I received earlier today from an old colleague a copy of the results NASA released, which purport to be the true election results for the presidency from the August contest, a 17 MB file with a summary and detailed results for every polling station. For the first time, if this file is a genuine NASA file (and not dirty tricks), this allows us to take a look at the basis behind NASA's claims, and to apply the same level of scrutiny to their allegations as has already been applied to the official IEBC results.
From a couple of hours review, on a sample basis I consider it likely that the summary page is a (clumsy) forgery.
Most obviously, the turnouts have been massively increased in these numbers, implausibly so in my opinion, as NASA claims turnouts of 97% and 98% across entire counties, which is simply impossible.
Picking one county, Lamu, chosen because it was small and marginal, we can drill down in more detail to understand the discrepancies.
The official presidential turnout for Lamu (East plus West) was 50,108 (71%) of which Kenyatta got 23,905 (48%) of the vote to Odinga’s 24,421 (49%)
(note that the Lamu East RO couldn't add up and the total valid votes of 13,524 is 243 less than it should have been)
Here, in NASA’s alleged summary Lamu has 55,970 votes cast (5000 more, an 82% turnout), of which Uhuru gets 21,553, two thousand less (38%) and Odinga 33,709 or 9000 more (60%):
A turnout of 80.2% would be very surprising in Lamu, when the parliamentary results had a 73% turnout, and the Governorship (which Jubilee won) 72%. For this to be true, it would require 10% of Lamu voters to only have voted for the presidency and not the other elections.
Also, I gritted my teeth and re-added the 167 individual polling station results for Raila and Uhuru in the NASA attachment. These give results of 22,120 for Uhuru, 27,016 for Raila, an entirely different result to that in NASA's summary. I rechecked and the number of polling stations is the same as in the official results. I may have made some minor typing mistakes, but the totals in the NASA attachment at polling station match neither those given by NASA in the summary nor the official results.
In other words, the results in the summary page appear not to be the total of the individual polling stations in Lamu but to be a clumsy fake, with at least 6000 votes added in the summary to Odinga's vote, which is contradicted by the file's own detailed attachment.
This is only a couple of hours work, and I am very happy to be disproved if someone would like to take up the challenge. A test of Tana River, also small, should allow verification of whether the issue I have seen is indeed (as a I suspect from the document's turnout claims) systemic