Blocking J, Part 2: Vodhanel’s Recount Against NOCCCD’s Measure J School Bond at 54¼% for 59 New LA County Votes

It’s a reference to the Hunger Games movies and if you don’t get it don’t worry about it.  P.S. It’s GLENN VODHANEL!!??

Part 1, giving the background for this story, and lots of links for even more background, is here.

If you’re a fan of the phrase “whom gods destroy they first drive mad,” you’re going to love what’s happening with the Measure J recount.  It’s cosmically comical.

It had certainly begun to look to me as if the recount of Measure J ballots in Los Angeles — focusing only on provisionals ballots although other ballots had to be machine counted and there may even have had to be a 1% hand-count of county wide ballots, which is why I suppose they quoted a starting price of about $5000 per day — was not going to happen.  But it did.  The results are now posted and Orange County stands ready to do a much much larger recount than the mere 16 precincts in the North Orange County Community College District that are found in and around La Habra Heights.

Here’s what they posted:

As of Friday, December 12, 2014, the total election results that were certified  on November 28, 2014 for the above precincts are 2,029 YES votes and 1,918 NO votes. The total manual recount results for all of the above precincts recounted to date are 2,061 YES votes and 1,945 NO votes.

The statement is a little ambiguous.  The notice above mentions “precincts recounted to date” but it also does call it a total.  I don’t think that it is a mere subtotal for one day’s worth of work, but I can’t assert without contacting them that these are the final results.  Looks likely, though.

I note that in almost all cases where a precinct tally changed — and this is unlike what I had expected and what I unwittingly convinced Chris Nguyen over at OCPolitical to believe — the number of votes tallied in each precinct went up for both YES and NO votes in all but one instance, where NO votes went down by 1.  I don’t know whether to expect the same in OC — different counties may apply different standards to including provisional ballots in the initial count — but let’s assume for now that it does.

For those who don’t like doing their own math, this means that 32 YES votes and 27 NO votes were ADDED in this process. taking the YES total from the first year of Chelsea Clinton’s first term as President to the first year of Sasha Obama’s first term as President (Republicans, you may substitute your own hopes) and the NO total from the last year of World War I to the last year of World War II.  But that sort of side fact (or “fact”) is not why you are reading this.

You are reading this because you know that it takes 55% to win the election and the total across both counties stood, prior to these results, at a just a smidge under 55.01%.  (All right, all right — it was 55.0097976874862%.)  And so the first thing that you want to know is: Is 32 Yes and 27 No more or less than 55%?

And the answer is: it’s less — barely.

We’ve just stirred 59 votes that went 54.237288135593220338983050847458% (that’s 54.2372881355932% to you, Chris Nguyen) for YES into the pot.  So what you really really want to know is the new total across both counties.  Here it is, with lots of scary numbers, so people who can’t handle the match can just skip ahead to where the text is no longer indented:

We now have 84,812 YES votes across the two counties.

We now have 69,365 NO votes across the two counties.

We now have 154,177 OVERALL votes across the two counties.

So: Measure J currently has 55.009502065807481012083514402278% of the vote.  To accommodate Chris Nguyen’s calculator, we’ll round it off to 15 digits, so that’s  55.0095020658075%.

So we need to take 55.0097976874862% and subtract 55.0095020658075%.  Easy!

This is a loss, meaning a negative gain, of –0.0002956216787% of the –0.0097976874862% reduction in margin necessary to reverse the result.

How much of the distance did the Los Angeles recount carry Tony Bushala Glenn Vodhanel towards his goal?  We divide the first number directly above by the second.

And: –0.0002956216787% over –0.0097976874862% is … 0.03017259727%.

In other words,  in Los Angeles County Tony Bushala Glenn Vodhanel has moved about 1/33 of the distance he needs to go.

So what we want to know now is: how many times more ballots were there in Orange County than Los Angeles County?

Well, we don’t know when it comes to uncounted provisionals!  But we can estimate based on the total number of pre-recount votes in each county.

Los Angeles County had 3,947 votes counted pre-recount (now up to 4,006, of course.)

