Sorry for the long post, but want to run by a concept. I recently installed nightscout and the cob-iop module and bwp. I understand it has not been kept up with people moving on to Open APS and I have only just started reviewing that work. In looking at my daughter's cgm tracings over the last month, and also knowing something about endocrinology and metabolism (i am a doc), I wonder if instead of worrying about carb absorption rates, which are tricky and primarily related to glycemic index, we model it differently. If she eats and doesn't take insulin, it is not like the glucose goes up and down, it goes up, stays up and comes down very slowly. Probably related to residual pancreatic function slowly dealing with the carbs or glucosuria (peeing it out) if it is very high. I notice that the bwp is pretty accurate for predicted blood glucose if all the carbs are dealt with, but it does not work so well if you try to use the carbs on board to mentally adjust. Instead what if we define a concept called residual insulin activity (ria) which is the amount of insulin left (based on carb factor) after the carbs are "dealt with" at whatever rate the insulin and carbs actually absorb. the ria is then used to calculate the eventual glucose result (egr) and that is displayed. Would then change the bwp to continue to list units if over the target and Gs of carbohydrates if under the target. Thoughts?