Chess.com Fair Play Report: Are There Too Many Upsets in Titled Tuesday?

This is the first of several planned reports from Chess.com’s Fair Play team on cheating. In this report we try to answer the question: are there too many upsets in Titled Tuesday... possibly suggesting widespread cheating? To answer this we use publicly available data from Chess.com’s own Titled Tuesday events and from over-the-board (OTB) blitz events.

Our analysis compares the rate of upsets in over-the-board blitz games to the rate of upsets in Titled Tuesday games, measured by how well underdogs score against higher rated opponents. We use FIDE blitz ratings to determine underdogs because they are completely independent of Chess.com ratings and best reflect the 3+1 time control we explore.

Our research indicates that upset rates in Titled Tuesday games and OTB blitz games are similar, typically within 5% of each other. What stands out in the research is that these similarities hold across all rating ranges. For example, when 2200s play against 2400s we see an upset rate of 33% in OTB blitz and 30% in Titled Tuesday. Likewise, when 2500s play against 2700s we see an upset rate of 32% in OTB blitz and 30% in Titled Tuesday.

While upset rates are broadly similar in OTB blitz and online play, we do identify some small discrepancies. These may be the result of skill differences in OTB blitz and online play, inaccurate blitz ratings, normal variance, and cheating. While we can’t yet say what portion is due to cheating, we know it exists, and in the next report, we will attack this question in greater detail.