The Elections Department (ELD) will give people an early indication of the possible result of today's Bukit Batok by-election shortly after polling closes at 8pm.
It will announce a sample count of the votes, which it did for the first time in last year's general election.
The ELD's aim in releasing the result is to prevent unnecessary speculation and reliance on unofficial sources of information while counting is still under way.
Tonight's sample count result will be derived from a random count of 100 ballot papers from each of the nine polling stations in Bukit Batok, and then weighted to account for the difference in the total number of votes cast at each station.
The result will be given in percentages, rounded off to the nearest whole number.
The sample count process has been done as an internal check against the final result in past polls, although GE2015 was the first time the sample count was made public.
At last year's general election, Singaporeans had a fairly reliable indicator of how votes went in each of the 16 group representation constituencies and 13 single-member constituencies by 11.15pm.
The first set of sample count results - which had included Bukit Batok - came in at 9.37pm.
Comparatively, the first official result came in at about 11.30pm, and the last after 3am. The official result for Bukit Batok was announced at about 1.20am.
The sample count results had correctly indicated all the eventual winners, even in the close contest of Aljunied GRC where there was a recount.
Its difference from the full count was below two percentage points in all but one of the 29 constituencies.
In Bukit Batok, the sample count showed Mr David Ong of the People's Action Party getting 74 per cent. It was 26 per cent for Mr Sadasivam Veriyah of the Singapore Democratic Party, and zero per cent for independent candidate Samir Salim Neji.
This proved comparable to the final result, in which Mr Ong took 73 per cent of the votes, Mr Sadasivam got 26.4 per cent, and Mr Samir had 0.6 per cent.
The sample counts have a confidence level of 95 per cent, with a margin of error of plus or minus four percentage points.
This means it is still possible that, despite a sample count result of 54 per cent to 46 per cent, the latter candidate could end up winning.
But the chance of such a swing taking place is less likely than when the difference is smaller, say, 51 per cent to 49 per cent.