Topics

Knock-Out Competition - Grades

February 2017

Preamble

There has been some discussion about how players’ grades are derived for the purposes on the calculating the handicap. Referring to rule 7 the current system uses the player’s current ECF grade where available, else his/her rapidplay grade. Where neither is available a grade is allocated depending on the highest division in which he/she has played:

Division 1  150

Division 2  135

Division 3  120

Division 4 (or not previously played)  105

The following comments have been edited to remove references to actual players:


Barry Maufe

Lewes

The current rule allows does not always reflect the strength of players with no ECF grade. In particular, it does not use the FIDE grade when available. There is a reasonably accurate conversion formula to calculate an equivalent ECF grade from it. There was an occasion this season where a player competed at about 40 points below his equivalent ECF grade.

I suggest that the rule is changed to allow the use of FIDE grades when no ECF grade is available.

There is a “Grader’s database” (not published) available to the Sussex Grader, which may contain an estimated grade for the player in question, but it could be wildly inaccurate.

If we assume that our league is based in honourable conduct, then I think it's fine to obtain the estimated grade from the player/club, where no ECF or FIDE grade is available.

Barry


Julie Denning

...Whilst we could propose an amended rule to use the suggested conversion factor for FIDE (or national ELO) ratings, there will always be the chance of anomalies. (Incidentally, the conversion given on the ECF grading website is an estimate based on a best fit line through a cross-plot of ECF grades and FIDE ratings for a number of players. The systems are quite different and there's no absolute mathematical link between them.) I would recommend that if we change the Rule we do so by requiring a club to seek an estimated grade from us (and probably ... the grader). Where there is a current FIDE or national ELO rating it would certainly make sense to use that. However, judgement will be required where someone has played in the past but doesn't have a current grade or rating...

Some years ago I did publish on the website estimated grades obtained from my Grader's database. These were based on whatever games had been played, but which were too few to qualify for a formal grade. I made it absolutely clear on the website what these values were, yet I was quite ferociously taken to task by one person accusing me of putting out misleading, or indeed dishonest, information! You can never please everyone.

Julie


Barry Maufe

Since the conversion formula was changed again around 2008, I’d assume ’s fairly accurate today

Surely the only criterion should be to make the competition fair, and can't we assume honourable conduct by the clubs? You could say you have to vet the submitted estimate if you think it's necessary.

Barry


Doug Stevenson

I would have thought that new members with only a FIDE grade will be a fairly rare occurrence.

However, I suggest rule 7 could be amended to "it should be estimated from any reasonable sources but shall not be less than ....... determined by the team nomination derived from main rule 4.3"

Doug


Barry Maufe

New members with only a FIDE grade are the ones that seem to cause the problem though.

I don’t see a reason to specify a minimum because for the case in I gave the grades specified in the rules are ridiculously low and therefore meaningless. However, on a different point entirely, there is a case for specifying a hard minimum for board 5, overriding their actual grade. This would avoid the case where a team could field a player rated at, say, zero, expecting him/her to lose, but gaining a grading average advantage overall.

Barry


Julie Denning

I agree that the conversion of a current FIDE rating to an ECF grade should give a result in the right ballpark, but with quite a bit of scatter. In the BCF grading list for 2002 - 3, an amended formula for conversion of ELO ratings below 2326 was given. This was based on a survey of players with both a BCF grade and an ELO / FIDE rating and the data was shown graphically ... (The conversion formula was subsequently changed again around 2008. This appears to have been related to the adjustments made to ECF grades to account for a perceived systemic deflation in these grades, but I've not seen any published data since that in the 2002 - 3 list.)

We need to be clear on what we're trying to achieve. I'm not too worried where the all-play-all divisions are concerned. Rule 5.2 is quite explicit that team order is to be based on "current playing strength" with "considerable discretion" permitted to team captains relative to actual grade. There is provision for this to be questioned by the Match Recorder. I don't recall an instance of this provision being used. There has been the occasional past instance of a new player being inappropriately placed initially, but rapidly corrected once their real strength was appreciated...

In the past, no account was taken of an estimated grade in showing an average grade for the team. The average was simply based on those players with an official grade. I stated previously my experience of trying to give an indication of estimated grade and being quite viciously attacked for my efforts! You've chosen to try to give an estimated grade, and that's fine by me, but it doesn't have any formal status where the league matches are concerned.

