PROPOSAL FOR THE GUIDANCE OF
TOURNEY ORGANISERS, COMPOSERS AND JUDGES

0. DEFINITIONS

conversion - the consequence of a move in an odb solution (ie series of optimal moves by both sides) where the force present changes, ie by capture or promotion.
database - see odb
depth - the number of consecutive optimal moves needed by the winning side to checkmate or to conversion.
metric - the component of the algorithm used to generate on odb whereby the program can refer to all target winning positions. There are in general only two metrics: checkmate; and (winning) conversion. The first task of the algorithm is to obtain or generate all winning positions according to the chosen metric.
odb - 'oracle' database, otherwise known as total information database or tablebase. An odb is generated by algorithm implemented by computer program for specific chess force. When complete an odb can yield on request the true result of any position for the force in question, the optimal move or moves (if any) where there is a win, and similarly the depth. Odb's for five chessmen (the kings included, but with no more than a single pawn) have been in the public domain since the 1980's. Pawnless odb's with six chessmen have been generated and many results published, but with rare exceptions six-man odb's are not yet (in 1999) publicly available. It is commonly assumed that odb's for one more chessman will be produced every ten years. Although not in general use the term 'oracle database' is useful to distinguish it from other types of database.
optimal - an optimal move in a position that is a win is a move that will win in fewer moves than any other move (if any), or a move of the defence that will delay loss longest. A similar definition can be applied to a position that is drawn. Technically, an optimal move by the winning side reduces the depth by 1, and an optimal move by the defender maintains the depth unaltered. See also metric.
table-base - a compressed odb.

1. In recent years judges and editors have been faced with the insurmountable difficulty of distinguishing between a 'malyutka' (five men only) study composed traditionally and one that may have been extracted from an odb. The studies subcommittee has considered the situation and reports as follows.

2. The skills needed to extract 'studies' (ie optimal series of unique moves to win or to draw) from an odb are distinct from the talent, creativity, technique and persistence needed for traditional study composing.

3. The subcommittee proposes two principles:

3.1 Traditional and odb studies should not compete in the same tourney.
3.2 However, the use of computers should be encouraged, because they can both assist in ensuring soundness and be a source of ideas.

4. The subcommittee therefore makes the following recommendations.

4.1 The use of odb's to verify the correctness of variations is acceptable.
4.2 The use of odb output in a main line is acceptable, provided only that the initial position has at least one more chessman present than is the maximum supported by odb's on the market.
4.3 Obscure lines of play should be accompanied by prose text explaining what is happening. This explanation must satisfy the judge. The source of the text may be collective.
4.4 Separate tourneys for odb 'studies' should be organised.
4.5 Judges of study tourneys should familiarise themselves with odb technology and with current publicly available odb's.

Signed:

John Roycroft (speaker), London
Hillel Aloni, Netanya
Ofer Comay, Israel
Gady Costeff, Israel and USA
David Gurgenidze, Tbilisi
Nikolai Kralin, Moscow

Netanya, October 1999