Five separately bounded political records
Political Evidence Register
The record already says something.
Five case assessments expose the strongest result in the frozen record, the decision it changes, the inference it does not support and the next official comparison. The underlying workpapers stay controlled; the public findings, denominators and primary-source links do not.
Dated source entries across the five cases
Frozen public-record objects mapped to the ledger
Bounded findings produced from more than one record
Five public assessments
What the record shows—and what changes because of it.
Every assessment pairs the signal with its decision use, next official test, inference boundary and primary-source path.
ASSESSMENT 01
Open case status
PRE-ELECTION ASSESSMENT · Primary pending — August 18, 2026
One contest. Two entirely different financial machines.
Daniel Kowal's non-candidate receipts were 124.83 times Daniel Zegarac's. Zegarac's campaign was 98.77% candidate-originated versus 34.81% for Kowal. The material distinction is outside financial breadth versus candidate dependence—not gross receipts alone.
124.83×non-candidate receipt ratioKowal relative to Zegarac
Decision readA serious field brief should test money origin, administrator overlap, vendor dependence and the incumbent's exact motions separately. A single fundraising total erases the useful contrast.
Next official testFreeze the certified August 18 primary result, then compare the fixed money-origin and spend-concentration findings without changing their definitions.
Inference boundary
Financial structure is not voter support. The record does not measure volunteers, nonfinancial networks, persuasion or turnout cause.
ASSESSMENT 02
Open case status
PRE-ELECTION ASSESSMENT · Primary pending — August 18, 2026
Casey Askar's real profile is concentration at every layer.
The current committee was 97.36% candidate-loan funded, and two vendor clusters received 97.84% of operating spend. At the same cutoff, 62.43% of active District 22 Republicans were outside Collier County. Capital scale, execution concentration and district geography point in the same direction: the war chest is large, but the operating system is narrow.
97.84%operating spend in two clusters$1,095,212.80 at the cutoff
62.43%district Republicans outside Collier113,676 of 182,079 active registrations
Decision readThe useful questions are whether the campaign can diversify execution, convert beyond its Collier base and show broader financial participation before the disclosure clock closes—not whether it can quote a large receipt total.
Next official testAfter the August 18 primary, compare fixed finance rank with certified result rank and Collier-versus-rest-of-district performance. Preserve the known one-candidate finance gap.
Inference boundary
Self-financing and vendor concentration do not establish reach, ownership, coordination, service quality, wrongdoing or electoral strength. Registration is not turnout or support.
ASSESSMENT 03
Open case status
SINGLE-CANDIDATE ASSESSMENT · Primary pending — August 18, 2026
Byron Donalds combines mass small-entry volume with concentrated dollars and counterparties.
Transactions of $50 or less accounted for 81.93% of positive contribution rows but only 14.36% of contributed dollars. Response America received 47.19% of reported spending, while the Republican Party of Florida supplied 97.38% of reported in-kind value. The accurate map has three layers: transaction volume, cash vendors and in-kind counterparties.
81.93%rows at $50 or less80,436 rows; 14.36% of dollars
47.19%spending to Response America$2,305,072.34 reported
97.38%in-kind value from RPOF$1,201,081.39 reported
Decision readSmall-entry volume is real, but it should not be used to imply dispersed dollar power. Vendor and party concentration create distinct operational dependencies that require separate review.
Next official testAfter the August 18 primary, compare the frozen Donalds record with the official result while keeping the eleven-candidate field gap visible. Do not manufacture a full-field model after the fact.
Inference boundary
Rows are transactions, not unique donors. Payment and in-kind concentration do not establish ownership, control, coordination, effectiveness, violation or misconduct.
ASSESSMENT 04
Open case status
RETROSPECTIVE DIAGNOSTIC · Retrospective result comparison complete
The simplest money rule failed—and the winning bloc fractured by motion.
A top-three receipts rule selected none of the three winners. Across all eight candidates, receipts and vote selections had a descriptive Pearson correlation of -0.449. In six selected early split votes, agreement within the elected trio ranged from 20% to 80%. The result rejects both a money-equals-rank shortcut and a permanent-winner-bloc shortcut.
0 / 3winners selected by top receiptssimple retrospective rule
r = −0.449receipts-to-vote correlationeight-candidate field
20–80%agreement within elected triosix selected split votes
Decision readFuture Naples work must combine vote mode, precinct geography, money origin and exact governing motion. Receipts alone did not explain this result, and election cohort alone did not explain the early coalition record.
Next official testPreregister the next municipal comparison before results and expand the coalition matrix beyond the six selected split votes.
Inference boundary
Eight candidates and six selected motions cannot establish a general electoral law, a causal money effect, ideology, motive or durable alliance.
ASSESSMENT 05
Open case status
TRANSITION ASSESSMENT · General election pending — November 3, 2026
The headline election is more governance transition than five-race competition.
