Defined scope. Written terms. Evidence before conclusions. JOEY VICTORINO.
Field Note 027 / Political economy & systems research
What is Soulja Boy’s 2028 campaign likely to become—and how would we know?
PoliFi means politics mixed with crypto. The question here is simple: does a presidential headline turn into campaign work, paid opportunities, both—or neither? We use seven advanced tools to test those paths. Follow the everyday examples, then open the full math if you want it.
Start here / The campaign forecast in plain English
Working forecast · Low confidence in the 2028 oddsThe clearest thing Soulja Boy has built so far is a public conversation about 2028. Our working forecast is more headline-driven activity before a broad, state-by-state campaign. That is a forecast about visible activity, not a claim that his interest is fake or that he has a hidden plan.
Through 2028: Intermittent political publicity and possible commercial activity remain the working case; broad ballot access remains the harder, less supported branch.
The biggest reason we could be wrong: The election is still two years away. Candidates can build serious operations later, and an early absence of an indexed filing is weak evidence about what will happen by 2028. Paid work and a genuine campaign can also happen at the same time.
The labels below express our current lean. The numbers are the original analyst estimates, not results fitted from this research. None is a probability of winning the presidency.
Original rough bet: about 35 in 100
Filing is easier than building a national campaign, and he has publicly expressed interest. The bounded current search has no match, and there is no authenticated operational plan in this dossier.
Low: an analyst judgment, not a measured base rate.
Exact event, deadline and scoring rule ↓Original rough bet: about 8 in 100
This requires a route, staff, money, deadlines and state-by-state work beyond an announcement. This dossier has not established those capabilities for this campaign.
Low: the research establishes hurdles, not an empirically fitted eight-percent estimate.
Exact event, deadline and scoring rule ↓Original rough bet: about 65 in 100
There is a reported prior crypto-event performance and a history of public promotional work. A repeat opportunity is plausible, but the future calendar and fees are unknown.
Low: historical precedent supports plausibility, not a validated sixty-five-percent rate.
Exact event, deadline and scoring rule ↓Original rough bet: about 25 in 100
Token activity has been reported, but the contract-to-person mapping and compensation reconciliation required by this prediction have not been established here.
Low: public disclosure choices and future token activity are unmeasured.
Exact event, deadline and scoring rule ↓These are early checkpoints, not legal deadlines or claims about events that have already happened.
More public attention is easier to generate than a functioning national organization. Treat new posts and interviews as attention unless they are accompanied by identifiable campaign work.
What would make us upgrade the forecast: A matching official filing plus a named team would be a stronger sign of organization than another viral post.
Check: Look for an identity-matched candidate filing, a named treasurer or authorized committee, and a campaign website that explains who receives any donations.
The useful question becomes whether the interest repeats as real work, not whether one headline was popular.
What would make us upgrade the forecast: Repeated, documented campaign operations would move the assessment toward a durable campaign. A performance contract would support the separate commercial-activity forecast.
Check: Verify a recurring team, dated policy work, actual campaign events, and any required financial reports. Keep commercial bookings on a separate list.
Expect a clearer split between a continuing public persona and an organization capable of doing the paperwork and work of an election.
What would make us upgrade the forecast: A funded plan with named owners and state-specific requirements is evidence that the campaign can turn attention into election work.
Check: Look for a published, jurisdiction-specific ballot strategy, assigned state organizers, a budget, and a timetable. With nearly two years still available, missing these items remains an early warning, not a final verdict.
October reporting describes an exploratory 2028 message. The FEC candidate index returned no match for four tested 2028 name variants.
Read the exact search and its limits ↓Attention is established more clearly than a national organization. State ballot work needs money, people and time. The records reviewed do not establish that capacity.
Matched filings, named staff, reported campaign money and a funded state-by-state plan would move us toward a more serious ballot forecast.
The math can be calculated correctly while the forecast is wrong. Right now we can check the calculation, but we cannot claim a measured real-world hit rate.
