That system can certainly be faulted if you like. And you do like: <<I would urge NAQT to consider strength of region and head-to-head victories for future at-large invitations. Statistics as one of the few comparative methods across regions, but stats should NOT be used to accept lesser-placed teams in the same region.>> Stength of region is absolutely _crucial_ to the comparisons that we do make. Head-to-head victories are meaningless for us, a decision we have made that you may dislike, but I honestly see little possibility of a workable system which would make that a possibility for us. Nor do I agree with that as even a philosophical goal -- I would not want to see a head to head victory on a particular packet, possibly by 5 points, or a similarly small margin, outweigh in consideration other overall wins and losses, or large differences in scoring average across all the packets of an entire tournament. Yes, within the context of a single tournament, it is the wins, however narrow, that matter for the standings, not who is statistically dominant. But for ranking teams for another purpose, that is not necessarily the means to the best result. That last statement, though, that "stats should NOT be used to accept lesser-placed teams in the same region" is very much a possible conclusion for us. Certainly it goes against the grain to have invitations not follow the order of final finish within a tournament. But rankings that are based on point-scoring statistics are the only means we have of fairly comparing teams across different tournaments, where won-loss record comparisons become meaningless due to differing formats and field strengths. If our approach was to be that we would compare teams across tournaments by statistical methods, but refuse to issue invitations out of order of final finish within individual tournaments, how would we handle the teams whose statistical performances are markedly better or worse than their numerical placements would have suggested? [continued again - man I hate these message size limitations when you have a lot to say.]
This archive was generated by hypermail 2.4.0: Sat 12 Feb 2022 12:30:43 AM EST EST