Five of seven contested candidates reported zero monetary receipts at the cutoff. Unopposed candidates accounted for 91.89% of reported expenditures and 64.65% of receipts. Separately, the Authority's recognized-operator count rose from 16 to 41 between Q1 and Q2 while reported voluntary-curfew compliance moved from 98.4% to 98.0%. The political file must join seat structure to the operating record, not treat raw fundraising as a complete viability map.
5 / 7contested candidates with zero receipts71.4% at the cutoff
91.89%spending by unopposed candidates64.65% of reported receipts
16 to 41recognized operators, Q1 to Q2compliance moved 98.4% to 98.0%
Decision readThe sharper brief identifies which seats are truly contested, where institutional continuity survives the elected-board transition and which operating issues carry comparable denominators.
Next official testFreeze qualification and certified November 3 results, keep unopposed seats outside contested calculations and compare Seats 4 and 5 on fixed finance and operating definitions.
Inference boundary
Zero recorded receipts is not zero campaign activity. Raw operator counts are not rates, and the record does not establish normalized performance, support or viability.
Portfolio status
Nine ready. Fifteen bounded. Fourteen blocked.
These are frozen candidate-record counts, not certified ballot counts. Field completeness and case depth are different measures; a deep file on one subject does not make every record in that case usable.
The candidate-specific public record supports a bounded principal brief with named limitations. It does not imply a client, endorsement or electoral advantage.
The record supports a field comparison, retrospective diagnostic or provisional brief only inside a named evidence gap.
A material identity, filing, result, vote, source or other proof requirement is missing. No candidate-specific conclusion is issued.
38candidate records
9brief ready
15bounded useful
14blocked
Metric clarity
Every headline number carries its denominator.
The count and its non-claim travel together. This prevents a workflow measure from being mistaken for evidence, accuracy or electoral performance.
208 / 208 hash checks
MeansTwo hundred three raw public-record objects plus five case-control files reproduce their recorded hashes.
Does not meanIt is not 208 validated claims and does not prove every assertion inside an underlying document.
123 source rows
MeansThe five case ledgers contain 123 dated source entries with case, record and limitation context.
Does not meanA row is not automatically a fact, finding or preserved underlying document.
35 multi-record findings
MeansThirty-five bounded findings join more than one frozen record and retain an inference limit.
Does not meanThey are not allegations, predictions or proof of motive, influence, coordination or misconduct.
1 retrospective diagnostic; 4 comparisons pending
MeansNaples Council has an official result comparison. Four pre-result evidence packages await their defined official result comparisons.
Does not meanNo registered win probability or prospective prediction model has been scored.
9 ready; 15 bounded; 14 blocked
MeansNine of 38 frozen candidate records can support a bounded brief now, 15 support a narrower use and 14 remain blocked.
Does not meanIt is not a certified ballot-field count, candidate ranking, favorability measure, evidence grade or forecast.
Operational readiness index
MeansA transparent workflow judgment combining source integrity, finding yield, analytical breadth, validation state and event proximity.
Does not meanIt is not a probability, risk score, accuracy score, field-completeness rate or evidence grade.
Five defined case fields
Depth is shown beside completeness.
The operational readiness index rewards integrity, finding yield, breadth, validation state and event proximity. It is not a probability or field-completeness rate.
Frozen case file · 2 candidate records
Collier County Commission District 4 — 2026 Primary
76.1operational readiness
2 brief ready
0 bounded
0 blocked
Primary pending — August 18, 2026Pre-result evidence package; no outcome model registered
Strongest current useComplete local money, vote, governing-record and decision-clock join for the two-candidate field.
Binding limitThe incumbent governing sample contains five material votes; challenger biography and prior-policy claims retain bounded gaps.
Frozen case file · 7 candidate records
U.S. House District 22 — 2026 Republican Primary
86.2operational readiness
1 brief ready
5 bounded
1 blocked
Primary pending — August 18, 2026Pre-result evidence package; explanatory calibration only
Strongest current useCurrent and prior identity reconciliation joined to field money, geography and the disclosure clock.
Binding limitOne qualified candidate lacks a located federal finance record; nonincumbents do not carry a comparable governing record.
Frozen case file · 11 candidate records
Florida Governor — 2026 Republican Primary
91.8operational readiness
1 brief ready
1 bounded
9 blocked
Primary pending — August 18, 2026Deep single-candidate file; field prediction unvalidated
Strongest current useTransaction-scale state finance joined to public funds, filings, legislation and selected federal votes.
Binding limitThe case does not contain an eleven-candidate field model or a complete outside-committee and independent-expenditure universe.
Frozen case file · 8 candidate records
Naples City Council — 2026 Backtest and Coalition Sample
90.0operational readiness
3 brief ready
5 bounded
0 blocked
Retrospective result comparison completeBacktested diagnostic; not prospective validation
Strongest current useCertified results falsify a simple top-receipts rule and a unified-winner-bloc assumption.
