Grade Categories

Home Page	Yorkshire Chess Association	Last update: 27/08/00
Grading Matters Contents	Grade Categories

Grade categories represent an attempt to combine the relevance of two variables into one parameter. They are in part determined by the total number of games used to calculate the grade, which is reflected in the chart below by the diagonal banding effect, and they are in part determined by how many of those games were played in the latest season, which is reflected by the horizontal banding effect. The resultant parameter, the grade category, is intended to be an indication of how accurate, by and large, the grade will be as a reflection of average playing performance of the player over the period concerned, which period may be one, two, or even three seasons.

In the chart below, the figures down the left represent the number of graded games from the latest season, while the figures across the top represent the number of graded games from the previous two years together. The letters A to E in the individual cells are the resultant grading category.

The letter U represents cases where no grade would be published, and the U can be thought of as representing "unpublishable", "unreliable", "useless" etc. The normal grading calculations can still be used to produce a "grade", in which case it is often referred to as an "estimated" grade, though that term is a little misleading as it suggests someone's opinion rather than the result of an arithmetical calculation.

Note that grade category E is not "E" for "estimated", though the implication is similar, i.e. "best regarded as probably only approximate". However, a player who normally plays a good number of games in a season, but happened not to play much in the latest completed season, may well end up with an E grade which is reasonably accurate. Reference could usefully be made to the previous season's grade and grade category.

D grades may be expected to be a bit more reliable, but there is still much room for a D grade to be misleading. Playing a 5-round congress each year, and no more, could yield a D grade. Unless it is on the cusp, a D grade is a reasonable basis for determining the section to which to allocate an entrant to a congress.

A to C grades are all based on 30 or more games and can be regarded as statistically quite sound from that point of view.

C grades draw their data from three seasons, while D grades from the latest two seasons. For this reason they are long term averages rather than a measure of "form" in the latest season. For established players whose strength varies little from season to season, a C grade is as good as an A grade. Such players may notice, however, that a "good" or "bad" season has limited effect on their grade, due to the averaging-out effect of including results from two or three seasons. For such players this still gives a reasonable guide to strength in the following season is that is more likely to be average than particularly good or bad. The averaging-out effect is more misleading for relatively new players for whom a B or C grade may be less accurate than an A grade if the player is improving fairly steadily. For such players a B or C grade with be an understatement of their strength in the new season. For juniors this problem is recognised and an attempt is made to remedy this by applying to the calculated grade an increment determine by the age.

As regards the applicability of grades generally, it should be remembered that they are a measure of a player's past performance over a year or more. In practice grades are used to estimate a player's probable current playing strength for the purpose of a team's board order, or for allocating players to graded sections of congresses. Current playing strength may well be different from the latest published grade. Grade calculation is not prophetic.

Top of Page

Grading Matters Contents

Home Page