Dj @djsoke.bsky.social · 19hLet me put some bad vibes out in the universe and say imagine if the final round of NYT’s polling just looks like their June polls again 😶 1230
Dj @djsoke.bsky.social · 03/10/2026So Lakshya pointed this out wrt to his polling, but Rs could have a serious youth turnout problem this year These huge 18-44 numbers for Talarico and Peltola are prob partially explained by that - two very young states with extremely fickle electorates 0523
Dj @djsoke.bsky.social · 03/10/2026Here are the demo changes in the NYT Siena polls from AK and TX from June to September In Alaska: Trump and Sullivan approval/fav dropped a ton and the age breakdown looks more normal In Texas: Talarico has firmer Black support and seemingly a kinder likely electorate 0504
Dj @djsoke.bsky.social · 03/10/2026Also NYT is my preferred polling aggregator right now so I went to their website and the gov races are all there But looking through the polls themselves it seems like they might have meant to have been withheld until tomorrow 😭 2280
Dj @djsoke.bsky.social · 03/10/2026NYT/Siena in AK: Peltola +7 | JKT + 4 | Begich +6 TX: Talarico + 6 | Hinojosa +4 OH: Brown +3 | Acton +4 IA: Hinson +1 | Sand +11 KS: Sen Tied | Gov: Masterson +4 1113518
Dj @djsoke.bsky.social · 03/10/2026mini AK primary thread: 2026's primary saw a large D turnout edge a la 2024, but different under the hood This year, the TO gap was super weighted to younger voters. How much higher Dem 18-24 TO was vs Rs in... 2024 - 1.3% 2026 - 12.4% The TO gap among 65+ voters shrank! 1277
Dj @djsoke.bsky.social · 02/10/2026There is no functional difference between a Talarico +0.5 poll and a Paxton +0.5 poll But the the difference in psychological impact on this site if the eventual NYT poll printed at a rounded +1 vs -1 is at the cosmic level 0392
Dj @djsoke.bsky.social · 01/10/2026It’s a curious quirk of the Fox polls that their LV screens benefit Rs RVs are Talarico +7, Turek +4 But while I don’t actually think that 2026’s electorate will be better for Rs than the full RV pool, the LV results are more reasonable imo 3275
Dj @djsoke.bsky.social · 01/10/2026As we get closer to more states opening up EV, important to remember that things aren't static or uniform For example, in 2024 in OH's biggest and bluest counties, IPEV was meaningfully bluer than VBM If you're ever looking for vote modality context, vote-mode.com is an excellent resource 0181
Dj @djsoke.bsky.social · 01/10/2026Precinct breakdown of IPEV in the nominally competitive CDs in VA 4192
Dj @djsoke.bsky.social · 01/10/2026This interactive is really awesome, and you should check it out, so much detail in there Since it’s comparing to *final* 2025 partisanship we should expect the electorate to drift redder until satellites open in big localities, then lurch bluer projects.statenavigate.com/interactives... 1295
Dj @djsoke.bsky.social · 30/09/2026Yeah, Texas is the EV-heaviest state in the country and it has been forever Big piece of that is cultural, TX had the first no-excuse ev system in the country (bc it’s such a physically large state, voting could have been a bit of a trek, people wanted flexibility) & it’s super ingrained now 4413
Dj @djsoke.bsky.social · 30/09/2026Turnout continues to increase across the board relative to 2025 in VA Harris +>15 localities now at 101% of same-day 2025 and Trump +>15 at 79% vs 94% & 72% last week 3544
Dj @djsoke.bsky.social · 30/09/2026Since apparently I’ve just enjoyed being stoned the past few days, here is how I would currently rank the 5 in the order that I think Ds win them *right now* AK > TX > ME > OH > IA (Obv subject to change) 7280
Dj @djsoke.bsky.social · 30/09/2026This is a pretty large RV-LV gap, but I think all the evidence we have (going back to last year’s November elections) suggests that it is dynamic playing out right now Texas’ turnout numbers relative to 2024 are gonna like insane 🤪 3321
Dj @djsoke.bsky.social · 30/09/2026imo the central problem from Marist is that their RV polls look like very optimistic LV polls Maybe it’s pedantic (but also think it’s true) that even if their numbers were “right,” I’d still think they flopped bc all evidence points to a Dem turnout advantage, which are an LV thing, not RVs 2161
Dj @djsoke.bsky.social · 30/09/2026Fwiw I think the Michigan Marist poll is the first reasonable one they’ve put out in a hot minute 1170
Dj @djsoke.bsky.social · 29/09/2026This is also how I see VA1/2/5 though I’d make VA-01 tilt D for now Will revisit in about a month 0210
Dj @djsoke.bsky.social · 29/09/2026People reading this as “Ds can’t have a huge year” instead of “I don’t think people understand what D+15 would truly look like” a self inflicted woundstatic.klipy.comKlasik Clasik: Chilling Duck Smokes CigaretteALT: Klasik Clasik: Chilling Duck Smokes Cigarette 3530
Dj @djsoke.bsky.social · 29/09/2026These numbers are very early still, and obv very bad for Rs - but even making huge assumptions that they 1) cleanly reflect vote intent 2) are sustained and 3) get mirrored nationally They still wouldn’t suggest a D+15 year 4404
