In Contest 27 Sky Blue Hawk moved from rank 11 to rank 1 in 2 days (from 8/25/2017 to 8/29/2017)
Reviewing Sky Blue Hawk performance metrics in Leaderboard csv files for the above dates one metric stands out as improbable - annVol_pt decreased from 0.059188114 to 0.042750392.
According to formula to calculate Annualized Volatility theoretically max possible volatility change in 2 days could only be -0.8% vs Quantopian calculated -27.8% or a difference of ~34 times. This bug also triggered excessive blow up of Sharpe and Sortino Ratios, and as a result, drastically adjusted leaderboard ranking - ie my own algo moved from 2nd to 4th place.
I believe this was not done intentionally!
Please fix this bug ASAP as we only have 1 day left in Contest 27