Friday, October 30, 2026

i've made my point.

if we had decent poling with big sample sizes, we could make some better projections, but that rarely happens in senate races. when low quality polls with small sample sizes, bad methodologies and high margins appear to produce surprising outcomes, that may say more about the low quality of the polling than the likeliness of the result being projected.

if i did a poll of 500 people online that told me that piss tasted better than apple juice, i might question the accuracy of the poll before i changed my grocery list.

if these polls are completely wrong and trump wins kansas by 10 points, that should not surprise anybody. these polls suck.
what about michigan?

poll source landline or internet end
date
outcome valid
poll?
https://ippsr.msu.edu/sites/default/files/soss/soss95.pdf internet oct 19 in the margin no
https://www.clickondetroit.com/news/local/2026/10/27/democratic-candidate-el-sayed-leads-michigan-us-senate-race-new-poll-finds/ landline oct 22 in the margin yes
https://insideradvantage.com/insideradvantage-senate-surveys-mi-el-sayed-47-rogers-46-ia-hinson-47-turek-45-oh-brown-47-husted-44/ internet oct 22 in the margin no
https://polisci.msu.edu/_assets/MI_2026_pre_election_report.pdf mail-outs
sms
oct 23 these results
were manpulated
to "correct for bias".
this is invalid.
the outcome should
not be reported.
no
https://afightworthhaving.com/polling/2026/michigan landline oct 23 in the margin yes
https://www.documentcloud.org/documents/28856720-cnn-poll-conducted-by-ssrs-michigan-october/ internet
and
landline
oct 27 in the margin low quality
mixed sampling
https://www.cbsnews.com/news/abdul-el-sayed-haley-stevens-feud-michigan-senate/?intcid=CNR-02-0623 internet oct 27 in the margin no
https://x.com/joshorton/status/2115883461076545615 ? oct 28 in the margin no
https://www.rasmussenreports.com/public_content/politics/elections/election_2026/election_2026_michigan_senate internet
and
landline
oct 29 in the margin low quality
mixed sampling


the only poll with a result that showed el sayed ahead and wasn't in the margin was conducted by a researcher at the university of michigan who claims to have "corrected the results for biases". right. i would describe that as rather open political manipulation.

everything else is pretty weak, given that michigan is historically democratic.

in theory, el-sayed should benefit in michigan from the things republicans have going for them in texas and kansas. if the cultural inertia and ground game is stronger for the republicans in texas, it should be stronger for the democrats in michigan.

however, it's not clear he can rely on that. el-sayed has struggled to win over key components of the democratic coalition and he may struggle to get the vote out. the outgoing governor and his opponent in the primary have been silent or hostile.

you also have to take a bradley effect into account.

i consider michigan to be in play in ways that kansas and texas aren't. i will not claim the data supports a flip, but it certainly doesn't support a hold, either.
what about kansas?

less results, but they're better.
 
poll source landline or internet end
date
outcome valid
poll?
https://afightworthhaving.com/polling/2026/kansas landline oct 23 in the margin yes
https://www.documentcloud.org/documents/28856723-cnn-poll-conducted-by-ssrs-kansas-october/ internet
and
landline
oct 26 in the margin low quality
mixed sampling
https://www.washingtonpost.com/politics/2026/10/30/we-polled-kansas-democrats-wanted-talk-republicans-not-so-much/ landline oct 26 d+1
past
margin*
yes
https://emersoncollegepolling.com/october-2026-kansas-poll-hamilton-maintains-edge-over-marshall/ internet oct 28 in the margin no


the washington post survey is notable but the analysis posted with the survey indicated that democrats in kansas were eager to talk about politics on the phone and republicans weren't. that makes sense to me. i did phone work as a kid. in kansas, democrats are far less numerous, and perhaps a little lonely.

these results are neither conclusive nor convincing. again, it's extremely unlikely that the democrats flip kansas, given the on the ground difficulties. the data does not support the idea that this is going to flip.


if i was to describe a "high quality poll", it would have random sampling and be conducted with phones. online sampling is not polling, it's political manipulation. averaging out 100 bad polls (which is what silver does) and trying to weight it by applying a pile of arbitrary personal biases is the worst approach possible. snapshot polling is preferable to aggregate analyses.

i guess i need to do this.

there's a list of recent polls at the texas senate wiki site. let's cut it off on oct 19th, the day before two weeks before the election. which polls are high quality and which aren't?

poll source landline or internet end
date
outcome valid
poll?
https://www.foxnews.com/politics/fox-news-poll-texas-senate-race-galloping-toward-photo-finish landline oct 19 in the margin yes
https://emersoncollegepolling.com/texas-2026-poll-toss-up/ internet oct 19 in the margin no
https://cdn.votehub.com/pdfs/VoteHub_Verasight_Oct_2026_Texas_Report_D8A925D0.pdf internet oct 20 in the margin no
https://www.univision.com/noticias/politica/poll-univision-texas-latinos internet oct 21 in the margin no
https://www.activote.net/polls/2026/tx-u-s-senate/2026-10-21/ internet oct 21 in the margin no
https://www.rasmussenreports.com/public_content/politics/elections/election_2026/election_2026_texas_senate internet
and
landline
oct 21 in the margin low quality
mixed sampling
https://www.wfaa.com/article/news/politics/elections/new-poll-talarico-abbott-lead-as-early-voting-in-texas-nears-close/287-7f4cca13-6e1d-433c-9041-727931b591cd internet
and
landline
oct 25 in the margin yes
https://254labs.io/news/2026-10-tx internet oct 26 in the margin no
https://www.documentcloud.org/documents/28856724-cnn-poll-conducted-by-ssrs-texas-october/ internet
and
landline
oct 27 in the margin low quality
mixed sampling
https://drive.google.com/file/d/1UACnmnM84Dd1hU0l6iqHNgXkyh5oCelO/view internet oct 27 in the margin no
https://www.activote.net/polls/2026/tx-u-s-senate/ internet oct 28 in the margin no
https://quantusinsights.org/polling/f/texas-survey-paxton-up-one-abbott-ahead-by-five internet
and
landline
oct 28 in the margin low quality
mixed sampling


if there is further polling in texas, it's not at wikipedia.

