An analysis of upset matches in Overwatch League 2019

Some of the most interesting and exciting matches in Overwatch League happen when an underdog team beats the odds and defeats an opponent that was expected to win. How often these results occur, do they happen more often in the regular season or in playoffs, and are some teams better at causing upsets than others?

if you would like to know more about the methodology behind the analysis, click here

Results

Of the 239 matches played between stage 2 and the Grand Finals, 82 were upset matches, around 34% of matches. Of these 25 were classed as major upsets, i.e. where the gap between the teams was seven places or higher, with 58 minor upsets. Overall approximately 10% of matches were major upsets.

Major and minor upsets per stage

The results differ slightly in different stages. As you can see the likelihood of upset results (both major and minor) increased throughout the regular season, with a corresponding slight reduction in the number of expected results.

The increase in upset matches in stages 3 and 4 can be partly explained by the upsets to the meta during those stages. Stage 1 and 2 had a fairly stable triple-tank triple support (3:3 or GOATs) meta, which meant that teams that did well in Stage 1 were likely to repeat the performance in Stage 2. Perhaps the best example of this was the two stage finals both being between the same teams, Vancouver Titans and San Francisco Shock.

Whereas Stage 3 was a very transitional stage. In anticipation of the introduction of role lock teams introduced more DPS heroes, leading to a loose 3:2:1 Sombra meta developing. Then Stage 4, played with 2:2:2 role lock saw the almost complete dominance of the Orisa Sigma double shield meta. These changes gave different teams the opportunity to shine. In partiuclar Shanghai Dragons winning Stage 3, and previously middle of the pack Atlanta Reign achieving a perfect 7–0 Stage 4.

Upsets are much more likely when the meta is in flux like this. The definition of an upset is based on the assumption that the games are played under the same conditions. When those conditions change it will benefit some and disadvantage others in ways we could not predict.

Regular stage vs stage playoffs

There was almost no difference in the overall likelihood of upset matches in stage playoffs when compared to regular season matches, across both major and minor upsets. This is reflected in the experience of watching the playoffs, where most matches were as competitive as the regular season.

However there was an appreciable difference when considering the upset factor (i.e. the seed difference between upset winner and loser)

The average upset factor per match (including non upset matches) is 30% higher for the regular season matches compared to the stage playoffs.

There is a simple explanation for this. Higher seeded teams are more likely to make it into the playoffs. This means the range of seeds in playoff matches is smaller than for regular matches. Since there are fewer lower seed teams the likelihood of a significantly lower seeded team beating a higher team is reduced.

Keep an eye out for part 2 where we break down the data by teams to show which teams were the best at pulling off surprise victories, and unexpected losses…