State Treasurer · South Carolina · 2026
General election · 2 on the November ballot · Incumbent running
Result
CalledCurtis Loftis (R) won this race.
- Curtis LoftisRepublicanIncumbentWinnerno count in the source
Also on the ballot, no count in the source: Vincent Coe (D).
- Called
- Called by Wikipedia editors, Nov 4, 4:43 PM ET (revision)
- Last updated
- Nov 5, 1:33 AM ET
Unofficial until the state certifies the result.
Before the election
Polls, ratings and the Vugol Tensor below were published before votes were counted. They are not results.
Model rating Leans Republican (Tilting) on the Vugol Vector model, a snapshot of today, not a probability.
No polls on file for this race. Model inputs: 3 of 7, updated yesterday. Page updated 2026-11-15. Sources · What is validated
ExperimentalNews attention 54/100, how much coverage the race gets, not who is winning. Updated 8 days ago. How it works
On the November ballot (2)
From reported primary results, checked 2026-11-05. Each state certifies its own ballot; see the source below.
Ballot source: https://en.wikipedia.org/w/index.php?title=2026_United_States_state_treasurer_elections&oldid=1383813983
Latest verified field polling
Verified = published by the pollster that conducted it, with a public source we link to.
No verified poll has been published for this race. The Vugol Vector call from our model is below, it is not a poll.
Polling average
No polls on file for this race yet. See latest polls across every race.
In the news
Couldn’t load news right now.
LabVugol Vector & Vugol Attention
Vugol Attention 54 Vugol Vector R · Tilting
Vugol Me model signals, not polls. Verified polling is in the Polls section above.
Vugol Me Attention (attention intensity)
Vugol Attention intensity index 0–100 · not a poll · normalized nightly
Vugol Vector
Proprietary ensemble model · a call, not a probability · model confidence: medium
The breakdown of this call (the model's multi-signal input matrix, input by input) is on Ultra Pro plans.
Vugol Tensor
How far this call can miss · a range, not a prediction · no verified poll here, so the range is wider and untested
Centre R+7.9. Four times in five the result lands between D+1.5 and R+17.3; nine in ten between D+4.1 and R+20.0. Chance the Democrat finishes ahead: 14%.
Polls have missed by about four points in a typical race and they miss together. How Tensor works →
Desk data: daily telemetry
Per-source daily series, CSV export and the Vugol Vector breakdown · paid plans
Sources
- Roster, official filings; fundraising from FEC datapoints where on file (https://api.open.fec.gov/).
- Vugol Attention inputs, Wikipedia pageviews (https://wikimedia.org/api/rest_v1/) and GDELT news volume and tone (https://www.gdeltproject.org/).
- Polls and results, individual polls and primary results listed on the Wikipedia election article (CC BY-SA 4.0), each poll linked to the pollster's release; money and election dates from the FEC (public domain). How rows are checked.
- Models attention-1.0.0 · vector-1.1.0 · attention inputs SHA-256 be94d9fe2333 · methodology · license CC BY 4.0 (data license).
Vugol Vector band: R · Tilting. Have a say in the Vugol race poll; it is open to signed-in voters and never feeds these numbers.
The four Vugol signals, side by side
They answer different questions and are kept apart. Full method: Methodology.
- Vugol Vector
- Our call for every race if today’s polls hold. A snapshot, not a win probability or a vote share.
- Vugol Tensor
- How sure we are: ranges and chamber-control odds around the Vector, built from past polling error. An experiment about uncertainty, not a better guess.
- Vugol Attention
- How much news and online noise a race gets, 0 to 100. It measures noise, not support: a high score says nothing about who is ahead.
- Vugol Pulse
- Our daily thumbs-up or thumbs-down question for readers. Anyone can answer, so it describes our readers, not the public. It is not a vote and never enters a poll average or any other signal.