https://lichess.org/@/D2D4C2C4/blogD2D4C2C4's Blog2024-10-18T08:42:24.548Zhttps://lichess.org/@/D2D4C2C4/blog/why-opening-statistics-are-hard/9g61F9Uc2024-10-18T08:42:24.548Z2024-10-18T08:42:24.548ZWhy Opening Statistics Are Hard<img class="ublog-post-image" width="880" height="550" src="https://image.lichess1.org/display?fmt=png&h=550&op=thumbnail&path=d2d4c2c4:ublog:9g61F9Uc:NlRhGdr0.png&w=880&sig=91f04d99b4a588029f42b2ac5b2585485c345c0d" /><br />TL;DR There are pros and cons to adjusting opening statistics by players’ ratings. Under certain conditions, the two types of estimates are virtually identical. Statistics on the very first few moves should be completely disregarded regardless of the method used.D2D4C2C4https://lichess.org/@/D2D4C2C4/blog/why-opening-statistics-are-wrong/VKNZ1oKw2024-09-24T15:52:47.969Z2024-09-24T15:52:47.969ZWhy Opening Statistics Are Wrong<img class="ublog-post-image" width="880" height="550" alt="First-move statistics" src="https://image.lichess1.org/display?fmt=png&h=550&op=thumbnail&path=d2d4c2c4:ublog:VKNZ1oKw:Gfe01g0T.png&w=880&sig=c10e5fe6d8013b3a50f9359cb683d13a254d7e7d" /><br />Opening statistics, such as those in the opening explorer have a methodological flaw. They tend to significantly overstate differences in win rates among various possible opening moves. I propose an alternative methodology and present some data to compare the two calculation methods.D2D4C2C4