The KO is a different matter as the assumed grade of a player can affect the outcome of the match. For better or worse the current KO Rules have a simple way of treating ungraded players which has been accepted for many years, despite the occasional anomaly...

An argument can easily be made, as you are doing, that the current KO Rule 7 isn't fit for purpose in some cases. Playing devil's advocate, I could argue that any rule is likely to be vulnerable to anomalies, so is it really worth meddling with the simplicity of the existing rule that has generally been deemed adequate for many years? If we do introduce an amendment that allows at least a degree of discretion, then I'm clear in my mind that this discretion should be exercised by an Officer of the MSCL and not based solely on the view of the club concerned. This is because of the obvious significance to the outcome of the competition. This would be comparable to the County Championships where a player without a current ECF grade must receive clearance from the tournament controller to play in a grade-limited competition. I would see the person exercising this discretion looking at all the information that might be available, including foreign ratings, ECF grades from the past, recent ECF-graded games that were insufficient in number to qualify for a formal grade and any guidance the host club can give.

A situation that I believe is at least as common as a newcomer from overseas, is someone returning to competitive chess in England after an absence of some years. It is possible that an historic grade for the player is available, but I think all would agree that some degree of loss of strength, at least temporarily, is to be expected. How should this be quantified? I'm not aware of any published data on this aspect. I can certainly think of a number of instances of this over the last few seasons...

In summary, I'm quite relaxed about consideration being given to the adequacy of KO Rule 7, but advocate the following points:

Julie


5th June 17

Brian Stockham

Uckfield

Please forgive me for not entering the debate sooner, but it must be in everyone's knowledge and certainly has been to all other captains I have spoken to this season that we in Uckfield are responsible for this debate.

I can now add to that proudly that we are this season's knockout champions, but we can state confidently that we would have been champions whether our newly acquired top board had been recorded as his ECF equivalent of 192, or as I recorded him, as 150 as we managed in some cases more than the required score, and in the final it would have made no difference.

This was a dilemma which at the start of the season did not sit entirely well with me as the captain, so I consulted with Julie Denning. We were pleased to welcome our new player to the club, as he had also visited several others, and were mighty glad he chose us as we had lost our former top board to University. However as regards the rules we had no way of judging his actual grade as he was clearly significantly better than anyone we had in the club. His FIDE rating was known as he had been living in Sweden for many years, but he had not played competitively for a while. So, was his equivalent ECF rating of 192 correct, and how should we rate him for the knockout ? Julie's advice was as she has added to this debate, but in fact as our first team was only competing in Division 2 we could have recorded him as a 135, but I felt that was wrong, so conveyed to Barry for resulting purposes that in the knockout we would record him as a 150, as he was clearly a player capable of playing in Division 1.

For a while, I was of a mind that this situation could be prevented in the future simply by changing the rule 7 by adding "(or FIDE equivalent)" to ECF grade.

However as the season went on although the player has had successes commensurate with the FIDE equivalent, he has also had losses to a 157 and 186 and drawn with a 179. Despite some fantastic chess against higher rated players and some of the best in our League, how does this help us as a club with no-one anywhere near as good to estimate his actual playing strength ? The only security is in the eventual grade arrived at by the ECF.

I do not claim to fully understand the intricacies of FIDE equivalence to ECF grades, and bow to Julie's well articulated concerns on that.

I have been helped considerably by fellow captains and players around the League, some of whom have offered that this anomaly has surfaced for many years in the Knock-Out competition, with overseas champions having played for many clubs with some effect. That raises a very considerable issue as that past acceptance of such an anomaly has therefore gained ground as precedent in this League. In legal terms precedent is always an abiding concept.

It was interesting that in the Final our opponents pitted against me personally (pure luck I am sure !) a player who relied on this rule to be recorded as 120, though in the League she has a FIDE equivalent grade of 125. She has since returning to the UK in the last year beaten all she has played in Divs 3&4, including a 148. Having been well beaten, I would certainly question her FIDE equivalent rating gained whilst in Greece, as on the ECF system she was as high as 177 in 2011 and falling to 161 in 2013 and has apparently since been a regional champion in her home country.

Mindful of all of the above, I would now after considerable reflection suggest there should be no change to the rule. We should not want the spectre of what is happening in athletics, with past championships needing to be overturned as a result of a rule change.

Brian Stockham