When you first start brewing it's all guesswork.
There's a lot of numbers involved in brewing. Efficiency, volume, gravity, boiloff, time, and many many more. You need to know these things when you brew, but you'll only know them when you start brewing, so you guess. You use rules of thumb, estimates, and calculations to achieve approximate numbers that you can use to estimate your results. And then you brew with those numbers. If you're lucky, they're close enough to true values that you end up with decent beer. When you're just starting, "decent beer" is great, and you get convinced it's the best beer you've ever tasted. So that's fine.
As you grow more proficient as a brewer, though, you start noticing certain things. The OG is consistently a little low, the color is a little off, the attenuation is not what you were looking for, the beer is decent but it's not great, etc. At first you attribute this to mistakes in the particular batch, then you start blaming inconsistencies in the process. But after you improve the process, and repeat it consistently, it becomes a nagging feeling that you're missing something else. Eventually, you start looking at the underlying assumptions that you base your process on, and try to modify those assumptions to fit the factors of your particular system. This is called "Dialing in" your process.
For me the trigger was realizing that my OG is consistently two points lower than the recipe expected. This mean that the water to sugar ratio in my wort was slightly off. IE, I either had too much water, which meant my boiloff wasn't as much as I thought, or I didn't have as much sugar, which meant that the mash efficiency wasn't quite what I thought. Of the two, boiloff is the easier to measure, and is relatively constant. So lets start there.
There's two ways (that I can think of) to measure boiloff:
The first involves simply measuring how much liquid you had at the beginning of the boil, versus how much you have at the end. The difference is the stuff that boiled off. Simple in theory, but can have unforeseen problems. For example: If you use an immersion chiller, like I do, you need to put it into your wort 10-15 minutes before the end of the boil in order to sanitize it. Since you have to measure at the end of the boil you will either have to allow for the volume of the chiller (bigger than it looks) or take the chiller out (which negates the point of putting it in to sanitize in the first place. ) You can measure after you chill, of course, so long as you account for the shrinkage due to temperature drop. In my case, since I measure with a ruler stuck in the wort, this was not an option. I simply wasn't going to stick an un-sanitized, sugar coated metal rod into my wort.
The second method to measure the volume is calculating posthumously. That is, using the numbers you get to calculate what the boiloff was. This works as following: Lets say after I finished my sparge with 32Liters of 1.042 wort. After I boiled and cooled I sampled my wort and came up with an OG of 1.050. The amount of sugar did not change during the boil, only the amount of water. Therefore I can say this (Ov*Pg)/Og = Fv Where Ov is Original volume, Pg is preboil gravity, Og is Original gravity (at end of boil) and Fv is final volume. In this case: (32*42)/50=26.88 Given that my original volume was 32, I lost 5.22L of liquid during the boil, or roughly 16%.
The beauty of this method is that you can apply it retroactively. I take notes every time I brew, and I know my start volume, start gravity, and final gravity for each batch. By taking the last few batches, running them through this simple calculation, and averaging the result, I can get a pretty accurate measure of my boiloff. One I know that, I can go back to my calculations and figure out if I had too much water, or not enough sugar. Adjust, brew, test, adjust again, and so on until the numbers are consistent.
The sad part about this is that I'm going to go through all the trouble of dialing in my process, and then I'm going to change it again when I go HERMS this summer. But then again: I'll still be using the same kettle once I go HERMS, so doing this calibration now will make it much easier to isolate problems with the new setup down the road. So I guess it's worth it.