There are three key things that you have to have to calculate accurate efficiency numbers:
- You must be able to measure all of your volumes accurately. When you measure a wort's gravity, the volume of the wort is just as important as far as calculating efficiency.
- You must be able to measure gravity accurately (usually not a problem with a good hydrometer or a refractometer) and account for temperature variations.
- You must have the gravity potential of the grain you are using. If you think your grains are giving you 37 points per pound per gallon (PPG) and they are really only giving you 35 PPG, then your calculated efficiency will be artificially lower. Conversely, if you count on 35 PPG and your grain gives you 37, your calculated efficiency will be artificially high.
Using guestimations or approximations when trying to calculate efficiency will have a significant effect on the final result.
If you have changed grain batches or brands recently, it might be that the new grain just doesn't have the gravity potential that you think it does. I encountered this with a bag of Breiss malt. Efficiency measured about 5% lower than normal. In reality, it was just wimpy malt and I wasn't accounting for the lower yield. When I switched to Fawcett Marris Otter, my measured efficiency went right back up, all other processes remaining the same.
It sounds like all your processes are fine. Get the actual malt specs if you can and just for grins, try milling the grain twice on the same setting. The second pass will separate the husk from the kernel which is what you want. You know it's time for a grain mill, don't you?