Monday, February 18, 2008

Meet Katrin Wills

This is old news but I have to mention it because I had no idea about the central character of this post until this afternoon.

Chess Victoria has a new boss and her name is WFM Katrin Wills. Some of our readers may remember her as Katrin Aladjova, a former winner of the World Girls U18 in '89 and the World Girls U16 in '86. Back then she played under the Bulgarian flag.

Apart from running a state chess association, she also runs her own make up business. ("The key to my success is the constant improving of my skills to keep up to date with the latest changes within the fashion and the make-up industry".) Just pop on over to www.makemeup.com.au and check it out.

Now Katrin might be an ex world champion, but according to our Australian ratings officer, she's just, well, a little better than a complete beginner! Her ACF rating? 1667! She can thank Glicko for that.

49 comments:

DeNovoMeme said...

1667. Wasn't that the year Glicko was invented by a hatter on acid.

Anonymous said...

It's pointless poking fun at Glicko over this. The fact is that Katrin had played no tournaments for several years before she played in Bendigo last year and Glicko gave a reasonably accurate indication of her then form which is what a rating system is supposed to do. Rating systems don't exist to give some notional/theoretical idea of a player's worth if they were playing at their best.

DJ

Anonymous said...

katrin lost the 500 or so points at ballarat over 7 or so games in 3 days - a very small number - there should be a threshold on the number of points lost at a single event. because of this glicko creates a lot of anomolies. if you asked her to play fritz 50 times at the same time limit and get fritz to estimate her performance i'm sure it would be 1900'ish. a similar player lost 500+ points in 3-days - it put new meaning to 'having a bad weekend'.

Anonymous said...

It was the Ballarat Begonia tourney where she lost hundreds of points, not Bendigo.

Anonymous said...

If anyone seriously contends that she should be a 2100 rated player, then they must think she is just as likely to perform 500 points upstream as she is to perform 500 points downstream. Looking forward to that 2600 performance. :P

DeNovoMeme said...

Name an active 2100 player in Australia who could lose 500 points in 7 games.
Cannot?
Katrin has been Glickoed!
Why would anyone try to justify this assault on common sense? Because they are matey mates with their good mate Gletsos - the bloke with statistically statistical uncommonunsense.

The system functions like a person with bipolar, who needs their medication reviewed by the ACF (which has lost its license to practice.)

Anonymous said...

Matthew's comment is typically misleading. Katrin had not played a rated game for many years and so was not active. She had not played a FIDE-rated game since at latest October 2001. So the comment has no bearing on the result of Katrin's rating after Ballarat 2007. Her current ACF rating is 1667?? Indciating a very unrelaible rating that comes about only through inactivity of one kind or another. Posssibly the only thing less reliable than Katrin's rating is Matt's knowledge of the ratings system.

DJ

Kevin Bonham said...

Re the third post, Fritz11 currently estimates my performance at 3017 after 48 games (47 of which were blitz in "unleashed" mode, of which I drew five and lost the remainder) so I don't think Fritz's assessments would be much of an indicator of anything!

Glicko-2 is very sensitive to good or bad results by a player returning from a long absence, but what this reflects is that their old rating is statistically almost meaningless after so much time since they last played. If she plays in more tournaments and performs at anything like her former rating in them then her new rating will rapidly head upwards again.

doubleroo said...

Katrin 1667? After one bad weekend? After all the anomalies (like when I got to 2582 while in europe) there are still people defending this "rating system"

Then again there are people defending "Creation theory"

DeNovoMeme said...

Denis' comment is typically misleading, as is the case with (such) lawyers. But sometimes the truth slips out from the maw of law. For example

"Her current ACF rating is 1667?? Indciating a very unrelaible rating ..." - DJ

Why would such a STUPID number even be published? It is an insult to the woman AND it makes Glicko look exactly as stupid is it is, as STUPID as the ACF.

The ACF ought to tell Gletsos to Getlost, but it will not because by dumping a idiot, its mean IQ would drop even further.


"Matt's knowledge of the ratings system." -DJ

Well, matey, whenever you wan to go toe to toe with me on the strengths and weakness of Elo and Glicko, just email me. And don't forget to bring your trainer, Gletsos. I will wipe the floor with both of you. But then again, maybe I should pass it up - it is to much of a mismatch with all the charm of a man kicking a tethered dog.

DeNovoMeme said...

