Smart Sort is functionality in eBird Mobile for Android was added in late November 2017. This new functionality uses "Status in area" from eBird frequency
Just like eBird maps gives frequency worldwide for any bird species in the world, such as Barn Swallow, one can imagine that each of those grid cells can provide frequency for all species as well. That is exactly what this new option is doing when you have a cell connection (this does not yet work offline).
For any 20x20 grid cell on the planet, the eBird app figures out first what 20x20 grid cell you are in. Once it knows that, it checks to see that there are at least 25 complete checklists for that grid cell for the current 3-week period (i.e., this week, the week before, and the week after, as in eBird bar charts). If there are fewer than 25 checklists in that spatio-temporal window, then it zooms out to a 60x60 km grid and tries again. If it still has too few checklists, it goes to 100x100km. If that fails, then it falls back to the regional data entry checklist. This allows it to be flexible and give finer scale information in areas with dense eBird data and still give spatially relevant results even when the data are sparse.
Once it knows what grid cells to use, it calculates the eBird frequency (again see more on the eBird bar charts page) and gives the average for all species for that day. Those results are then binned into three groups: frequent (6-100% frequency), infrequent (6%> and >0% frequency; orange half-circle) and unrecorded (0%; red circle). Since eBird data entry checklists are regional, but bird occurrence may follow strong patterns even within a region (e.g., coastal species only on coast and mountain species only at higher elevations) this new functionality can quickly push those species to the bottom of the list if they don't occur where you go birding. This should make it easier to quickly work through a short list of likely species!
A few caveats are worth understanding:
1) Subspecies are not reported uniformly, so data summaries are at the species level and subspecies simply follow the parent species. Frequency data are not stored for subspecies, so you may see a very rare subspecies showing up next to the common one (e.g., Myrtle and Audubon's showing together under Yellow-rumped Warbler, even if one is very uncommon and the other common).
2) Since other taxa (slashes, spuhs, hybrids) are reported only by some users, those are grouped with the unrecorded taxa at the bottom, Remember to search for these if you saw hybrids or distant scoters you could not identify (e.g., scoter sp.)
3) Sometimes a rarity that shows up and is seen by many in an area will cause it to appear as more frequent than expected because of the large number of observations. This is expected and will change over time as more data on common species are collected.
As always, if something is not showing up, you can add it by searching for it in the header bar; see tips on doing this efficiently here. finally, click "Tap to search entire database" to find a species that is not showing up at all (this will almost certainly be a real rarity!).
The image below is for a site in England, has Smart Sort checked, and is showing the three frequency bins. Black Swan of course is an escapee here, but we encourage reporting of such birds to track potentially expanding populations.