P Grades question

[Brian Valentine]

Only if that rating was above the dummy rating.

[David Blower]

A further question I have got is does the dummy draw, eventually disappear from a player's rating, once they have played enough games?

[Roger de Coverly]

A further question I have got is does the dummy draw, eventually disappear from a player's rating, once they have played enough games?

The same question also applies to FIDE ratings which now has two dummy games.

One answer at least is that the earliest rated games eventually drop out of their own accord by the application of the k factor.

[Nigel White]

Perhaps he or you could ask the questions on this forum

We had this interminably before and during the MCF AGM last month

They believe that adding dummy game(s) against a notional 1800 player, rather than a notional 1400 player, is wrong; they'd be much better taking it up directly with the ECF grading team

To prove that there is nothing new, it appears that we had a similar debate at the AGM this week of the league for which I am the ratings officer. It was portrayed as a conspiracy to do down the new players of just one club! Fortunately, despite having no notice that this was coming up, I consider myself to be fairly familiar with how the rating system works and could explain the working of the dummy rating and that it doesn't mean new players start with a rating of 1800 - not that this appeared to assuage the particular individuals concerned. The debate was all a bit irrelevant, as the league does not use P ratings for eligibility purposes. Nonetheless, one club wanted the League to write to ECF to get it to change the rating system, which was overwhelmingly defeated when put to a vote.

Although, I understand the how of the working out of ECF ratings, I am not sure I fully get the why for all aspects. In particular, is there anywhere an explanation of the Elo performance table used to determine the dp factor in the calculation of a P rating. It is clearly non-linear, but out of curiosity, what is the basis behind it?

[Roger de Coverly]

Although, I understand the how of the working out of ECF ratings, I am not sure I fully get the why for all aspects. In particular, is there anywhere an explanation of the Elo performance table used to determine the dp factor in the calculation of a P rating. It is clearly non-linear, but out of curiosity, what is the basis behind it?

You may need to go back to Elo's original book where he sets out the theory.

The ECF copied the tables it uses from FIDE and doubtless those originated with Elo or possibly the USCF. For no doubt practical reasons the tables are discrete spot values rather than a continuous distribution.

I believe the method for a new player scoring x% against a field averaging Y is to establish what rating a player would need to achieve x% against that field. The draw against a dummy player trick ensures that you don't get the problems that can arise with this method when a player scores 0% or 100%.

FIDE now also use the dummy player idea, but deem there to be two games rather than one.

[Kevin Thurlow]

"To prove that there is nothing new, it appears that we had a similar debate at the AGM this week of the league for which I am the ratings officer. It was portrayed as a conspiracy to do down the new players of just one club!"

That has happened in the past...

[NickFaulks]

FIDE now also use the dummy player idea, but deem there to be two games rather than one.

After many years of hard work on the FIDE rating system, the 2024 changes were the point at which I finally lost all hope and gave up.