Bonham: Fritz11 currently estimates my performance at 3017 after 48 games .

DNM: When you finish with yourself, clean up the mess, and do something useful in the ACF, like run the Olympiad Appeal.

Kevin Bonham said...

Matt, I hate to break it to you, bearing in mind how fragile you are about these things, but an Olympiad Appeal Co-ordinator has in fact already been appointed.

I think this is the second Olympiad in a row where I have had some very silly people suggest to me that I should run the appeal as well as the selections, and make this peculiar suggestion *after* the appeal co-ordinator position had been filled!

The only time you tried to wipe the floor with an ACF rating system by submitting a prediction that you thought would effectively second-guess it, you were the one who ended up towelled, by a very wide margin indeed (see http://www.chesskit.com/auschess/cgi-bin/yabb/YaBB.pl?board=auschess;action=display;num=1070622981 for the details). That was over four years ago now and despite your various empty claims to be cooking up something superior, all we have had from you since is bile and theoretical bluster from a person who pretends that he understands ratings but is not willing to invest the hard yakka required to come up with an effective critique or alternative.

By the way, I don't personally know the new CV President who is supposed to be the subject of this thread, but I wish her all the best on the hefty administrative challenges facing said body, which she acknowledges in her report at http://chessvictoria.netfirms.com/cv_exec_announcements.htm and which are far more interesting than the level of her resurrected ?? rating.

Of course, the ratings publication issue you raise is a damned-if-you-do-and-damned-if-you-don't situation; if we did not publish the ratings of those who had ?? after their number we would hear no end of complaints from those who saddled up as an interclub sub for the first time since 1973 and didn't even get a fresh shiny mint new rating for their troubles. Matthew, are you saying that "??" ratings should not be published?

The Closet Grandmaster said...

The URLs in Kevin's post, in the order he posted them are (i) this and (ii) this.

- TCG

Anonymous said...

What Kevin Bonham and Denis Jessop continuously fail to acknowledge is this:

Several past and present people in chess administration claim that the Glicko system, especially in its current form, leads to significantly inaccurate ratings. The criticism is NOT limited to Matt Sweeney.

Who?

Oh, I don't know? Lets start with:

(a) myself
(b) Katrin
(c) Phil Donnelly
(d) Peter Parr
(e) Brian Thomas
(f) countless others.

One of the biggest criticisms of the inaccurate Glicko ratings is that the RD's (which loom large in the calculation of Glicko Ratings) are not made known to the general chess community.

Perhaps this is why the Queensland Chess Association unanimously passed a motion for the ACF Ratings Officer Bill Gletsos to release these.

If the RD's were made known chess players could, to an extent, manipulate their RD's and consequently their ratings. This would likely have the effect of increasing the accuracy of their Glicko Ratings.

However, whether the recalcitrant ACF will eventually bow to common sense is yet to be seen.

Best,

AO

Anonymous said...

Matt has delivered a characteristically moronic outburst about two question marks. His tiny mind seems unable to cope with the fact that a player who has had very few games rated has a very unreliable rating, though not necesarily an inaccurate one. That system of indicating reliability was used under the previous system as well as the Glicko system. Moreover, Katrin's rating before Ballarat would equally have had a "??" as she had played so few games in recent times. Rather than call for Bill Gletsos to stand down as ACF Ratings Officer - perhaps the stupidest suggestion I have heard in the last decade - Matt should go away and try to learn something about rating sysyems of which he is now abysmally ignorant.

DJ

Kevin Bonham said...

Matthew Sweeney has now claimed (in his usual place of disrepute - don't know why he cannot do it here instead) to have a new ratings algorithm that will only be unveiled once Bill Gletsos is no longer ratings officer.

However there is no reason why anyone should believe Matt's claim to have successfully tested a superior alternative, since even if he has one and believes he has tested it successfully, there is no guarantee he has not made calculation errors (as he sometimes has in the past.)

There is also no reason why he should not wheel out his alternative now, since if it was indeed successful and was adopted, it would surely be a substantial propaganda victory for him.

Furthermore, Matthew could show that his alternative was working even without revealing the herbs and spices, simply by posting a large number of predictions generated using it, that could then be compared with those of the existing system.

Matthew has now suggested that only !! ratings should be published. Just imagine the uproar if the ACF suddenly came out with a rating list containing less than 500 names!

