to the principle possibilities given due to anonymous posting.

Can we now end the meta-discussion?

RobertJasiek wrote:trolling is flooding without factual discussion, which is not the case for me.

Usually trolls them-self do not understand that their flooding does not have much factual content. Most of your "discussion" is based on that you have difficulties of understanding semantics. E.g. you have no idea what even trolling means. Often case trolling does not require that it is intentional activity, e.g. it is possible that some religious trolls do not mean to troll, but they do it because they do not have enough social intelligence to grasp what other people are thinking from the discussion and its factual content. And they continue and continue writing without delay and getting tired and are unable to read between lines cues for other people's emotions.

Understanding semantics and what is going on on other people's minds is of course very difficult and lot of people (me included!!) have great difficulties with it. Problem with you is that you do not have enough self-criticism that you could recognize and admit your flaws in the reasoning and behavior. But you act like you know everything and that your personal opinions would have more relevance than other people's opinions.

It was my mistake of course to react your trolling, but I did not suspect that your purpose (intentional or not) was mainly to flood this thread because it is not for your likings. But as I know other EGF rules commission members, I should have guessed it.

I hope that this (psycho-)analysis is not correct, because I do not know you well enough (it is mainly derived from other people's implicit attitude towards you). And certainly I would hope that it is not necessary to return this meta-discussion, at least not in this thread.

RobertJasiek wrote:Criticising somebody for hiding his identity must be possible

Who is hiding her identity? Perhaps it is hidden only from trolls. But if you are unable to connect my EGF rating to the EGD rating list, I can make it easier for you and I publish link to my home page.