To me the point seems that the benchmark is so misguided and missed the obvious error in the usage on LINQ that the results are not relevant. You should not take perf advice from the authors of that course.