Matthew also engages in personal abuse and trolling of an unfactual sort that is not at all helpful to his implied claim to be a sincere person, and makes me less inclined to take his claims about having developed an effective algorithm seriously.

Anonymous said...

The anonymous Alex Toolsie has made a significantly insignificant contribution here though I am at least amused to find that he and "countless others" are listed among the chess administrators who don't like Glicko. He then moves to a point that has been extensively canvassed elsewhere but is irrelevan to the current discussion, namely the public release of RDs. This was an issue raised in the ACF Council several years ago when there was no support for their release. When it was raised again recently by CAQ delegate Brian Thomas, ACF Ratings Officer Bill Gletsos explained that such a release is at best useless and more likely misleading. The assumption that knowledge of RDs allows one to calculate one's own rating is false and I believe that Brian accepted that explanation as it was made in a meeting at which I was present.

DJ

Anonymous said...

Katrin was in Sofia last year, she moved pieces for Topalov during the blindfold exhibition.
Boy, she talks more than Irina Krush!

Kevin Bonham said...

What a bizarre contribution from Alex Toolsie (AO) - yet again he has spectacularly little clue what he is talking about.

This can be seen especially where he writes "If the RD's were made known chess players could, to an extent, manipulate their RD's and consequently their ratings. This would likely have the effect of increasing the accuracy of their Glicko Ratings."

This is wrong for several reasons. Firstly players can manipulate their RDs anyway, simply by playing more or less chess, and it's not rocket science to work out that the more you play the more stable your rating becomes, unless your playing strength is rapidly changing.

Secondly very few players understand the Glicko system well enough to be able to determine the exact impacts of their decisions on their RDs anyway.

Thirdly while a player who thinks they are underrated might think that not playing for a while to bump up their RD would help their rating to increase, most likely it would actually increase faster (assuming that they were underrated) if they simply played more chess.

Fourthly the very idea that encouraging a player to manipulate their rating makes it more accurate is lalaland stuff typical of Toolsie's clueless buzzing and bumbling on virtually any issue he discusses. Rating systems where players can very easily manipulate their ratings are not more accurate at all, but are more prone to inflation and therefore less credible. Certain online rating systems suffer from this defect.

AO falsely claims that I do not acknowledge that several people have complained and in some cases whinged about the Glicko system. Of course I acknowledge it - just as at least as many people have complained about every other rating system the ACF has had. However, it's notable that a large proportion of those complaining (not all, but many) are (i) those like Sweeney and Toolsie who are anti-Bill-Gletsos for reasons other than ratings (ii) those who have an obvious bias in the debate because their own fossilised and nearly-irrelevant old ratings have become vulnerable to rapid points loss should they not play up to their former standards (iii) those whose playing strength appears to be becoming more erratic and who want to blame Glicko's handling of junior ratings for their woes.

It is notable that no-one complaining about ratings points losses for inactive players has produced empirical evidence to show that the new (Glicko-2) ratings are less predictive of their subsequent results than ratings systems allowing only a small drop per game (such as ELO) would be. The reality is that while some players will return after several years off and immediately pick up where they left off, many will not, and if they can't produce their old form in their first tournament back then the rating system is quite correct to not expect them to do so in their second. Of course there will be overshoots and undershoots where a player plays a single tournament after many years of inactivity, but the solution is simple: if you are an inactive player and you want to have a rating that reflects your new ability, *play more chess*. It seems, however, that an accurate rating is exactly what some of the inactive critics of Glicko are afraid of.

The publishing-RDs campaign is a misguided beatup in which opponents of the ACF have picked an issue to complain about and make noise on, but it is really just a pretext for continued contrived conflict with the Ratings Officer, and one that would not deliver what its supporters claim on its behalf. If one seriously thought Glicko-2 was so inaccurate, one would consider the publication or otherwise of the RDs as a tenth-rate issue at best.

DeNovoMeme said...

Denis says of me: "His tiny mind seems unable to cope with the fact that a player who has had very few games rated has a very unreliable rating, though not necesarily an inaccurate one."

The day you, a back-seat-driver at the ACF, can debate with me on matters relating to statistics and data interpretation, is the day I put a gun in my mouth.

You remind me of Paul Keating. He was an idiot savant for politics. He knew nothing about economics but talked the talk. You are an idiot savant for gas-bagging about ACF reform. Talk the talk but can do absolutely nothing. Now you want to add mathematics too your repertoire of embarrassing faux pas. Best that you shut up before you reveal that your mind is past its twilight.