There was a reported public 2028 signal. The tested FEC searches returned no match. Official SEC records document both the 2023 claim and its March 2026 dismissal.
More publicity before a broad ballot operation is our working call. A new team, fresh funding or new filings could change that picture.
We have not tested this forecasting system on a large, fair set of similar campaigns. We cannot honestly call it 90% accurate. The exact odds, earnings and vote share are not established.
Chance and confidence are different. A 65-in-100 bet says the event seems more likely than not. “Low confidence” says the evidence behind that bet is thin. It does not mean we studied 100 Soulja Boy campaigns.
What the advanced tools add: they let us account for limited resources, sudden headlines, noisy reports and changing evidence. They make assumptions easier to challenge. They do not supply missing facts, prove a motive or turn a guess into a measured rate.
How we will find out: keep the four dated predictions and rules, check what actually happens, and score the misses as well as the hits. We need many resolved forecasts and fair comparisons before claiming a reliable hit rate.
Select imagined new evidence. This changes the explanation only; it does not change our published probabilities or add a source to the ledger.
No new evidence selected. The working call stays at an early public signal.
A missing check means “not selected in this exercise,” not proof that the activity does not exist. These pieces overlap, so adding them does not justify multiplying odds as if they were independent.
Read each question, the small worked example, and what it means for the forecast. The example numbers teach the method; they are not Soulja Boy’s accounts. Open “Try the model” to change assumptions. Open “Inspect the advanced math” for the full derivation.
Two things can be true: a person can want public office and earn money from attention. A booking is not a vote. A missing search result is not proof of no campaign. We forecast observable steps, not a hidden motive.
October reporting supports an early presidential signal. It does not establish a staffed, funded campaign. The next useful clues are filings, named staff, spending records and state ballot plans.
What the advanced math does: Measure theory maps the possible activities. KKT asks where one more hour or dollar does the most good when time and money are limited.
Invented example: someone has 20 hours this week.
The plan needs more staff, more time, or a smaller scope. Attention alone does not supply those resources.
What this changes in the forecast: Watch whether public attention is followed by an organization. A real organizing footprint would strengthen the forecast of a sustained campaign.
How much to trust it: Low for his actual costs. The laws are documented; his campaign budget and workload are not.
To get closer to real life: Replace invented costs with dated budgets, filings and staff records.
What would change the call: Verified fundraising, experienced staff and a funded ballot plan.
Move the cursor across a boundary. The cost changes because the model assumes it changes—not because the map grants legal permission.
Drag the dot, or focus the map and use the arrow keys. Green means the example assigns fewer costs; red means it assigns more. A real legal boundary cannot be read from a circle on a chart.
What this chart means. Move the dot into red and the example becomes more costly. This shows why campaign paperwork, token-promotion rules and ballot work need separate budgets. The score is a teaching device, not a legal clearance or a campaign probability.
Political speech, campaign activity, securities promotion, and ballot access create overlapping obligations. A geometric map makes the scenario legible; it does not create a legal safe harbor.
Take a bounded measurable payoff Π, a finite-measure domain, and a positive density cap L̄. The cap prevents a formal optimizer from piling unlimited attention density onto an arbitrarily small point. Introduce the multiplier λ for the capacity constraint. Pointwise maximization of the Lagrangian gives the threshold allocation above; dual feasibility, primal feasibility, and complementary slackness complete the KKT conditions under a standard constraint qualification. The optimized value J is concave in capacity K, so λ is a supporting marginal value, and equals dJ/dK wherever that derivative exists. Neither the indicator premium nor the exponential cost formula follows from KKT. Their coefficients, monetary units, and relationship to legal risk must be separately estimated. The three colored circles are conceptual boundaries for candidate registration, paid securities touting, and ballot-access cost; actual law has neither Euclidean distances nor a single common coordinate system.
Attention is limited. Spend it where the return looks strongest, and budget for the rules that apply. The green patch shows a business assumption. It is not a lawyer’s green light. A campaign slogan does not cancel paperwork or turn a token promotion into a free pass.
