Yes and yes. Does that mean that I am wrong? Maybe, maybe not.
In real life, I have monthly meetings to discuss data analytics. One thing that recently occurred was someone else’s analysis of a case-control study with two subgroups of cases. Each “sample” had ~50,000 data variables per sample and >200,000 samples total. You can imagine this data requires supercomputing for analysis or filtering to make the dataset manageable.
As you can imagine, this type of data analysis requires considerable expertise and rigor to analyze correctly. The primary individual responsible for this analysis was a junior faculty member who has a PhD in statistics. He typically provides high quality analysis. Among our team, I was the only one to raise concerns about the data analytics because it did not make logical sense and failed what I would call a simple reality check.
Over the next several meetings and privately, I continued to voice my concerns to no avail. I also analyzed a subset of the data for myself. Low and behold, a sheepish e-mail was later sent out to the team that his prior reported analysis had a simple error, leading to erroneous conclusions.
The reason that I relate this story is there are time when someone says “You are wrong.” They may or may not be correct.
Blizzard’s data analysis may be completely correct. Their analytical methods and dataset that they used is proprietary. Their data in the API is not. In the case that I described above, there was enough transparency to make an independent analysis of the data.
There are very few people who post on these boards who really play around/analyze the leaderboard data. In general, the results of the analyses are reasonably similar even when using divergent methods. I am just stuck by game blog analysis that appears in opposition.