DeNovoMeme said...

ACF VP KB says: "Matthew has now suggested that only !! ratings should be published. Just imagine the uproar if the ACF suddenly came out with a rating list containing less than 500 names!"

The uproar is that the 500 with !! should be 5000 with !! and the ACF is 100% responsible for that state of affairs.

Kevin what is the ACF's target membership number for active chess players in Australia? Oh, that's right, the ACF doesn't have one. Go away you horrid little boy and learn not to throw stones.

And to end, I will be pulling out of this thread. If you want to be humiliated you can try playing in an away-game at http://www.ozchess.com.au/

Kevin Bonham said...

Matt, I certainly won't be playing an away-game when the opposition is fifth-rate at best, one of the refs is a do-nothing who is totally unsuited to the position, and the chief arbiter dabbles in the game to attempt to distort it while effacing evidence that he has done so. You evidently cannot cope with even a neutral venue.

Ten times more players with !! ratings would presumably mean ten times more players full-stop, which on a nationwide basis would mean that at any given time more than one in every thousand Australians (of all ages) would be active tournament chessplayers. That would put us on participation levels that very few nations in the world could match. It would be bloody nice but is very unlikely to happen and considering us 100% to blame for it not being the case is just another example of Sweeney's exaggerated and stupid trolling act.

As for throwing stones, I couldn't even make a (true) comment about the inaccuracy of Fritz-assigned ratings and a factual comment about the Glicko system, no personal attacks on anyone in that post, without Matthew Sweeney doing exactly that, with:

"When you finish with yourself, clean up the mess, and do something useful in the ACF, like run the Olympiad Appeal."

Whether Sweeney intended this as a suicide or a masturbation gag is unknown to me, but in either case the real mess that should be cleaned up is Matthew Sweeney's trolling.

Kevin Bonham said...

By the way, I had forgotten it (even though I posted on the thread!) but Matthew had yet another ratings system prediction failure, see

http://www.chesschat.org/showthread.php?t=1599

I'm betting whatever he's claimed to have cooked up is a similar failure and that his claim to have anything workable ready to go should his nemesis depart is pure bluff.

Anonymous said...

I see that Matthew has now, in typical fashion, explicitly run away from this debate and sought refuge on a BB populated largely by clueless clowns (to be kind to them). That's a pity because I am more than happy to see him put a gun in his mouth and fire it. Moreover, with a mouth like his, he could use a double-barrelled shot gun and discharge both barrels simultaneously without hitting anything but bits of his skull-bone. Unfortunately the offer may not be open as I was not "debat(ing) with (him) on matters relating to statistics and data interpretation". I do however acknowledge his expertise in the area of doing nothing as that's just what he has done when on a couple of occasions he promised to do something for the ACF and, so I am told, when he was an NSWCA Councillor.

DJ

Anonymous said...

I will get to Kevin Bonham's nonsensical self contradicting gobblygoop probably tomorrow or on the weekend. Wading through the sewerage in Denis Jessop's post is all I can do in one sitting.

Denis Jessop states:

When it was raised again recently by CAQ delegate Brian Thomas, ACF Ratings Officer Bill Gletsos explained that such a release is at best useless and more likely misleading.

Notice here how Denis doesn't even address the issue I raised that the CAQ unanimously passed a motion requesting the ACF Ratings Officer to release the RD's? That includes the very experienced CAQ Ratings Officer Patrick Byrom who voted in that motion.

All Denis is saying is that they fed Brian Thomas a lamb excuse about why the business as usual approach was okay, and then sent Brian Thomas off with his tail between his legs! Good effort Denis. But your insinuation that Brian is now a convert to the 'keep RD's secret' school of thought is either overly optimistic or outright disingenuous. When I spoke briefly to Brian about this at the Surfers Open he didn't seem too 'won over'.

Then Denis Jessop states:

The assumption that knowledge of RDs allows one to calculate one's own rating is false ...

How exactly is it false? Please elaborate here Denis.

Best

AO

Anonymous said...

I've changed my mind and will throw Kevin a little bone too, so he doesn't get jealous of Denis Jessop.

I wrote:

If the RD's were made known chess players could, to an extent, manipulate their RD's and consequently their ratings. This would likely have the effect of increasing the accuracy of their Glicko Ratings.

