I thought with EI dosing you are supposed to change 50% of the water per week to "reset" the tank.
50% is just simply a convenient example, you can modify it, say do 80%, weekly water changes, or say 30%, but the possible error will go up with less %/less frequency, and down with more frequency and larger % water change.
50% is just an arbitrary example, there's no "have to do 50%" about it. Never was.
If you wish to explore these errors without doing all the math, one may do so with this modeling calculator:
Graphing:
http://ei.petalphile.com/
Nutrient dosing
http://calc.petalphile.com/
I use 2-4 week water changes on some tanks.
2x a week on one and then weekly on most.
Still, you can easily target a routine you desire with a little eye balling the dosing and and the plant response. Simply start with a non limitin glevel for the nutrients, then SLOW and PROGRESSIVELY redeuce a little bit each week and watch.
Once you hit a negative response, then you adjust the last highest dosing level.
At that point, you can likely do very few water changes, since the uptake and the demand is specific for your tank.
This works well if the CO2 and the light are well balanced, so less light will help almost universally, typically in the 1.5w/gal range for T5 over a deep tank say 24" deep etc.
Less light = less CO2 demand, so CO2 is easier to dose.
Then less light= less nutrient demand as well.
So now you have the most wiggle room, or the highest degree of management stability(or "user neglect" without poor results).
Then dosing is even easier. It's only as hard as folks make it.
Less light, good CO2, and common sense dosing.
Still, good sized water changes never hurt a CO2/Excel enriched planted tank I've had or seen.
So it's a good way to remove errors we make.
But if you wanna use them test kits, calibrate, otherwise it's pretty much just a guess that they are correct.
This is counter to the dogma and advice typically given. But if you think about it and know anything about Science, you know they must claibrate their methods to make sure the test work correctly, no guessing.
That's not honest.
Regards,
Tom Barr