They Know Your Vote Before You Even Fill Out the Ballot

Mark Cook March 22, 2026

Are you ready to become a lot smarter? Here you go:

Imagine stepping into a polling place, checking in to vote, and in that very moment, someone else already knows how you voted. Imagine mailing in your ballot or dropping it in a drop-box, and someone else already knows how you voted. Not because they guessed. This is not science fiction. Very smart bad actors created a system that enables this, right under your noses. And you bought it under the guise of ‘convenience’. This is the core of modern election manipulation: I've named it ‘feedback loop predictive modeling’. It operates like a living algorithm—collecting data, refining predictions, and adjusting outcomes—all while staying completely hidden from audits, observers, and even many of the top experts.

Most people fighting for election integrity focus on glaringly obvious problems: broken machines, hacked software, or suspicious ballots. Those are real issues, but they are only the surface scratches. The real danger lies deeper, in a sophisticated, multi-layered process that bad actors—highly intelligent, well-resourced, and patient—have engineered over decades. And THIS is what they don't want anyone to know. They no longer need to tamper with ballots at midnight. They can manipulate the game before it starts, using data we willingly give away every day without our knowledge. And because this method works up front, it evades every post-election check. That is why leaders—politicians, officials, activists—must seek out and listen to the rare voices who understand the full threat model. There are not many. In fact, I think there may only be a couple of us who see it this clearly. We must be realistic: Partial fixes from good intentions will never close the gap. We need the whole picture.

Let us start with something familiar: weather forecasting. Everyone understands how it evolved. Back in the old days, people relied on simple patterns—winters get colder, summers warmer, days shorten in fall. That was seasonal modeling: basic, reliable, but limited. Then technology arrived. Satellites circled the Earth, radars swept storms, weather balloons measured pressure, and sensors dotted every coastline. All that data—temperature, wind speed, humidity—flowed into powerful computers running complex numerical models. The result? A 5-day forecast today is more accurate than a one-day forecast was in the 1980s. A 3-day outlook hits 97% reliability. Even 7 days out, predictions hold at around 80%. Why? Because we layered real-time inputs onto historical patterns. The models learned, adapted, and got frighteningly precise.

Here is a visual of how that data flows—like satellites and radars feeding a central brain:

Election modeling follows the exact same logic—except it is invisible, and the stakes are our Republic itself.

Bad actors begin with Phase One: baseline predictive modeling. They pull from decades of records: turnout rates by county, party registration, historical swings. In a polarized country, party affiliation alone tells you a lot. Democrats vote Democrat. Republicans vote Republican. Down-ballot races follow suit—especially on divisive issues like guns, abortion, or taxes. A simple model based on party? Already decent. But they add confidence metrics: the presidential vote might score 95% certainty. A local bond issue? Maybe 70% if it is murky. This is old-school profiling—voter files, registration lists, past election data. It’s a solid foundation, like knowing that seasons change.

Next, and much deeper, comes voter profiling—the real sensors and radars of elections. Big data is a trillion-dollar industry because we hand over our lives in fine print. Every app, every card swipe, every search logs something. GPS tracks where you go: church on Sunday, gym at six, coffee shop downtown. Credit cards reveal purchases—books, guns, organic food. Amazon knows your tastes. Social media? Public posts show your stance on every ballot issue—climate, immigration, schools. Even private messages, if accessed, paint a fuller picture. Firms like Cambridge Analytica in 2016 scraped millions of Facebook profiles to map psychographics: fears, values, triggers. They called it microtargeting. It was voter dissection. Combine that with party? Your predicted ballot sharpens: "This independent hates regulation—92% for tax cuts." Accuracy climbs into the high 90’s, just like a 10-day weather forecast. One person, one profile, every choice scored with eerie precision.

See how all that data layers around you—like a web of connections tracking every move:

And another view—your digital footprint broken down into pieces big tech harvests:

That is Phase One: baseline plus profiling equals a near-perfect ghost ballot. No votes cast. Just data shadows.