To which Kevin replied:

This is wrong for several reasons. Firstly players can manipulate their RDs anyway, simply by playing more or less chess

Its not wrong at all you blockhead, because you'd be able to manipulate your RD far more precisely if you actually knew what your RD was in the first place!

Best

AO

Kevin Bonham said...

Aaaah, more Alex Toolsie garbage - all theatrics and abuse, no substance, foot firmly lodged in mouth as always. Often you can obtain a lot of wisdom from these stupid Toolsie posts simply by assuming their negation to be true.

Since Alex believes that if one knows one's RD precisely one can manipulate it, let's see him show us how it's done:

The March rating list will soon be released. Alex, let's assume for the sake of argument that your new RD (whatever your new rating is) is 110, that that of every !! player is 50, that that of every ! player except you is 70, that that of every player with a blank is 140, that that of every player with a ? is 205, and that that of every player with a ?? is 300.

Now, given those assumptions, I want you to tell me in what way you would manipulate your RD so that it is somewhere between 75 and 85 by the end of the ratings period after next - and provide a proof that your solution is correct. If there is not enough space for you here feel free to waste a post in your gadget-obsessive junkbox of a forum on it; even if it was an abject failure it couldn't be any more useless than at least 90% of the rest of the posts there.

Since you maintain that "you'd be able to manipulate your RD far more precisely if you actually knew what your RD was in the first place!" you should find showing how it's done a simple task! Good luck!

Anonymous said...

Kevin Bonham demonstrates his jealously of OzChess' new features with this comment:

in your gadget-obsessive junkbox of a forum on it

Please don't be jealous Kevin. Its not my fault that Karthick is too lazy or apathetic to give Chess Chat any new features. When was the last time he bothered to add any new features to Chess Chat anyway? 2 Years ago? More perhaps?

Anyway, stop distracting me with your babbling hypothetical calculations. You need to look at the big picture to understand the simple concept I've referred to Kevin.

Namely, if you know what you're RD is, obviously you can manipulate your rating by playing tournaments when its high, and avoiding playing tournaments when its low.

The means if you wait until its high, and you have a good tournament, then your rating should skyrocket. Thats how RD's affect one's rating in Glicko.

Comprende Vous?

Best

AO

Kevin Bonham said...

Aaaah, the "jealousy card", how feeble. Attributing somebody's comments on the internet to jealousy when there is not the slightest evidence that that is the case is a certain hallmark of a very incompetent debater - so much so that one of my other forums has a thread where those foolish enough to wheel out the card in question can be named and their feeble efforts laughed at.

The distraction here is your waffling about trivial forum features, which are not having any discernible impact on the extent to which posters use your forum. Chesschat has not upgraded to the very latest upgrades of vBulletin because it has an excellent custom-built game viewer and the effort to recode the same with each upgrade is generally just not worth it. As for "spoiler" tags, we have had a perfectly adequate solution to the spoiler problem since we started. It's called color=white; have you heard of it?

You write: "if you know what you're RD is, obviously you can manipulate your rating by playing tournaments when its high, and avoiding playing tournaments when its low."

But you already know when your RD is high or low. If it is ?? or ? then it is high (?? being especially high) and if it is ! or !! then it is low. (!! being especially low). So you do not need to know your exact RD to be able to attempt to manipulate your rating on that basis. *That* is the big picture that is already there for everyone to see.

Knowing your *exact* RD will not help you go beyond that very approximate level in manipulating it unless you know enough to calculate exactly how a given strategy will affect your RD.
99.5+% of players do not have the mathematical knowledge let alone the data to be able to calculate this.

As for the idea of manipulating ratings by playing only when your RD is high, while this will increase the amount your rating goes up or down by, you will not stand to gain points on average by doing so unless you were underrated to begin with.

If you are underrated to begin with then the best way to go up is to play a huge number of tournament games right away, whatever your RD. The benefits of waiting for your RD to rise by not playing are very likely to be more than cancelled out by the missed opportunity to increase your rating (by smaller amounts per game but over many more games) in the games that you declined to play in that time.

Of course if a player who is neither overrated or underrated takes a long spell from playing then it is possible that a freakishly good performance on return could see them make a gain they could not have made through normal efforts. However since such a rating will have a ?? after it, it will not be taken seriously anyway (especially not by selectors), so what's the point?

Furthermore such a player is most likely risking a loss in playing strength through inactivity and hence a loss of points upon return (which could be magnified should they have a really bad weekend!)