Evidence & identification limit. An announcement does not establish an exemption from campaign-finance law. The federal candidate definition generally turns on contributions received or expenditures made exceeding $5,000, subject to testing-the-waters rules. Form 2 and Form 1 have different filers and deadlines. Financial disclosure is a separate regime. Ballot rules vary by jurisdiction. Political speech does not neutralize securities, fraud, or disclosure obligations.
The reported Crypto Ball appearance and earlier promotional work show possible commercial routes. They do not tell us what this presidential signal will earn.
What the advanced math does: Jump-diffusion models ordinary changes plus sudden shocks. Monte Carlo runs many possible paths so one dramatic result does not stand in for every outcome.
Invented example: 100,000 people see an offer.
The money depends on buyers and costs, not the headline count.
What this changes in the forecast: Commercial opportunities are plausible. Their size is unknown. The simulation explores what would happen if attention converts at different rates; it does not uncover hidden earnings.
How much to trust it: Low for dollar forecasts. No audited campaign-linked receipts or measured conversion rate are supplied.
To get closer to real life: Use unique reach, paying-customer counts, contracts and reconciled costs.
What would change the call: Documented new paid work tied to the campaign period.
Lower conversion and raise overhead. See how easily a lively audience can produce a weak cash result.
This is an invented money example, not a forecast of his earnings. The computer draws 60 possible 90-day paths using made-up income, costs and surprises. The 82–86% margin slider is a narrow assumption we were asked to test; it is not an estimate from his accounts. A setting of 0.18 is a model link, not a measured 18% buyer rate.
What this chart means. Thin lines show different imagined money paths; the dark line shows their middle result each day. A wider spread means more uncertainty inside this scenario. These paths do not measure how likely the real campaign is to succeed.
Separate earned attention from commercial receipts. Simulate how controversy, conversion, volatility, and jumps interact, while treating all cash-flow inputs as illustrative.
Assume independent Brownian drivers, an independent Poisson process, and i.i.d. finite-second-moment jump sizes. Substitute the correlated Brownian representation into the cash equation; summing covariance terms gives σ²C. Over a small time interval, draw one Gaussian increment and a Poisson jump count, then add one jump size for each arrival. Conditioning on the jump count and using the law of total variance gives λJ E[J²]Δ for compound-Poisson variance. This corrects the double-counting notation ΣJm dNt. A separate per-arrival commercial event model can be added, but its expected fees must not also be counted as fresh jumps without adjustment. Use percentile bands rather than a single path as the forecast summary.
A headline brings the crowd. A real product or booking collects the money. The sliders show what happens if more of that crowd pays. Free publicity can cut ad bills, but somebody still pays for staff, delivery, taxes, and mistakes.
Evidence & identification limit. Zero paid advertising can imply zero paid-media CAC when the denominator is positive. It does not imply zero total acquisition cost, guaranteed new customers, or zero overhead. The requested 82–86% pocketed margin is an unverified scenario range. Earned media, gross token receipts, realizable proceeds, and lawful net income are different quantities.
A viral day can create many stories from one post. Our campaign forecast should move more when new records appear than when the same story gets repeated.
What the advanced math does: KL compares probability profiles. Stop-gradient holds one comparison fixed during learning. Under strict conditions this can reduce training noise. Neither tool makes an unreliable source true.
Invented example: four news stories all quote the same announcement.
Count new evidence, not just more copies of old evidence.
What this changes in the forecast: Check for new filings, staff, money and state work at each checkpoint. This keeps the forecast tied to progress. The widget is a demonstration, not a trained campaign predictor.
How much to trust it: Unknown forecast accuracy. There is no tested teacher model or campaign backtest here. A 98.6% motive probability is unsupported.
To get closer to real life: Freeze predictions, collect independent outcomes and score them later.
What would change the call: Repeated accurate predictions on cases the model had not already seen.
Switch headline noise on and off. A calmer line is easier to read; it is not automatically more accurate.