Binding limitThe design is retrospective and the coalition record contains six selected split votes, not a preregistered forecast.
Frozen case file · 10 candidate records
Naples Airport Authority — 2026 General Election
89.0operational readiness
2 brief ready
4 bounded
4 blocked
General election pending — November 3, 2026Transition calibration pending; no general electoral model
Strongest current useLaw-to-seat transition joined to institutional roles, candidate money and operating records.
Binding limitLow-dollar field records and candidate qualification affirmations are not independently complete; no result exists yet.
Validation protocol
Result comparison is not automatically prediction.
One retrospective diagnostic is complete. Four pre-result evidence packages await official result comparison. None contains a registered win probability or outcome forecast, so later work scores correspondence, reproducibility and explanatory usefulness—not prediction accuracy.
-
complete
Naples City Council — 2026 Backtest and Coalition Sample
Comparison opens 2026-02-04. Eight candidates and six selected votes cannot establish a general electoral rule.
-
pending
Collier County Commission District 4 — 2026 Primary
Comparison opens 2026-08-19. A two-candidate result cannot validate a general money model.
-
pending
U.S. House District 22 — 2026 Republican Primary
Comparison opens 2026-08-19. Finance, registration and geography remain separate from voter support and turnout cause.
-
pending
Florida Governor — 2026 Republican Primary
Comparison opens 2026-08-19. A deep single-candidate file does not become an eleven-candidate field model.
-
pending
Naples Airport Authority — 2026 General Election
Comparison opens 2026-11-04. Unopposed seats remain outside contested calculations; certified identities and results are required.
Pattern library
Nine repeatable detection methods.
Each method keeps source class, time state and inference boundary attached. The value is not the graphic. It is the false conclusion the method prevents.
Money origin, not money total
Candidate money, outside individuals, committees, public funds, in-kind activity, transfers, refunds, spending and distributions remain separate.
- Test
- Candidate-originated share, outside breadth, cash and debt structure, and money-class reconciliation.
- Do not infer
- Support, influence, solvency or outcome.
Vendor and administrator concentration
Repeated vendors, addresses, treasurers and operating clusters while preserving legal distinctions.
- Test
- Top-one and top-two shares, address clusters and repeated role states.
- Do not infer
- Ownership, control, coordination, service quality or misconduct.
Disclosure-clock gap
What was officially knowable before voting, filing, hearing and decision dates.
- Test
- Coverage end, report due, notice window, voting start and result comparison date.
- Do not infer
- Concealment, noncompliance or voter knowledge.
Geography and vote-mode conversion
Where registration, county, precinct, mail, early and Election Day records describe different parts of the field.
- Test
- County share, mode ratio, precinct rank reversal and fixed field share.
- Do not infer
- Individual targeting, demographic persuasion, turnout cause or forecast.
Motion-specific coalition
The exact motion, amendment, authority, mover, seconder, abstention, recusal and eligible member vector.
- Test
- Pairwise agreement by shared eligible votes and issue-class matrix.
- Do not infer
- Ideology, friendship, coordination, motive or durable alliance.
Authority transition and continuity
How law, election status, seat rules, current roles and prior institutional service change formal power.
- Test
- Identity changes, prior roles, contested status and transition date.
- Do not infer
- Endorsement, incumbency advantage, certified result or policy continuity.
Statement-versus-record test
Whether a public statement is corroborated, partial, unresolved, future-testable or rejected by the controlling record.
- Test
- Bind the statement to the exact finance, vote, bill, employment, service or institutional record.
- Do not infer
- An unresolved statement is not automatically false.
Capital and payment chain
Budget, award, amendment, authorization, invoice, payment, balance, milestone, output and outcome as separate states.
- Test
- Compatible paid-to-authorized comparisons only when the record supports them.
- Do not infer
- Nonpayment, overrun, completion or missing private capital.
Entity resolution and false-positive suppression
Stable identities across people, committees, bills, vendors, organizations, roles, addresses and time states.
- Test
- Collision checks, current/prior separation and exact-name versus resolved-entity edges.
- Do not infer
- Same name or address is not proof of one legal person, ownership or control.
Public boundary
Named findings are public. Private work product remains controlled.
The assessments on this page are derived from frozen public records and retain their sources, denominators and non-claims. Candidate-specific workpapers, client material and nonpublic campaign data remain outside the register. Candidate, PAC, party and independent-expenditure work stays separately scoped. Intelligence Desk publishes no voter file, persuasion model, legal conclusion, influence claim, outcome guarantee or beneficial-owner guess.
The assessments and counts derive from a frozen August 9 command workbook built from five public-record case systems at the August 7 evidence cutoff. Source pages may later change; the public register states the captured cutoff and does not expose private work product.Private briefing
Bring one question. See the evidence boundary first.
The first conversation identifies the jurisdiction, record classes, decision date, missing proof and separation requirements before any engagement is proposed.