Biased Data in the LCS

In the past few days it was announced by Riot Games that their esports leagues in Europe and North America would change the Best-of-1 format for the summer half of the 2016 regular season to a Best-of-2, for EU, and a Best-of-3, for NA, format. After the season ends, Riot Games will review the data from both the NA & EU LCS and decide which format is best for the regular season to continue using.

First, we need to backup and understand a few things before continuing. When it comes to data analysis there is often confusion for those that are inexperienced, or lacking in knowledge, of the correct steps that need to be made throughout the duration of the process so as to have accurate results. My experience as a student, researcher, and analyst is that many people are lacking in an understanding of the initial and final steps of the analysis process. I assume that the majority of the esports community are not individuals that have much knowledge of regressions or other tests used in statistical analysis as well. The necessity of this point is that simply because a person learns and understands how to use statistical software doesn’t necessarily mean the results will be without bias, that they’re using the correct type of regression, and also that they properly understand the results from said regression.

This is where the issues arise specifically with Riot Games and their approach to deciding which format to use. To eliminate all biased data, the parameters need to be exactly the same. The parameters around the EU & NA LCS are not the same. The first issue is that the days that the EU LCS will be playing during the 2016 summer split are on Thursday and Friday. The days that the NA LCS will be playing are on Friday, Saturday, and Sunday. The excuse that Riot has used is that culturally, or Europe as a region, does better on these dates than on the weekend. This is weakly plausible, but for fairness I’ll explain the thought process of why it might be better. It is because of competition with contemporary sports. Football (in NA, soccer) might take priority over esports for many Europeans. Riot is interested in stable growth for its viewer-base, like any company should. The UEFA Champions League plays its matches on Tuesday and Wednesday. Individual leagues (Bundesliga, La Liga, etc) play the majority of their matches on Saturday and Sunday with a rare game, or two, occurring on Friday and Monday. To eliminate the possibility of people having to choose between watching the EU LCS and a football match, they play on Thursday and Friday. The added concern may be that some Europeans go out for the day if the match is relevant and thus will miss the entire day of LCS. This concern is far reaching but now can no longer be used as a vague excuse.

The second issue for the EU LCS this upcoming summer split is the time of day. Coverage in the past began at 6pm and typically ran until 11pm-midnight because of games starting about 45 minutes later than the time listed on the program. Within the Bo2 format, if the start time remains the same, the end time would be 10-11pm which makes it more feasible for people to watch until the end. The issue that has primarily been aroused is the weeknight broadcast. It is more difficult for individuals to forgo so much time on a night where the next day most people have work or school. The next issue is that on Friday people often are only getting home by 6pm and are hungry, etc. and an instantaneous transition to watch the LCS is tedious for them.

How riot can fix their biased data conundrum is to have the EU LCS do a Best-of-2 for one split then the following split have it run a Best-of-3. The data that they, in my opinion, should look for is when peak viewership occurs for both splits, for each hour of broadcast what is the peak number and average number of viewers, viewer number in relation to highlight matches (i.e. Origen vs Fnatic), and whether the Bo2 or Bo3 format does a better job of correctly ranking teams. The NA LCS should do likewise.

I will assume that Bo3 is the best option for both regions. It allows for more airtime for sponsors and advertising, more preparation time between Bo3s for all those working at the event, and holds a higher probability in properly ranking teams. There are some valid concerns about the Best-of-3 format for broadcasting and program schedule purposes. They can be resolved and will be discussed further ina future article.