"Hi everyone, I wanted to open a quick discussion to better understand the leaderboard scoring mechanism.The competition overview states the metric is a weighted average of MAE and RMSE. Typically, this means 'Lower is Better' (closer to 0 is perfect). However, looking at the top leaderboard scores (around ~0.96) versus the scores I'm seeing in my local validation (which are significantly lower/different), I'm finding the scale ambiguous.Could anyone (or the Zindi team) clarify: Is the leaderboard score strictly (MAE * 0.5) + (RMSE * 0.5)? Is there any scaling or normalization applied to the target columns before scoring?Thanks for the help!"