Pick a reference chance, then compare a second forecast with it. Both are made up for this lesson. Agreement tells us how close the two sets of numbers are. It does not tell us whether either forecast is right about the campaign.
What this chart means. Dark is the assumed reference, green is the test forecast, and red is a mathematical lower bound on their agreement. Turning off smoothing adds pretend daily shocks. Even high agreement is not a measured accuracy rate for Soulja Boy’s campaign.
| Time ahead | Reference / test forecast | Difference score (KL) | Agreement (α) | Minimum agreement bound |
|---|
A conditional-expectation teacher can reduce gradient variance under stated assumptions. That machine-learning result does not establish the probability of a campaign’s commercial purpose.
For cross-entropy with softmax logits, the derivative with respect to logits is p−eY. Add and subtract q and apply the Jacobian chain rule to obtain the gradient decomposition. Suppose p, q, and J are measurable with respect to the same conditioning sigma-algebra G and q is the true conditional class probability. Then the residual has conditional mean zero, and the covariance cross-terms vanish. The law of total covariance gives Loewner dominance. Detaching the teacher with stop-gradient prevents derivatives through q; it does not make an inaccurate teacher correct. The probability-vector overlap equals one minus total variation, and Pinsker bounds that overlap from below in natural-log units. A 0.986 overlap requires a sufficiently small KL value; no measured teacher, KL value, training experiment, or validated motive labels are provided here. The table is a scenario smoother, not a trained self-distillation model.
One loud day can fool you. Check the pattern after 30, 60, and 90 days: what got filed, what got built, and what got paid? A smooth chart helps you see the pattern. It cannot read somebody’s mind or turn a guess into 98.6 percent certainty.
Evidence & identification limit. Loewner dominance can fail with a biased teacher or mismatched conditioning. Daily headline shocks are not automatically a martingale. Forecasts out to Q4 2028 require an explicit outcome model and out-of-sample calibration; the requested 98.6% / <1.4% case split is not established by this derivation.
Reporting says payment mattered to his explanation of the Crypto Ball booking. That supports a narrow commercial point. It does not establish his entire politics or campaign strategy.
What the advanced math does: HJB compares choices today while counting their possible future benefits and costs. It asks what choice works best under specified assumptions.
Invented example: two otherwise similar bookings.
Compare the whole deal, including future costs—not just the advertised fee.
What this changes in the forecast: Paid appearances remain a plausible next step. A past paid booking does not show whether a serious campaign will also develop. Both can happen.
How much to trust it: Moderate for that booking; low for future choices. The performance and payment-focused remarks were reported. Future contracts are unknown. The SEC dismissed its pending claim in March 2026; that does not prove political protection.
To get closer to real life: Use actual terms and dated costs across several decisions.
What would change the call: Verified new contracts or a different pattern of decisions.
Change the assumed cash return. The default gives affiliations the same hazard; a party label is not evidence of regulatory protection.
This is an invented money example. We compare three choices using income, waiting costs and possible legal losses. Every choice gets the same legal-risk assumption. A higher bar means a better result under those assumptions, not proof of what he will choose.
What this chart means. The green bar wins this money example. Change the assumed income and another choice may win. Political names label the scenarios; the bars do not measure political beliefs, regulatory favors or the real campaign’s future.
An HJB model can compare hypothetical commercial strategies. A public statement about compensation supports a narrow revealed-preference inference, not a solved control problem or a reduced enforcement hazard.
Write the dynamic-programming recursion for a short interval Δ, including discounting and an independent terminal hazard with zero recovery. Apply the diffusion generator and the jump expectation to the continuation value. Expand e−δΔ to first order, cancel V, divide by Δ, and take the limit to obtain the HJB equation. A well-defined solution also needs state dynamics, a domain, boundary or terminal conditions, admissible controls, regularity, and estimated parameters. If cash flow and hazard are constant and state dependence disappears, integrating π exp[−(δ+Λ)t] over the infinite horizon gives π/(δ+Λ). That is a closed-form special case; the public quote does not supply its parameter values or solve the full equation.