Phase Two: Voter Snapshot Modeling. Voting begins. They do not need your actual choices—just confirmation you voted. Mail-in ballots? Tracking systems sold to you on the premise of notification when yours is printed, mailed, returned, accepted—real-time, precinct-linked. But that tracking is not for you. It’s for ‘them’. In-person? Electronic poll books—tablets at the door—log you instantly. Scan your ID, and turnout ticks up. Tie that to your profile: "Early mailer from this zip code—blue-leaning, high confidence." The model updates: "60% blue so far." Early voting drags this out—days of live previews. The feedback loop starts: predict, observe, refine. Like weather sensors feeding storm data as it builds.

Phase Three: Vote Leak Tuning. Machines turn guesses into facts. Direct Recording Electronic (DRE) terminals? You touch the screen—choices hit digital memory immediately, by vendor design. No subtlety; data broadcasts live. Ballot Marking Devices (BMDs)? They print paper, but the touchscreen knows every selection before ink touches. Software can log it covertly—buffered, uploaded, tied to precinct and ballot style. Voters believe "paper keeps it secret." Wrong. The computer has already tallied. Early tabulation? Lol, they’ve GOT IT ALL. The model tunes: "This candidate is down, need more injected votes. This measure is down, need more injected votes." Confidence peaks. Subtract real from desired: exact number of fakes required.

Here is what a DRE touchscreen looks like—your vote going straight digital:

And despite a BMD (Ballot Marking Device) that most DRE’s have been swapped out for, where you touch first, it prints a ballot, and we’re told that the ballot isn’t tabulated until later, what most people don’t realize is that the machine already knows the choices. Clever, aren’t they?

Phase Four: Injection and Cleanup. With a tuned target, they act—mules drop ballots in boxes, late mail-ins arrive, extras scatter across counties.

And since nobody at the county has any way of knowing who filled out a received mailed-ballot (or drop-box ballot), they have no way of actually verifying eligibility of the voter. How do you verify eligibility of a voter when you don’t have any way of knowing who the voter is? That’s the beauty of the rig they’ve built. They’re wicked-smart.

The final tally? Matches the prediction perfectly. Audits see clean numbers because the loop engineered them. Manipulation happens upfront—data feeds, profiles, leaks—all before counting. Officials notice nothing. They’ve then got most experts chasing glitches, not the system. That’s by design.

This is why audits fail. The fraud is front-loaded. No mismatched ballots. No suspicious patterns. Just engineered outcomes. That is the sophistication most miss—and why we cannot rely on surface fixes.

The solution is brutal simplicity: eliminate the preview. One-day voting only—at your precinct, on Election Day, with hand-marked paper ballots. No early voting. No late arrivals. Count them right there—bi-partisan teams, video-recorded, batches of twenty-five, witnesses watching every mark. Use translucent boxes, locked chains of custody. Electronics? A dumb scanner and simple inexpensive cameras only after polls close—for video and PDF images and records only, never for deciding winners. The key is NO ballots accepted post-close.

And observers right there—people you can see, not predictive feedback-loop-based algorithms you cannot:

That process exists. I built it. Watch how real, visible counting looks—no machines, no secrets, fully transparent at https://a-count-able.org : batch randomization, public tallies, full transparency.

And the complete ecosystem? Paper registration every two years, fourteen-hour national window, precincts capped at 1500 voters, even unlimited open public audits from ballot images and video footage. It is all at https://HandCountRoadShow.org/solution . Read it. No data pipelines. No leaks. No loops. Just human eyes on paper, in public, before a ballot even leaves the room.

I even co-authored a proposed executive order to address this current ELECTIONS NATIONAL STATE OF EMERGENCY: https://HandCountRoadShow.org/eo .

We do not need smarter machines. We need no secrets. Listen to the voices who understand the full depth—not only the fragments. And what do we do right now with only months left? Vote ON ELECTION DAY and IN-PERSON to deprive bad actors of their feedback loop. And vote in massive numbers, like your ballot matters, Because when we do, it finally will.