Finally players temporarily withdrawing themselves from the system in order to make their rating more sensitive to a small number of games certainly doesn't improve the accuracy of the system. All it does is increase the chance that their own rating will become unreliable, and decrease the effectiveness of it in establishing the ratings of others.

Anonymous said...

Kevin,

Please take your calm down medication. You are in fact agreeing with me that one can manipulate their rating by taking advantage of their RD when it is high and avoiding playing when its low.

Secondly, where did you get this fanciful statistic of 99.5%? More self serving speculation on your part?

Lastly, you wouldn't have raised the subject of OzChess' new features if it wasn't really bugging you inside. That strongly suggests jealousy.

Best

AO

Anonymous said...

In case it needs to be said, I would like to point out that AO's representation of my views do not actually represent my views whatsoever.

Brian Thomas

Kevin Bonham said...

Ah, the "medication" card, another incompetent cliche deployed by overexcited losers in between one froth of foam and bile and the next.

Alex's original claim was "If the RD's[sic] were made known chess players could, to an extent, manipulate their RD's[sic] and consequently their ratings."

We need to distinguish the two forms of manipulation here. A player can easily manipulate their RD, but whether they succeed in manipulating their rating in the desired direction after doing so is another thing altogether.

My point has been that to the very limited extent that a player can *try* to manipulate their rating (which is more likely than not to fail anyway), they can do it without knowing their exact RD.

Therefore Alex's "If the RD's[sic] were made known ..." is total rubbish, because whether the RDs are made known is more or less irrelevant to your ability to try to manipulate your rating, unless you are a mathematical genius.

I have been to Glickman's site and seen the maths of the system. It is way beyond most people to calculate their own rating doing it. Furthermore the system as explained on Glickman's site is complicated further by various practical Australian changes such as the use of intermediate ratings in calculation, and iterative processes used to generate internal ratings for unrated players. I could have said 99.8% or even 99.9% instead of 99.5% and I would have been quite safe.

Alex Toolsie has no idea how to do it or he would have answered my hypothetical question.

Alex Toolsie has no idea about ratings generally. Indeed he has not much idea about anything to do with chess admin, and the howlers permeating his so-called tactics puzzle threads suggest he has not much idea about chess, fullstop.

He also has no idea about psychology. Obviously the reason I expose his ridiculous obsession with trivial gadgets is not out of jealousy or because it is bugging me, but to further humiliate him and share my amusement at his defective and delusional nature with others.

Anonymous said...

Brian Thomas stated:

"In case it needs to be said, I would like to point out that AO's representation of my views do not actually represent my views whatsoever.

Brian Thomas"


The only thing I know for certain Brian is what's already on the public record. Namely, that the CAQ passed a motion requiring the ACF Ratings Officer to release the RD's. You voted in favour of that motion.

I am unaware whether the CAQ has passed a motion re-visiting the issue, and if it has not, it would be reasonable to assume the CAQ (of which you are the secretary) is of the same opinion now as it was when the motion was originally passed.

Best,

Alex Toolsie

Anonymous said...

AO-I am unaware whether the CAQ has passed a motion re-visiting the issue, and if it has not, it would be reasonable to assume the CAQ (of which you are the secretary) is of the same opinion now as it was when the motion was originally passed.
-----------------------------

Another possible interpretation is that CAQ was satisfied with the answers given as to why the RD's are not published and decided not to revisit the issue.

Anonymous said...

Anonymous said:

Another possible interpretation is that CAQ was satisfied with the answers given as to why the RD's are not published and decided not to revisit the issue.

That is incorrect.

There is nothing in the most recently published Minutes of the CAQ to suggest that the issue was re-visited, and it would have been referred to in the 'Other/General Business' section of the Minutes if it had been discussed.

Best,

AO

Anonymous said...

Kevin Bonham has been caught out!

He is, in this discussion/thread, advocating no changes to the current Glicko rating system used by the ACF. But in this quote, he is singing a very different tune!

Kevin Bonham Quote

After initially defending the (Glicko 2) system against the initial flurry of concerns (mostly raised by players who had performed very poorly and whose rating drops proved on later performance to be justified) I have looked at the changes since the system was instituted and find there is overwhelming evidence of serious deflation at the top of the list.

