I can see you’ve never caucused before.
This morning’s top stories:
*Late-reporting precincts give Sanders huge boost, put him close to lead in Iowa.
*DNC chair calls for full recount.
Just because you’re paranoid doesn’t mean they aren’t out to get you…
He did much better than expected so it’s at worst a win in that respect. 
If they do recount, they really should only check two things:
One, the number of “final alignment” votes each candidate got at each precinct, so the party can make sure the SDEs are calculated correctly - I found four precincts where candidates that did not meet the 15% minimum vote count still got SDEs (in three of the four, it appears that the 15% number was rounded to the nearest integer instead of being always rounded up);
Two, the results of any coin tosses - a recount of a precinct that results in confirming that a coin toss was needed should keep the original coin toss result (well, maybe except for that one on video…).
Yeah, if anything warranted a careful recount, it’s this shitshow.
Recounts rarely shift the results by more than half a percent. Unless there’s some reason to believe the reported vote count is RADICALLY wrong, why bother? By any reasonable standard, this was basically a tie. Let’s move on. It’s not a winner take all State, we don’t need to obsess about getting the vote count exactly right to eight decimal places.
By all accounts, they are taking all the time they need and being very careful to get the results right before they release them. A recount would only fuel suspicion that there was something fishy going on. If one of the actual candidates feels strongly that they were screwed, fine, let’s recount. Otherwise, let the DNC chair STFU about it.
Next time, Iowa needs to ditch the ridiculous caucus system and just be a flat-out primary.
Except we know for a fact that some of the released results are incorrect.
“The results released by the Iowa Democratic Party on Wednesday were riddled with inconsistencies and other flaws.”
This link was posted earlier by another poster: Iowa Caucus Results Riddled With Errors and Inconsistencies
From your cite:
So I still think a recount would be pointless and only drag this embarrassment to the Party out longer. Definitely retracting my statement about them being very careful, though.
I mean, seriously, Deval Patrick sweeping central Des Moines? How the hell do you not look at that and think “Maybe we should double check this”?
Right?
I largely agree with you that a recount wouldn’t accomplish much, but there’s something to said for having ‘final’ results that do not contain obvious and easily detectable errors.
Question for you.
I follow PredicitIt. Bloomberg certainly has statistically significant odds in many of the states you’re tracking, pretty much higher than Warren throughout. PredicitIt has him as 2nd in CA and third in NC ,MN, TX and MA, just to name the ones I checked.
But he’s not on your tracker … I have no claim to any sort of experience in statistics but how do you make that omission and keep it accurate? Are you just tracking 4 of the candidates relative to each other instead of as a percentage of the entire field?
This is the difficulty with saying that the errors left are probably small. This is pretty glaring. It took a bunch of people screaming about it. They changed it in real time without an announcement. I’m not saying the results are wrong, but I’ve lost faith that they’re accurate.
Good question.
When I started tracking, Bloomberg wasn’t a factor so I just left him out.
I think I may need to start collecting data with him included and switch to that dataset when when there’s enough data in there to show meaningful results.
Ok, data collection with Bloomy in the mix is in place.
I don’t think I answered your question explicitly earlier. Numbers are normalized so that they sum to one.
As of the latest count, I have found 90 definite errors in SDE calculation, with another 24 “likely.”
These include the previously mentioned four precincts where somebody got SDEs even though they did not have 15% of the final vote.
Example: Des Moines-72 in Polk County shows 6 people voting for Sanders and only 3 for Biden, but Biden got SDEs there while Sanders did not.
Meanwhile, in Graham Township in Johnson County, 56 people voted in the first round, but the data shows that nobody voted in the second round, even though SDEs were counted (and it was not a case of everybody who got votes in the first round got at least 15%, so a second round would be unnecessary as nobody could change their vote).
The problem is, a recount can plunge the shit out of this shitshow because the shitshow is a caucus, and because caucuses don’t operate according to the straightforward rules and conventions of primary and general elections, there’s always going to be a cloud of doubt - recanvassing, especially at the behest of the DNC, might actually make matters worse.
I say just move on from Iowa. The damage is done anyway and there’s no fixing the kind of damage that was inflicted, but bookmarking it actually keeps it lingering in the consciousness of voters and pundits. Move on to races that matter. And from now on, Iowa will have to share the stage with other states.
And the DNC may want to encourage other states to have their primary on or around the same date - the days of allowing Iowa to get out-sized attention for all the wrong reasons are over.
I asked back in the 90s “Why can’t they START with a Super Tuesday? Have a dozen primaries in diverse states!”
No one listened to me…
(Now I feel lucky to have been part of The Last Caucus Ever)
So it turns out Republican operatives jammed the phone lines to prevent the Democratic precincts from reporting their results to the statewide party. Apparently Republicans sabotage the Democrats and then belittle them for not being able to get the results in a timely manner.