My count (obviously not necessarily the same as the battery maker) is by the BMS, a Daly 48 volt version. i highly doubt that they have had any contact with CALB, (or Eve, Lishan etc.) and are setting a value that their calculations show as a good approximate. they might be averaging charge vs discharge to come up with an approximate wear and tear cycle that roughly corresponds with a full cycle according to the battery maker.
this would also explain why some folks get 10,000 cycles on a battery pack that is only supposed to get 3000 (if you count each day as a cycle).
this would make the most sense to me for what they are offering. if 100 amp cell draws out a total of 90 amp hours over x period of time even though it was recharged say twice... it aggregates the same amount of wear and tear as a full cycle hence it counts it as one???? i don't know as I am not privy to the programing that went into these devices and they are not about to tell us peons what could be propriety info to them. Remember they are not bad engineers, they are just stifled by a government and a socio-economic model that encourages uniformity/mediocrity not standouts or shining stars. unleash them and we should be very afraid for our socio-economic model....