In August 2000 there were 49 players rated 2200 and above on the ACF rating list, of whom 42 are still listed as active. Now there are 28 players rated 2200 and above on the list


After this self contradiction, how can we take anything he has to say seriously. His credibility on his issue has just been permanently tarnished.

Best

AO

The Closet Grandmaster said...

It would be preferred if posters include a hyperlink in instances where they quote another poster. Or if a link is not available, then make mention of the exact source.

I will leave AO's post for now because I can see it in a Google SERP, but the page seems to result in a 404.

- TCG

Kevin Bonham said...

It would be even more preferable if incompetent time-wasters not attempt to create fake contradictions by quoting posts that are five years old, and giving no indication that they are of that vintage.

The quote Alex Toolsie provides is from the old ACF BB and was made on March 4 2003 in the thread "Rogers vs Glicko". There were some errors in the initial post which were corrected in follow-up posts to that thread, without much changing the overall conclusion.

At the time I believed there was strong evidence of compaction affecting the ratings of players at the very top of the list, although there was no evidence of general deflation. Shortly after this, changes were made to the handling of the ratings of the very top of the rating pool, and since then, the problem I referred to has no longer been apparent. Therefore, there is no contradiction, as the criticisms I made in that post were of an earlier version of the system, that has since been improved. Additionally, the compaction at the top end of the rating pool may have been a correction of over-stretching by the ELO system, as I acknowledged in later comments on the issue.

To quote a post that old without indicating its age is an act of intellectually dishonest trolling. I would ask, after *that*, how can we take anything *Alex Toolsie* posts seriously, except that nobody who matters was in any danger of doing so anyway.

Anonymous said...

Kevin,

Your last post was a complete flip flop. The reference to the year 2000 provided readers with a time frame to date the post. After this extrodinary flip flop, the question becomes are you genuine in your support of Glicko 2? Your own analysis of Glicko 2 demonstrated with 'overwhelming evidence' as you called it, that ratings are highly inaccurate as a result of Glicko 2.

Sorry mate, your own words have incriminated you.

Best

AO

Kevin Bonham said...

Oh dear, the troll is at it again. How he can misrepresent so shamelessly and still now and then give lip service to Christianity is beyond me (unless that is also insincere), but perhaps his use of "flip flop" steals from another in the same situation, George W "Mr 19%" Bush.

The quote comparing August 2000 and "now" gives no indication of when "now" was, and hence does not provide a time frame to date the post. It only provides one end of the period under discussion.

As I have already explained (but Alex Toolsie was far too thick to follow) this early reservation about the functioning of Glicko concerned something that has since been fixed and is no longer an issue, so using it as evidence that I currently hold concerns in that area, or that my views on the system are genuinely inconsistent, is totally invalid. Anyone who does so is an idiot, a troll, or both.

What I have done on this thread is point out that all the objections raised on it against the way Glicko-2 is currently implemented are failures. For the record, my view is that Glicko-2 is a massive improvement on any known version of the ELO system. This does not mean it is incapable of improvement, but the onus is on its detractors to demonstrate flaws that are significant and improvements that would actually work. In nearly all cases they have dismally failed to do either.

Anonymous said...

Kevin Bonham Quoted:

"How he can misrepresent so shamelessly and still now and then give lip service to Christianity is beyond me (unless that is also insincere), but perhaps his use of "flip flop" steals from another in the same situation, George W "Mr 19%" Bush."

Kevin, mate, please show me where I have given lip service to Christianity. If you can't, well, we will just deduct that from your overdrawn credibility rating.

Best

AO

Kevin Bonham said...

Plenty of support for my comment in multiple posts on these threads:

http://chesschat.org/showthread.php?p=109018&highlight=Christ%2A#post109018
http://chesschat.org/showthread.php?p=107180&highlight=Christ%2A#post107180

(These run over the edge of the post so you may need to scroll right to pick up the full URL. The threads are "life as an atheist in Australia" and "Religious sponsorship".)

So, far from my credibility rating being "overdrawn" at all, looks like another deposit has been made from your credibility account into mine.

Or it would if there were any funds in yours to begin with. :)

The Closet Grandmaster said...

See first link here and second one here

- TCG

Anonymous said...

I don't know if the linked posts can be considered giving lip service to Christianity so much as they are merely pointing out an observation I made that Australia is a pre-dominantly Christian country.

The fact I didn't complain about the pre-tournment prayer at the Gold Coast Open, again, I don't think that constitutes lip service to Christianity so much as it demonstrated that Brian Thomas was being a big cry baby for complaining about it.