He said the event paid him. That tells you money mattered for that booking. It does not tell you the full deal, who controlled the money, or what he truly believes. Price the job on its terms. Do not price imaginary protection into it.
Evidence & identification limit. The January 2025 quote indicates that payment mattered to his stated decision. It does not establish an endorsement contract, identify the payer, prove all political preferences, or show that a Trump-associated event reduced SEC risk. No political enforcement discount is inferred; the simulator holds hazard equal across affiliations unless the user changes an explicitly hypothetical input.
Follower totals do not count unique people or paying customers. The supplied 61.7% token concentration claim is not verified insider ownership, and it is not an audience share.
What the advanced math does: Network topology separates different groups. The dwell-time model tests how demand might fade while an offer is delayed. Market depth limits what can actually be sold.
Invented example: 1,000 distinct followers.
A small paying group can matter more than a large watching group. These group sizes are assumptions.
What this changes in the forecast: A huge audience can support attention without supporting a broad campaign or large sales. Look for measured donors, volunteers and buyers rather than treating followers as all three.
How much to trust it: Low for audience-to-cash estimates. Unique reach, buyer groups, pool depth and wallet ownership are unverified.
To get closer to real life: Measure separate groups and actual sales; identify contracts and block dates.
What would change the call: Verified turnout, repeat buyers, donors or sustained volunteer activity.
Increase the delay. Watch the assumed demand fall. The rate of that fall has not been measured for this campaign.
This is an invented money example. Waiting longer can let interest fade before a paid offer appears. The audience mix is assumed: 85 out of 100 watch, 14 make smaller purchases, and 1 makes larger purchases. The 61.7% token setting is unverified and is not a count of people or proof of common ownership.
What this chart means. Dark shows all imagined receipts, green larger buyers and gray smaller buyers. Red marks your chosen delay. The curve shows the cost of waiting under the example’s assumptions, not his observed conversion rate or future earnings.
Three audience segments show why raw follower counts are poor substitutes for paying demand. Latency and capacity constrain any conversion model.
Choose disjoint segments so the same modeled payer is counted once. Multiply transaction velocity, average ticket, an explicitly defined dimensionless multiplier, and take rate to obtain a monetary rate. Multiplying that rate by an effective duration produces a dollar quantity. The exponential dwell function is flat before τ0 and has slope −γT afterward; doubling the post-threshold delay multiplies retention by another exponential factor. Sum across nodes or segment representatives with explicit segment counts. Finally cap the estimate by executable market depth and actual delivery capacity. Leverage can increase notional exposure but does not mechanically increase cash receipts; a conservative scenario sets that multiplier to one.
A million viewers do not mean a million buyers. A few big buyers may matter more than the crowd, but a big token balance is not cash in hand. If people lose interest before the offer arrives, sales fall. Test that with real receipts, not follower math.
Evidence & identification limit. The 85/14/1 segmentation is invented for sensitivity analysis, not measured audience data. Social followers overlap across platforms and may include inactive accounts. A 61.7% wallet-cluster concentration claim is not a 61.7% audience share, a count of independent whales, or proof of beneficial ownership. Contract, chain, block height, supply basis, and attribution method are required.
An announcement is one clue. A verified filing, paid booking or state campaign team is another. We should weigh what each record establishes, not call every clue proof of one hidden plan.
What the advanced math does: SPRT updates an evidence score one observation at a time. A separate payment gate checks whether agreed commercial conditions are met. Those are different tasks.
Invented booking rule: a $25,000 payment is required before delivery.
That rule protects a proposed booking. It says nothing about a person's sincerity or legal clearance.
What this changes in the forecast: A verified organizing milestone should strengthen the campaign case. A commercial booking supports commercial activity. Neither cancels the other, and repeated coverage is not fresh evidence.
How much to trust it: Low for the current score. The likelihoods and minimum fee are hypothetical; no real PAC wire history is supplied.
To get closer to real life: Estimate weights from comparable cases and track independent records.
What would change the call: A sustained sequence of verified campaign milestones.
