Not defending the use of Google, but you can't say "math is a bitch when it doesn't go your way", when the math you use (in exaggeration, I hope) involves rating the number of incidents on two different constructions as the same or (constant) over time. That's a terrible way to extrapolate data and fundamentally flawed.Math is a bitch when it doesn't go your way. I'm not arguing in any way, form or fashion that the EB is more reliable than the 5.3. What I am saying is that using Google as any type of indicator of reliability is poor at best. Now, if there was some type of report that showed the 3.5 EB was suffering 5% failure rates versus an industry standard of 1% (numbers are made up) than I would absolutely agree that the EB has problems.