So, to conclude Kevin, I think you failed your homework assignment, and if you passed it was by the barest of margins.

In any event I see you are trying to raise topics like religion to detract people from the remark of yours I quoted which is directly at odds with your current stance on Glicko 2.

Kevin, please tell us again about your 'Road to Demascus' experience leading you to stray from your previous anti-Glicko 2 position.

Best

AO

Kevin Bonham said...

I said "multiple posts on these threads" but in his typical short-attention-span fashion AO refers only to the content of the individual posts that come up first when the links are clicked. So what AO concludes is irrelevant.

His comments on the threads in question do not merely "point out" the (very dubious) claim that Australia is a Christian country but also take the Christian apologist side on a wide range of issues. AO also clearly claims to believe in "God" on the "God Related Discussions" thread on his own little forum. Of course, none of this makes AO necessarily a Christian but it does support my claim that he gives "lip service to Christianity".

My comments from 2003 were hardly anti-Glicko-2. They simply indicated that I believed there was an issue with how the system affected certain players at that time. That issue is no longer apparent following changes to the system. To portray my view then as a polar opposite to my view now is just more cardboard-cutout trolling.

My reason for raising the religion issue is not specifically to distract AO from wasting time on his clueless wafflings about ratings but to indicate how strange it is that he panders to a religion which would generally (and for once correctly) view his online conduct extremely dimly indeed.

Anonymous said...

Kevin Bonham Quote:

My reason for raising the religion issue is not specifically to distract AO from wasting time on his clueless wafflings about ratings but to indicate how strange it is that he panders to a religion which would generally (and for once correctly) view his online conduct extremely dimly indeed.

The Christian perspective wouldn't just judge my online conduct dimly, indeed, it would judge most aspects of my life dimly.

I do believe that "All have sinned and fallen short of the glory of God", myself included.

That maxim applies to everyone though, and is not just limited to myself.

But that's all I will say on this topic because I do not want to succumb to Kevin's thread hijacking attempt. I am much more interested in discussing when Kevin completely changed his opinion on Glicko 2 after uncovering the overwhelming evidence, as he called it, showing that Glicko 2 leads to inaccurate ratings.

Probably when he started sleeping with Bill.

Its almost reminiscent of Brendan Nelson's "I have never voted Liberal in my life" only to become the leader of the Liberals.

Kevin Bonham said...

Alex, you don't get an excuse in any Christian doctrine for repeatedly sinning and doing nothing about it. The point, according to such doctrines, is to realise you are doing it and try to stop it. In your case, quitting the internet entirely would be a good way to try to save yourself, since you seem so addicted to fibbing and trolling otherwise. Of course, this is none of my concern as an atheist, but your inconsistency is, as always, quite extreme.

Once again, in 2003 I made one particular criticism of the impact of Glicko on the ratings of a very small (but important) section of the ratings pool. Other comments I was making about the system at that time were generally positive. I stopped making the criticism in question when the problem went away, probably as a result of changes that were made to the system.

As for your use of "sleeping with" (without any substance in either fact or metaphor, of course) to denigrate an opponent, it looks like Matt's homophobia has finally rubbed off on you, which is a pity.

There's nothing to see here Alex, and repeating yourself when your question has already been answered will only make you look like a damaged trollbot. It's well past time for you to move on from this discredited line of attack and stop trolling.

Unknown said...

Kevin Bonham Quote:

... as an atheist,

Since have Satan worshipping Goths considered themselves atheists?

In any event mate, I have so thoroughly cleaned your clock in this thread (and exposed your hypocracy) that there is simply nothing else left to clean.

Therefore, I suggest you join Grandpa in the rocking chair I banished him to when I simply asked him for some further explanation.

Yours Sincerely,

AO

Kevin Bonham said...

Alex's latest post is just lame unsubstantiated chest-beating with no connection to the reality of this debate or any other (a familiar condition for him).

If he's really up for diversions it would be possible to have a fascinating discussion about how some variants of "Satanism" are in fact atheistic. However, not being a satanist of any kind myself (and indeed, the proportion of goths that are satanists of any kind is low) I am not sure what this has to do with anything. Perhaps such completely off-topic trivia would be best explored elsewhere (he says, with knife poised above the sacrificial (arro)goat while reciting the scriptures backwards :) ).