Add a filing checkpoint, then a payment checkpoint. The score moves because the widget assigns invented evidence weights.
This invented booking example needs two things: payment at least equal to the fee, and an attention score of at least 40. Those are pretend house rules, not his known terms. The evidence buttons compare two made-up stories: commercial activity and ballot work. They do not update the real campaign forecast.
What this chart means. The line moves up when a pretend record favors commercial activity and down when it favors ballot work. It stops at a chosen evidence threshold. Repeated clicks are practice observations, not new research, and the result is not a verdict about his motives.
No observations yet.
A deterministic payment rule and a sequential evidence test solve different problems. Keep the business gate separate from the hypothesis score.
Define H at zero explicitly so the payment and reach thresholds have unambiguous equality behavior. Its product unlocks a simulated commercial workflow only when both thresholds pass. Independently, Wald’s test accumulates the log likelihood ratio of observed events under two fully specified hypotheses. Likelihood products become sums of logs; the thresholds approximate target type-I and type-II errors, with possible overshoot and model-dependent operating characteristics. A candidate filing may weigh toward one hypothesis while a documented commercial booking weighs toward another, but those probabilities must be estimated. Correlated repeats of the same news item cannot be counted as independent evidence. The posterior formula additionally requires a prior and the same likelihood assumptions.
Get the scope and price in writing. Confirm the money before the agreed work starts. That is a business rule. Separately, keep a scorecard of real evidence. One payment can explain one job; it cannot explain every choice a person makes.
Evidence & identification limit. All simulator event likelihoods are illustrative. The hypotheses may coexist: a person can pursue commercial activity and a political campaign. A threshold crossing is not a finding of deception. No evidence supplied establishes a real minimum fee, refusal of unpaid Democratic requests, PAC wire history, or a hardware-locked endorsement system. Unlocking is not legal clearance.
Public filings, authenticated contracts and correctly dated chain records can test the story. The present dossier lacks a complete negative FEC audit and verified ownership of the claimed wallet cluster.
What the advanced math does: Counters measure changes. Persistence tracks repeated observations. Cash-flow accounting subtracts obligations. Merkle hashes let another reader check that saved records have the same bytes.
Two invented examples; neither describes his actual accounts.
A busy counter is not a sales total. Upfront money is not all free cash.
What this changes in the forecast: Upgrade the forecast when records can be independently checked. Keep missing information marked unknown rather than turning it into zero activity.
How much to trust it: High for byte checks; low for missing real-world inputs. A valid hash checks consistency, not truth, ownership or future accuracy.
To get closer to real life: Publish dated queries, full contract addresses, block numbers and reconciled receipts.
What would change the call: Independent records that confirm—or contradict—the forecast.
Change the counter dates or inventory haircut. Then inspect a source and verify its inclusion proof.
The numbers here are invented. Subtracting two account counters can count outgoing transactions; it cannot count buyers, sales or profit. The wallet chart uses 24 made-up addresses, not identified insiders. The money example starts with $200,000 in tokens, then adjusts for sale losses, fees, an advance, costs and work still owed.
What this chart means. Bars count invented wallets by how long they keep tokens; the red line is your chosen cutoff. More persistent wallets need not mean more people. None of these settings measures the real campaign’s finances or ownership.
Public filings and chain records can support reproducible observations. A snapshot, wallet attribution, and enterprise valuation each need their own evidence.
The proposed sequential-count expression simplifies algebraically to the observed counter difference and is undefined at zero before simplification. For one monotone EVM account nonce, the difference can count consumed transaction nonces between correctly aligned snapshots; it does not measure users, profitable trades, token swaps, all chains, or FEC filings. Solana does not provide an equivalent universal EVM account nonce. The actual German-tank estimator assumes a simple random sample without replacement from consecutive labels 1 through N; its unbiased maximum-based estimator is m(n+1)/n−1. That sampling model is absent from ordinary chain observations. A persistence indicator counts labeled observations; cluster aggregation and address-to-owner uncertainty must be handled separately. Finally, token quantity times spot price marks an inventory, while free cash flow deducts costs and obligations and valuation discounts dated cash flows. Upfront receipts produce useful float only to the extent those funds are available after restrictions and promised delivery.
