Tempdrop's algorithm takes your whole night's data and calculates your sleeping temperature. It works to both learn your unique sleeping temperature patterns and filter out disturbances to temperature readings.
Your first 14 results
At first, Tempdrop gives you an accurate core body temperature without filtering variables* that may have affected your result. During this early data-gathering phase, some users may experience rocky results but, for others, their cycle pattern is discernible right out of the box.
15 nights of data
The real power of Tempdrop starts at 15 nights of data collection. At this point, the variable filtering starts, and your previous two results will begin to adjust retroactively with each new sync. The algorithm is making these changes based on the new data as well as previous patterns.
You will see your cycle charts improve, getting better and better from this point on. The retroactive updates to your last two results will continue with each new reading but will likely be less significant over time. The adjustment size will be most notable when you have a "questionable" or out-of-the-ordinary night (poor sleep, alcohol, etc.) and for women who have very disturbed sleep patterns.
As Tempdrop is a new technology, some may feel a little uncomfortable with the changing results at first. These users may choose to wait 1-2 days before reading anything into the most recent temperature as it may change tomorrow. Over time, you will become more comfortable and be able to recognize right away if your first day is likely to significantly change with the syncing of new data.
The algorithm will continue to learn and adapt to your circumstances throughout life, so don't worry if your sleep habits, environment or, cycle patterns change.
*Variables can include, but are not limited to, blankets, environment temperature, drinking alcohol, disturbed sleep, etc.