Thanks for reaching out. As far as I know, there are currently no plans to release the test data for the 2015 Challenge. We keep this data private for the reasons you probably read about in the previous threads. Keeping the test data secret discourages cheating (especially accidental cheating) and learning from the test data. However, we may be able to run your code for a previous year’s Challenge, including one shot at the test data. This policy is meant to ensure that the method is truly repeatable and the model is truly generalizable.