Dj @djsoke.bsky.social · 29/09/2026Unfortunately still no precinct data, but here are the statewide numbers through day 7 in VA VBM still being returned very fast (>140% of same-day 2025) - need some time to see where that settles 2232
Dj @djsoke.bsky.social · 29/09/2026Still hoping we get something from them this week but it’s looking rough 😔 0130
Dj @djsoke.bsky.social · 28/09/2026At least in VA it seems like the Trump admin VBM actions are just making people return ballots faster Loudoun is at 99% of 2025 total requests, 106% of the same day. Returns are >5k vs 1675 same day 38314
Dj @djsoke.bsky.social · 28/09/2026NYT/Siena has been pretty consistent about starting their waves of poll releases on Tuesdays this year, so we’ll see if anything drops tomorrow I think if they don’t release a state this week it’s prob unlikely they’ll be re-polling it before the election 3222
Dj @djsoke.bsky.social · 27/09/2026i am not convinced Cubans will behave like other Latinos in this case and FL otherwise is a much tougher lift than any of the other states imo I just flatly don't think it's competitive before the year is already a massive blue wave 4200
Dj @djsoke.bsky.social · 27/09/2026did a little thought experiment and tried to make the simplest arguments for each side in the big senate races, excluding money and big environment deviations (e.g. "actually this a D+5/D+15 year") 7529
Dj @djsoke.bsky.social · 27/09/2026I mean I think this is self-evident atp To me, the majority-makers are clearly AK, TX, OH, ME, IA GA/NC aren’t competitive. I don’t think MI or NH are feasible tipping points, and I certainly don’t think FL or SC are 3362
Dj @djsoke.bsky.social · 26/09/2026Cumulative IPEV in VA through its first full week of EV! [so not only is the precinct data for this year not updating but the locality data for last year was being a pain in the ass z_z] 1264
Dj @djsoke.bsky.social · 26/09/2026Swim complete, thank yall who donated, it’s really generous beyond imagining ❤️ 0160
Dj @djsoke.bsky.social · 25/09/2026Made this graphically both for the cumulative and the day-5 only vote Def a visible dem TO advantage relative to 2025 (we're missing some data from yesterday, including 3 of the best-performing dem localities: [circled] CVille WBurg Albemarle) 1241
Dj @djsoke.bsky.social · 25/09/2026Posted ab this last month (tysm to those who donated!) but reposting since tomorrow I'll be swimming in SAA's Houston open water swim to raise money for MD Anderson cancer research! If you're up to donating a few $ to my fundraiser, you can do so here: www.swimacrossamerica.org/site/Donatio...swimacrossamerica.org Houston Open Water Swim - Swim Across America 1131
Dj @djsoke.bsky.social · 25/09/2026Not saying I think either are favored, but Hamilton (KS) and Hinojosa are two of the campaigns I’ve been most impressed with this cycle Hinojosa had no money and no name-rec against a massive warchest and another race taking up all the oxygen 4556
Dj @djsoke.bsky.social · 25/09/2026The update today is missing data from a few important localities for the competitive CDs (like Albemarle), so just gonna drop the locality statewide gradient Day 5 IPEV: 1201
Dj @djsoke.bsky.social · 25/09/2026Looks like they just released the gov races for yesterday’s states so discourse nuke paused 😌 1131
Dj @djsoke.bsky.social · 24/09/2026Can we at least wait until NYT’s October polls before melting down so we know exactly what to melt down about? What if their final round shows something different and you have to recalibrate your takes 😔✋ 2301
Dj @djsoke.bsky.social · 24/09/2026Did the one-word to describe focus group thing for Paxton and Talarico 🤠 3210
Dj @djsoke.bsky.social · 24/09/2026Day 4 update: VA02 had an issue where Isle of Wight only had one day of 2025 pct data, so it looked like it was waaaay outpacing 2025 when it wasn't Also so far VBM looks totally fine, requests are ahead of same-day 2025 but we'll have to see what returns eventually look like 2201
Dj @djsoke.bsky.social · 24/09/2026I’m assuming we’re getting the NYT Texas poll tomorrow and I am not ready for that discourse nuke 2351
Dj @djsoke.bsky.social · 24/09/2026AK posted its primary turnout report by party reg (timing, ik) and Ds put up a massive turnout edge. If translated to Nov, Peltola would def win - but if turnout normalizes, there would be a decent inherent swing towards Rs and that'd help Sullivan a lot 3304
Dj @djsoke.bsky.social · 23/09/2026Just got polled by YouGov for Texas, though it wasn’t really recognizable to me as a Hobby or UT/TPP poll, so not sure who the sponsor was 1270
Dj @djsoke.bsky.social · 23/09/2026Statewide precinct numbers, keeping in mind that the baseline is 2025 here 1172
Dj @djsoke.bsky.social · 23/09/2026Here is precinct data through the first three days of EV for the three nominally competitive VA CDs, using the same day in 2025 as the baseline Still extremely early 10301
Dj @djsoke.bsky.social · 23/09/2026Would be one of the most disastrous policy outcomes imaginable jfc 79110
Dj @djsoke.bsky.social · 23/09/2026Because Texas is so EV heavy, and Talarico/Hinojosa are likely to do better on eday, we’ll know if we are in a universe where Hinojosa can win (that ≠ *will*) pretty soon into our EV window Travis would be supercharged and the rurals flopping massively 2393