Orange County has had 150,171 votes counted so far.  Does their sum equal 154,118?  Yep, it does — we’re golden.  So let’s divide them.

150,171 divided by 3,947 equals … 38.0468710412972.

So that is our expected result.  IF Orange County acts EXACTLY THE SAME WAY as Los Angeles County — meaning a little over 38 times as many OC ballots come in at the same 54.2372% YES as LA — we’d end up with about 1,218 more YES votes and 1027 more NO votes.

Add those to our totals and you get 86,030 YES votes and 70,392 NO votes.  That’s 156,422 votes total.

What percentage would YES have?  It would have 54.9986574778484%.

In other words, it would lose by -0.0017454140790%.  That’s about 1/75000.  (To be precise, it’s about 1/74487.)

Out of 156,000 — unless I’m missing something, that’s a margin of about 2 votes.  Measure J would be expected to lose, if everything is counted and the LA trend continues and the Orange County precinct yield the exact same number of ballots per precinct as those in LA — by 2 votes.

(I literally laughed out loud when I saw this.)

OK, that’s enough arithmetic.  Back to politics — and law.

[NOTE: I’ve been corrected — it’s Glenn Vodhanel of Brea paying for the recount, or at least fronting for someone if not paying personally.]

This microscopic projected margin makes one want to laugh and cry at the same time.  I don’t see how Bushala can’t go ahead with it (not so sure now that it’s Vodhanel instead), but he’s got an almost exactly 50% chance of losing his whole stake — and that’s making all sorts of assumptions that, while the best we have without comparing the size of the precincts and the historic numbers of provisionals (which, Dear Reader, I am not going to do for you — or for anyone for free) almost surely aren’t true.  Furthermore, if you extrapolate from LA and there was one more YES vote, it would be a pretty easy win; if one more NO vote, a pretty easy loss.  That’s why you don’t extrapolate and put your money on something — you do it just to arrive at a “best guess” — and in this case the best guess is one that makes Bao Nguyen’s victory this year look like Ed Royce’s.

It would be like taking the “not-black” bet on a roulette wheel almost 40,000 red numbers, one green, and 40,000 black.  Yes, the odds are in your favor, but you’d better not love your $50,000 (or whatever it turns out to be) too much.

Now here is where we reach, and dent, the bounds of my knowledge of the law of recounts.  (And I have been learning as I go, I admit.)  I believe that you only get to count a county if you recount the entire county — meaning only those ballots from within the NOCCCD region, of course.  (In other words, it’s not like the Perez recount against Betty Yee where he could stop the count if he ever ended up ahead in completed counties.  There’s only one county left here!)  Bushala Vodhanel  can’t just pull ahead in the count one day, pull the emergency brake, and then declare victory and dare NOCCCD to take over paying for it.  NOCCCD is happy with the existing election results; the danger for them was if Bushala pulled ahead in LA.  And Bushala Vodhanel’s inability to pull ahead based on the LA vote means that NOCCCD will never have to take over asking for the count.  If Bushala Vodhanel wants it, he has to go all of the way through until the end.

So imagine this: Bushala Vodhanel could pull ahead in the count right away, hold it for days — and then lose it again on the last day.  If he wins, you and I (if you live in OC) pay the tab.  If he loses, he pays the tab.  And it’s a big tab.  And with the projection — however flawed the initial assumptions — being a narrow victory, mathematically he can’t expect to pull ahead until the last day (although of course it might happen.)  At that point, all of the money is spent.

So: Bushala Vodhanel is going to have to go all-in, pretty much right away,  on essentially a coin flip.  It’s good for him that he could afford the loss, because once he gets on this horse I don’t think he ever jumps off.

(And if NOCCCD doesn’t have a really good election lawyer present there — as good as Bushala’s non-lawyer recount savant Tim Whitacre is — then they are nuts.  My guess is that they are not nuts.)

If you’re a math geek or a political geek (and especially if you’re both), you just have to enjoy this — because something like this doesn’t come along all that often.  A 2-vote projected margin out of 160,000 votes!  That’s … wow!  What are the odds?