I added up the weights of all of the ingredients and got 835.5 grams. I believe you meant to say that the input was about 836 grams rather than 846 grams. I calculated a loss of about 3%. That would not be unusual for a hand kneaded dough. It is also the reason why the dough calculating tools have what is called a "bowl residue compensation" input to compensate for such losses. Those losses are typically due to small amounts of the ingredients and dough that stick to surfaces of the bowl, measuring cups (mainly water and any other wet ingredients), the agitators (hook, paddle, etc.), work surfaces, dough scrapers, spoons an spatulas, and the hands. In your case, you perhaps ended up with more of the dough sticking to your fingers. I have found that when I use a standard stand mixer, a value of 1.5% for the bowl residue compensation works well. For a food processor or a bread maker, a value of 1% is a good value. For hand kneading, 3% or more would be quite normal. Recently, I have been using a food processor to make a dough with a lot of sticky, gooey molasses. I have been using 4% as the bowl residue compensation. I thought for sure by scraping everything and getting all of the ingredients into the processor bowl and off of my fingers that I wouldn't have such losses, but I was wrong.
If you use one of the dough calculating tools, such as the expanded dough calculating tool at http://www.pizzamaking.com/expanded_calculator.html
, you can use the bowl residue compensation feature to modify the ingredient values to compensate for the losses that are likely to occur. If it turns out that the final dough weighs more than the dough without the bowl residue compensation, which can occur if the value of the bowl residue compensation was too high, then you simply scale the weight of the dough ball back to the calculated value. So, my view is that it really doesn't matter which method you use so long as you select the appropriate bowl residue compensation value.