Thread by @Fromagehomme

A few (very early) thoughts on the Azimio petition.

Let’s start with the strengths:

1. the KPMG audit did indeed raise a number of concerning points that need to be investigated
2. is true that the absence of elections for …

🧵 1/7

#KenyansDecides2022
#KenyanElection2022

… Governor in Kakamega and Mombasa may have reduced turnout in the presidential election for Odinga, though some of the accusations about why this happened are speculative and unevidenced.
3. If they can provide forms to substantiate the claim that 23,500 more people vote in …

… the Pres election in Kirinyaga than in the governorship election, that seems like a very high number that needs investigation.
4. If the claims about forms 34As being tampered with can be substantiated, this could cast doubt on the wider process - though many of these …

… claims are again unevidenced, at least so far as the written petition is concerned.
5. There is some legal ambiguity about exactly what happens if the commission is split when it comes to the results, and the appropriate course of action for the Chair, and further clarity …

… is required, which the Court may be able to provide.
6. There is a similar question about whether all constituency results need to be read out in Bomas before the Pres result can be declared - though one might have sympathy for the Chair’s actions here given the unrest in …

… Bomas at the time.

Taken together with issues related to the late opening/closing of some polling stations, and inconsistencies with the use of the manual register, these points merit serious consideration.

However …

Let’s move to the weaknesses.

The petition itself has many flaws. It presents a table of alternative results that only seems to add up to 99%. Given that results are shown to 3DP, this is unlikely to just be a rounding error. The table also seems to imply that Ruto won the …

… most votes. This seems to contradict a number of public statements by senior Azimio leaders, which some may feel calls into question their credibility.

Just as concerning, the whole table appears to be based on the idea that the Chair announced a voted turnout of 65.4%. …

… this is bizarre as this has already been shown to be a verbal error that was corrected at Bomas.

Basing a major petition claim on something that has already been explained and discredited seems to be a major strategic error.

It also suggests a touch of desperation where…

… the numbers are concerned. One reason for this may be that the main numbers documented in the petition, eg when it comes to discrepancies between different tallies, don’t seem to be enough to put Ruto below the 50% +1 threshold. Eg the discrepancies in this table only …

… add up to 33,208, which is well below the numbers circulated on social media before the petition was submitted, and - at least on their own - would not be enough to suggest the wrong outcome was announced.

The petition also seems to make basic errors. For example …

… the 33,000 votes come from 8 counties, which are said to represent 21% of all counties.

But 8/47*100=17 not 21.

This doesn’t mean the claims in the petition are false of course, but when challenging someone’s numbers it helps to get your own right.

There are also a number of claims in the petition that are hard to evaluate because they are based on evidence that will presumably be provided during by personal testimony.

This included various accusations of hacking and digital manipulation.

We will have to wait to see …

… what is supplied later in this regard, but cynics may wonder why, if the petition team has such good information on how the rigging was all done, they has so few hard figures and so few concrete examples of forms being changed at any level …

… the final point to make is that of course this petition tells one side of the story.

It naturally does not deal with the many accusations of unfairness made by the Kenya Kwanza camp.

These will need to be balanced against the Azimio claims.

They are likely to include …

…

- allegations that Chiefs and state officials were used to mobilise for Odinga
- allegations that the media was unduly influenced to provide favourable coverage for Odinga
- allegations that the Chair of the IEBC was physically assaulted at Bomas …

… my sense at this stage is that much will depend on three factors:

1. how the Court rules on whether the Chair had the right to declare the result under these circumstances.
2. whether accusations of hacking and the swapping of forms 34A can be supported by hard evidence …

… as opposed to hearsay
3. whether, when the IEBC releases the vote tallies for the other 5 elections, there are large discrepancies with the presidential tally.

Given the limited evidence suggested by the petition on point 2, point 3 could be a big one …

… finally, it is important to note that these petitions are put together under great time pressure. To be fair that means that mistakes and a lack of systematic analysis are more likely.

It may be that more information will come out during the proceedings, and so it is not …

… easy to come to a firm conclusion about the merits of the petition at this stage.

More soon …

Some interesting (and provisional) work by Charles Hornsby on his blog.

He generally finds that the turnout in different elections is pretty similar and there are not major concerns with the exception of Kirinyaga.

If sustained ...

charleshornsby.com/kenya-blog

... , this would make things harder for this part of the Azimio petition, as it would be more difficult to claim that the presidential vote was massively inflated, and evidence this on the basis of higher turnout in the presidency than in the lower level races ...

To quote Charles: "Generally the match is very good and much better than previous elections. The only case ... which is clearly anomalous + relevant is Kirinyaga, where there genuinely seems to be 20,000 more presidential votes than any of the other polls."

This therefore still leaves an important question: what happened in Kirinyaga, how was it possible, and what does this tell us about he quality and credibility of the electoral process ... Any thoughts Twitter mind?

And now tragedy: Kirinyaga returning officer Geoffrey Gitobu died on Monday, August 22, at around 2pm.

"Laikipia County Director of Criminal Investigations (DCI) Onesmus Towett has, however, dismissed rumours of possible killing of Gitobu."

citizen.digital/news/fresh-det…

Worth noting, though, that these (very provisional numbers which are not all from official sources) also suggest differences over 10k in votes in a number of other areas including:

- Nairobi
- Bungoma
- Kisii
- Vihiga

But note 1) these are not official results and 2) Charles

will have been looking at the percentage difference between the races with the 2 highest turnouts in the county + whether we see more than a 5% discrepancy between them

E.g. Nairobi is a big amount, but as a % it is relatively small given the vast number of votes cast...

A final question raised by the analysis of Charles Hornsby, and I will leave this here for now, is why there are so many more votes for Governor than any other position in Homa Bay.

Does this suggest ballot stuffing in the Governorship election? Or is it a typo (5 should be 3)?

If someone has alternative results for these areas/confirmation of results from formal IEBC forms/counts, please do share it here ...

So we seem to have indications that the tally for Governor for Kirinyaga was incomplete and missing a candidate. If that was true, the gap between the governor turnout and presidential turnout may disappear. In that case, there would not be that many major discrepancies ...

... there are, though, some 34A forms that are worth investigating.

Eg this form 34A on the IEBC portal from Gorgor Primary Polling Station (2/2), which has every single vote cast for the same candidate, and only has agent signatures from that party.

That said, ...

... I would suspect that we might find some similar examples from Nyanza (etc) for the other side if we looked through all of the forms 34A on the portal from that area.

I am trying to check the claim that the forms 34A and forms 34B do not match in some areas, but that may ...

... take a while so please bear with me

In the meantime, check out the update of Charles Hornsby's blog, which suggests Kirinyaga may not be problematic, corrects a typo for Homa Bay, and finds (again) there are far fewer discrepancies than last time

charleshornsby.com/kenya-blog

This is the link for the form 34A, for those that wish to check:

forms.iebc.or.ke/form/formA/1_3…

Some examples of this in other areas already provided eg from Uradi Primary School Polling Station (2/2).

As noted, though not common, all of the votes going to one candidate is something that sometimes happens in "strongholds", and while it is worth looking in to does not ...

... equate to evidence of rigging.

Also worth noting, of course, that further investigation may be very hard in stronghold areas where there were no agents for candidates from other parties, and media/observer coverage is unlikely.