all of the polling, both real and fake, is in the margin, and it is all mixed sample or worse and consequently of low quality. the only high quality poll was done for fox news two weeks ago and reported a result in the margin of error. the cnn poll intentionally generated a higher than normally acceptable margin of error by using weird methodology and then put the democrats on the outside fringe of that large margin, which inflated the spread through bad sampling practices, and is suspected political manipulation for that reason.

given the myriad of real world factors working against the democrats in texas on election day, it is highly unlikely that they will win this seat. that will not be a polling error. the polling is inconclusive.

i consequently see no reason to expect this seat to flip.
 
we've been here before, as well. remember beto o'rourke?
i actually don't think that gay people should be allowed to get married, but i don't think that straight people should be allowed to get married, either. i support the total abolition of marriage as a secular concept. the state should tax people as individuals. (personal) property should be owned by individuals. things like health care coverage should apply to individuals. marriage, as a state-sanctioned idea, is a relic of feudalism and should cease to exist.

if people want to get married as a religious sacrament, i have no interest in this, and the state should have no interest in it, either. it should have no legal relevance. heterosexual religious marriage contracts should be null and void and irrelevant in a legal sense.

it would follow that if gay people can find a religion that supports their ideas about marriage, that's their prerogative, but the state should have no interaction with such things. religious gay marriage contracts should also be null and void and irrelevant in a legal sense.

i do not support allowing gay people to adopt children and think that when children that are born into heterosexual relationships find themselves with gay parents the court should award sole custody to the heterosexual partner explicitly on the basis of the gay parent being gay. i agree with arguments about homosexuality being a mental health issue and would support a limited reintroduction of homosexuality to the dsm.

while i support the state leaving the gays alone and not attacking or persecuting them, i have no interest in the politics of homosexual rights, and don't support initiatives to integrate the state with gay politics. 

i am transgendered. i am not gay. i support maintaining heterosexual norms in society, broadly. subversion has a place in the underground, not in the mainstream. homosexuality belongs underground.
i would likely vote for the democratic candidates if i lived in ohio or maine, but no in texas, michigan or kansas. and i'll say again that i'd rather vote for ashley hinson than josh turek, and it's not a difficult choice for me.
all that a poll that has the democrats ahead by +2 with a 4 point margin of error says is that the outcome is in the margin of error. if the republicans win, that is not a polling error. rather, thinking it is a polling error is an error in reading the polls.

if you have 20 polls that have the democrats ahead but in the margin of error in 20 states and they lose all of them, that is still not a polling error. it's arguably reflective of a slight polling bias, but the polls don't say the democrats are barely ahead. the polls says that the outcome is unclear.

then, when you pile on incumbency effects, get out the vote, voter turnout demographics (age, income) and cultural inertia, these in-the-margin polls in red states all tip back towards the republicans. every single one of these flips is independently extremely unlikely, except perhaps north carolina (where the democratic candidate is popular and has an edge for that reason. that's local, not national. it has nothing to do with trump). that's the right mathematical model - every senate race is an independent experiment and a weighted coin toss in the republicans' favour, and the likelihood that the democrats will win more than two seats becomes extremely remote.
i believe this is footage of a young adam hamilton, from the 90s.

i just want to summarize what the democrats are running on

- in texas, the democatic candidate has refused to criticize fossil fuel development or refuse money from oil companies. as a leftist, i would not vote for the democratic candidate if i lived in texas.
- in kansas, the democrats are running a bigoted christian pastor that thinks non-christians are all going to hell. as a leftist, i would not vote for the democratic candidate in kansas.
- in iowa, they democrats are actually running transphobic ads on tv and the republican candidate is clerly the less conservative option. if i was in iowa, i would vote for the republican candidate, because she's less conservative - because i'm a leftist.
- in michigan, they're running a muslim with a history of supporting right-wing social policies and extreme right, fascist groups in the middle east. if i was in michigan, i would not vote for the democratic candidate, as a leftist.

these aren't left-wing political candidates. they're on the moderate right to the far right. 

i will reiterate that the data does not support the idea of a democratic "blue wave", which is an idea that is nonsensical in the american political system. the data says that a republican hold is overwhelmingly likely. an extremely good night for the democrats would be a two seat pickup.

however, if the democrats manage to win these seats - and you should not listen to nate silver. he's always wrong about everything. why does the media keep quoting this guy? he has the worst track record in polling that i've ever seen. his model sucks. you'd be better off asking a poli sci major. i've spent hours correcting him, debunking him and ripping him apart in this space - it's not reflective of some political shift to the left, it's because the democrats are moving hard to the right and picking up issues of concern to far right voters.

the senate election is not a single national election for one candidate. trump is almost irrelevant. the senate election is a series of independent experiments taking place in isolation of each other. the polling (which is low quality across the board. don't accept the label "high quality polling". that's right out of orwell.) does not support the projection that several unlikely to very unlikely events will happen at the same time.

when media messaging all focuses on a handful of talking points - which is what the democratic media s doing - it suggests an attempt to control a message and manufacture a narrative. that's actually a voter suppression tactic. what democratic media appears to be doing in projecting a landslide is attempting to suppress the vote. it's not based in reality, polling or science - it's nonsense, projection, faith and magical thinking.