Lies, Damn Lies, and Statistics
Statistics can lend an aura of scientific credibility to any sizing or benchmarking exercise. But don’t be fooled by your own statistics. The line between credibility and the appearance of credibility can be exceedingly fine.
I’m not immune from being fooled, either—I have more than once wished I had paid more attention during Statistics 101! For that reason, before drawing conclusions from statistics, I try especially hard to understand the following factors:
The raw data used to calculate the metrics, how the metrics have been calculated, and what they mean. Do low average response times signify satisfied users, for example? Maybe not, if response times are averaged across both quiet periods with low response ...
Get Configuring and Tuning Databases on the Solaris™ Platform now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.
