We use Survey123 connect to conduct salmonid monitoring. We bin fish by basin, species, length, and whether they were microchipped that day or a recapture. For this repetitive data, I use a repeat. Within the repeat, I had a whole host of nested if statements that tallied the binned fish. Unfortunately, the counts were not always accurate.
The errors occurred when a new repeat was added on an incorrect entry. For example, if a steelhead was marked as a coho, this error was tallied even after the initial error was fixed before sending the data. Furthermore, if the incorrect repeat is deleted, the tally remained. Is there a way to tally binned fish that recalculates when the repeat is corrected?
Without seeing your calculations it will be hard to pinpoint where the breakdown is, if you can share snips of you xlsx it will be easier to trouble shoot.
Also have you experimented with the different calculation modes, They are all set to auto by default, you may want to try setting to 'always' for these tally calculations that way they always update when data is updated or deleted.
Survey123 Tricks of the Trade: Calculation modes - Esri Community