
Funny, when criticism is discussed, you're the one to call it quits ("smart" pump performance automatically reduced, an FX5 user who added fluidized sand biofiltration to his FX5 set-up, another one reporting shaky parts, ...).:
You guys sure know how to beat a dead horse to death

There have been polls, more threads, fights, difference of opinions and everything in between....
Here you sense a victory, and you're bouncing up the walls...
But: If you run a statistical analysis, the results are not significantly different. People call that
deuce. (I used GraphPad InStat V 3.06, contingency table). Output below.
HarleyK
FX5 vs Eheim
Fisher's Exact Test
The two-sided P value is 0.4672,
considered not significant.
The row/column association is
not statistically significant.
Relative Risk
Relative risk = 0.7895
95% Confidence Interval: 0.4875 to 1.279
(using the approximation of Katz.)
Difference between the two proportions
Top row (Eheim):
Fraction in the left column: 0.4412
95% Confidence Interval of that fraction: 0.2716 to 0.6214
Bottom row (FX5):
Fraction in the left column: 0.5588
95% Confidence Interval of that fraction: 0.3786 to 0.7284
Difference:
Difference between the fractions: 0.1176
Standard error of the difference: 0.1213
95% confidence interval of difference: -0.1201 to 0.3554
Data analyzed
yes no Total
Eheim 15 19 34
( 22%) ( 28%) ( 50%)
FX5 19 15 34
( 28%) ( 22%) ( 50%)
-----------------------------------------------------------------
Total 34 34
( 50%) ( 50%) (100%)