| |
BGonline.org Forums
3rd BMAB-USA Tournament PR Results - Additional Comments
Posted By: Rick Janowski In Response To: 3rd BMAB-USA Tournament PR Results (Timothy Chow)
Date: Thursday, 21 July 2016, at 12:13 p.m.
At BMAB we have been aware about occasional bugs in BG analysis since we started operations in December 2014. For this reason, two independent sets of analysis are always carried out for each tournament and individual application batch. It was also apparent that small differences are a common occurrence. Xavier is aware of this bug and will resolve in XG3 it us to be hoped.
To moderate the results for a typical batch (50 experience points or more typically for all players) the following methodology has evolved:
1. For any player, accept the lowest batch PR result, when the two results are within 1% of each other. The overall calculation of aggregate PRs makes use of XG Profile files so taking an average value will not work. Using the lower value, provides a consistent approach for all players, without any significant loss of accuracy and avoids unnecessary discussions and complaints. For batches less than 50 experience points (eg, 1 day BMAB tournaments), a difference of 2% may be accepted.
2. Where the two batch results for a player are different by more than 2%, a third "moderating" analysis is carried out to resolve the issue. The median PR (and its corresponding profile file) is accepted provided that this is within 1% of either the maximum or minimum PR values.
| |
BGonline.org Forums is maintained by Stick with WebBBS 5.12.