PRATCOOKERPratCooker is designed to be very very easy to use for a tournament organiser. It works very well on a USB stick. Pratcooker uses a Windows dialogue to find the file that you want to process, and assumes that the prior data comes from RATING.DAT which should have been loasded onto the USB stick (or the working directory) beforehand. It calculates the new ratings, copies the TOU file and the TXT file to the USB stick and displays the TXT file. The TXT file shows a warning that the processing is interim, pending processing by the NRO. The display shows the tournament placings, with wins and spread, together with the rating change. This text file is quite well suited to displaying in an email message to the OzScrabble mailing list.
The reliance on RATING.DAT means that it is not possible to process two sequential tournaments.
PLINRATPLINRAT is really only intended to be used by the National Ratings Officer. It does not have any Windows dialogues and just works at the DOS command line, in a working directory which contains recent files. The tournament file is chosen from a list of recent files (but will miss files created with ancient versions of AUPAIR). It allows the prior data to come from RATING.DAT or any recent RT2 files (this allows several files to be done one after another). It writes a traditional STA file, an updated rating file .RT2, a .WHY file for checking and debugging, a CSV file for checking and debugging and the TXT file.
Aggregation can work without extra storage because the equation which equates expected wins to actual wins can be rearranged to give:
n_games x rating = sum of opponent ratings + 1200 x all wins - 600 x n_games
The right hand side is easily updated with a new tournament.
The old value of the left hand side is already available from data in RATING.DAT as we always have had it.
An alternative methd to get the starting value would be to store extra data at the end of the line
At the moment the program can read such extra data but doesn't write it.
Change log15.04.2010 Better treatment of the 500 floor (now moved back to be inside the recursive loop). Earlier tests seemed to be unsatisfatory, but new tests seem okay. Verified by Martin.
WHY file now shows details of provisional caclulation eg
| 1.0 -493 | TAS&Laura Dimento | 7 | 1.0 | 685 556 Prov. (g=7 -> 14) Prov rating 556.3 = ( 7 * 685 + 977+ 786+ 703+ 977+ 786+ 703+1061 + 1200 * 1 - 600 * 7 ) / 14
16.04.2010 Error in calculating ratings of established players now fixed (had to ensure that new rating of opponents was
NOT used if the opponent was established).
26.04.2010 Two errors where novice players have a bye. (1) Error created in an eval statement while displaying the calculation in the WHY file. (2) Calculations in some places used last round played instead of actual games played. Bugs exposed nicely in the 2007 Janboree tournament (good bug catching by Martin). Fixed now.
06.10.2010 A bug in Pratcooker was fixed (opponents of provisional players not being calculated correctly due to a mismatch of a variable name). Pratcooker now displays the version number and a warning that the results are pending NRO processing. The bug was not present in PLINRAT, but for consistency both are now given as version 10.10