Osun declared Accord candidate Ademola Adeleke the winner with 511,067 votes, against APC candidate Bola Oyebamiji’s 444,815 [E1]. The difference is 66,252 votes, an arithmetic margin fixed by those two declared tallies [E1]. That declaration is the result that crossed the Returning Officer’s line [E1]. It is also the number against which the election’s separate transparency machinery has to be judged.
The published declaration leaves five electoral quantities unavailable as distinct accounts: registered voters, BVAS-accredited voters, valid votes, rejected ballots and total ballots cast. Two other figures permit arithmetic checks, but do not supply any of those quantities. Accord’s 511,067 votes minus APC’s 444,815 gives the declared margin of 66,252 [E1]; 3,292 visible result images plus 471 outstanding gives the stated 3,763 polling units [E2].
IReV remained at 3,292 of 3,763 polling-unit results during the observed pause, leaving 471 outstanding from public view [E2]. The observed stasis ran from about 19:48 UTC until approximately 22:50 UTC, when uploads resumed at about 23:50 WAT [E2][E3]. The fixed 3,292 was therefore a count of polling-unit result images visible on a public portal, not a vote tally [E2]. The pause by itself establishes neither fraud, intentional suppression, a backend cause nor any change to votes [E2][E3].
IReV’s public object is a scanned EC8A polling-unit result, one layer in a longer election chain. BVAS accreditation is a different record; voters cast ballots before paper counting produces the polling-unit result, while certified forms continue through ward, LGA and state collation. Treating a stalled public upload as proof that any of those other stages stalled would collapse distinct systems into one unsupported allegation [E2][E4]. The transparency screen stopped moving while the certified count kept its paper shoes on [E1][E2].
Yiaga’s PRVT drew on 289 of 300 sampled polling units. Its estimated Accord range was 50.3–54.0%, containing the official share of about 51.9%, while its APC range was 43.0–46.8%, containing the official share of about 45.2% [E4][E5]. That is evidence of consistency between the sampled count and the official result, not proof that every electoral layer functioned correctly. Yiaga still called for a BVAS audit and greater transparency around the transmission protocol [E4].
The strongest counter-case is therefore operationally ordinary: polling-unit images can reach the public display late while the certified paper chain still produces the correct declared result [E1][E3][E4][E5]. The agreement between the declaration and the parallel sample gives that explanation real weight [E4][E5]. It does not account for why the upload queue stopped advancing or what happened to the 471 outstanding units during the pause [E2][E4]. A timestamped INEC upload-and-validation incident log, paired with a BVAS and transmission audit, is the discriminator that could test the public pause without guessing at its cause [E4].
Because uploads later resumed, 471 describes the portal at the point of stasis, not its eventual terminal state [E3]. Because the declared shares sat inside Yiaga’s parallel ranges, the pause alone supplies no basis for saying the declared vote was altered [E4][E5]. Both facts matter: a certified count can survive while its public transparency layer fails to explain its own interruption. Osun’s result was certified; the transparency mechanism still owes its own result.