|
|||||||||
|
The Black Hole in Chess Ratings – Part II By Russ Mollot (founder, Chess Express Ratings)
In recent years, these distortions have been slowly percolating through the entire national rating pool. Like a celestial “black hole”, invisible but dragging massive objects (e.g. stars) down into its gravity well, there is a black hole in chess, warping the rating system, causing all established ratings to be pulled steadily lower. This “Black Hole Effect” (BHE) is caused primarily by the haphazard manner in which young players are being assigned initial "club ratings" by many scholastic chess programs.
So why doesn’t the world of scholastic chess employ this method? There are several reasons that scholastic chess programs choose to assign initial (provisional) ratings to their students. But let’s face it – even if they wished to follow the proper procedure, the logistics of tournament play (travel arrangements, costly entry fees, USCF dues, and perhaps even local federation dues; the long, grueling hours of tough competition) are difficult even for an adult player, let alone for a large group of children. It is far more convenient to say to the young chess students: “Let’s give everyone an equal rating to start with: 600 since you’re all sixth-graders.” There are all sorts of ways such ratings are assigned, but in general, kids are assigned ratings that are way too low. They then acquire varying degrees of skill depending upon the number of classroom hours in the program, the skills of the instructor, the relevance of the curriculum, and the ability and desire of the kids themselves. Those who graduate from the world of scholastic chess and move into competitive tournament chess tend to be the strongest scholastic players. They are typically under-rated by 400 to 600 rating points. The ensuing interaction between them and veteran adult players results in a net draw-down of the “established ratings.” The scope of this rating deflation is very large and growing geometrically.
There is good reason for this; the author
of this rule understood the lack of statistical validity in the scholastic
ratings, and recognized the negative impact that could result from mixing
the two. The trouble is, most tournament directors (based on our survey of
50 TDs) and the overwhelming majority of chess clubs allow youngsters to
play rated games against adults. So the dreaded mixing of "good" ratings
and "bad" ratings happens with regularity, resulting in the massive BHE
deflation mentioned above. The simple answer which presents itself is enforcement of the USCF's ban on kids playing in adult tournaments. This keeps the adult ratings "pure", and when a scholastic player is ready to join the adult world, he simply throws his or her scholastic rating into the dung-heap, purchases a regular (adult) USCF membership, and starts all over again as an unrated player. Now, whether you like this idea or not, it has one great thing going for it: it works (at least in theory). But enforcing such a distasteful "apartheid" policy would be as controversial and unpopular as it would be difficult. The separation of scholastic chess could easily become a permanent rift; not just in the ratings and the events, but in the hearts and minds of the players. Is the USCF prepared to offend (perhaps even jettison) the scholastic chess community which has become its fastest growing source of revenue? That would be bad business, and almost as certain, bad for chess. If we are to dismiss the "apartheid
solution" (and I am not necessarily saying that it should be dismissed),
then the BHE deflation problem remains. It is in everyone’s interest to
solve this problem. We really need to face up to it, and stop hiding our
heads in the sand. Can we bring the two worlds (adult and scholastic)
"in sync"? Even if we cannot ensure that there is some standard, objective
meaning to a rating of 1200 (for example), can we still, at least, reduce
the disparity between scholastic and adult ratings? No one would object to
such an achievement. With that in mind, Chess Express Ratings, Inc., an independent chess rating authority, has been among those voices calling for:
The first item on this list is simple to execute, as there already are some suggested curricula which have been published, and one or more of them could be officially adopted by a national board of scholastic chess organizations (“NaBoScO?”). It is in everyone’s interest to establish standards not only for the assignment of initial ratings, but also for the basic curricula being taught among the varied scholastic programs. Every other subject taught in the schools has well-established curricula and standardized tests. Teachers have to demonstrate a certain mastery of their subjects, as well as the ability to teach. I don't propose that only professors of Chessology be allowed to run scholastic chess clubs. But there is a solution out there, and it lies somewhere between the current rampant anarchy and strict standards. It is obviously better to have an instructor knowledgeable in chess than one that is not. It is better to have a suggested curriculum than none at all. We also wish to preserve the right of each scholastic chess program to use its own individual style of teaching that distinguishes it from others (as long as it is doing a decent job for the kids). There is certainly much subject matter which can be standardized without jeopardizing each program’s right to individuality and independence. The second item will certainly be the most difficult and the most controversial. Yet, if chess is being taught in schools, why should there not be standardized tests as with any other school subject? The controversial part is the idea of making the scores on these tests an input factor affecting player ratings! The monumental task of developing such tests has, to a large extent, already been performed by reputable chess educators. Choosing the appropriate tests, and devising a method for using test results to adjust scholastic ratings, should be done under the authority of the national board of scholastic chess organizations (referred to above). The third and last item -- special
treatment of scholastic ratings -- will stir up the usual controversy and
hostility from many players, but will also be welcomed by many others. It
essentially involves a method to ensure that scholastic ratings take into
account the very rapid improvement in play that occurs as students absorb
large doses of learning and training. The solution may not be all that
difficult to implement. In fact, Chess Express Ratings has already
implemented what we believe to be an excellent solution. This solution will
be discussed in Part III of this series.
|
Advertisement
The
The
Advertise to Single insert:
|
||||||||
|
|||||||||
In case you
missed 
Although the non-statistician may find
it logical to assign low ratings to children who know nothing about the
game, this was not the intent of Professors Arpad Elo and Frank Brady,
pioneers in modern rating theory. One of the axioms of the Elo system
is that a rating of 1500 would represent an “average” player. Also, a new
player was not meant to be given an initial rating; he or she needed
tournament play in order to establish a provisional rating. In other words,
new player Joe Blow was considered unrated until the results of
over-the-board play (against players with established ratings) were
available to calculate his initial provisional rating according to the
special formula for that purpose.
It is worth noting that the USCF
officially sanctions the use of scholastic "club ratings", including the
somewhat arbitrary assignment of low initial ratings by school clubs. The
USCF offers "scholastic memberships" at a discount from the regular
("adult") membership. However, if you read the fine print, scholastic
players are INELIGIBLE TO PLAY in rated adult tournaments! They are
restricted to scholastic tournaments.