Count what you can show. If the counter rose by 20, say 20; do not dress it up as a secret formula. If a wallet holds tokens, do not call the whole balance spendable cash. Money paid up front still comes with work, bills, and sometimes a refund.
Evidence & identification limit. No private access or live feeds are used. The brief supplies no verified token addresses, block snapshots, filing-search export, or ownership graph. Missing records are reported as missing observations, not zero activity. Commercial prepayments, token sale proceeds, and political funds cannot be presumed unrestricted or cost-free. Zero ballot spending is an unverified assumption.
The internet can rent you a crowd. A crowd is not a cap table, a bank balance, or a ballot. Here is what the seven models mean when the math leaves the room.
A slogan can attract attention. It cannot cancel campaign rules or securities law. Put the cost of doing things correctly in the budget. A green patch on a chart is an assumption, not permission.
A headline brings people in. A booking or product may bring money in. Track the conversion and every cost. Free publicity can lower the ad bill; it does not guarantee an 86 percent margin.
At 30, 60, and 90 days, check filings, staff, contracts, and receipts. A quieter chart may help you think. It cannot tell you why a person made a choice.
The Crypto Ball statement says payment mattered for that appearance. It does not prove the price, the payer, an endorsement agreement, or protection from regulators. Do not pay extra for imaginary cover.
Some people watch. Some buy. Some write bigger checks. Find out who actually pays without counting the same person twice. A token balance is worth only what can really be sold after costs.
A payment milestone can protect a real booking. It cannot buy legal clearance or turn a model score into evidence. Keep political activity, commercial work, and token promotion on the correct books.
Save the filing query, contract address, chain, date, and block number. Say what each record shows and what it cannot show. A digital fingerprint proves the saved bytes match; it does not prove the story is true.
A proposed commercial appearance term sheet, not an endorsement-buying mechanism. No message has been sent. The economics below are negotiating terms, not claims about Way’s rates.
These four estimates were first published on 08 November 2026. We keep them unchanged so we cannot move the goalposts. They are rough analyst judgments, not measured success rates or probabilities learned by these models. “35%” means a bet of about 35 chances in 100—not “35% of the work is done.”
The score in everyday math: suppose we say a coin has a 50% chance of heads. Write 50% as 0.5. If heads happens, the score is (0.5 − 1) × (0.5 − 1) = 0.25. If we had said 90% and heads happens, it is (0.9 − 1) × (0.9 − 1) = 0.01. Lower is better. A confident wrong call gets a larger penalty. We score only events with a documented outcome.
| Observable event & resolution rule | Our original rough estimate | Outcome / evidence URL or note | Brier loss |
|---|---|---|---|
| F01 · A 2028 presidential Form 2 is filed for DeAndre Cortez Way.Deadline: 2028-11-07 (UTC)Resolve Yes on an FEC filing image or official record matched to identity and office. Otherwise No after deadline and a logged search of aliases and committee records. | 35%About 35 in 100 · untested judgment | — | |
| F02 · Way appears as a presidential candidate on general-election ballots in at least 10 U.S. states.Deadline: 2028-11-07 (UTC)Count state-certified printed ballot listings; exclude write-in eligibility, territories, DC, primary ballots, and withdrawn listings. Ten distinct states resolves Yes. | 8%About 8 in 100 · untested judgment | — | |
| F03 · At least one new paid appearance at a crypto-industry event is documented.Deadline: 2028-11-07 (UTC)Event must occur after 08 November 2026. Require an organizer or management record that identifies both the event and compensation; attendance or a photograph alone is insufficient. | 65%About 65 in 100 · untested judgment | — | |
| F04 · An issuer-authenticated creator-token contract has a public, independently reconciled compensation disclosure linking it to Way.Deadline: 2028-11-30 (UTC)Require full chain and contract or mint address, issuer/management authentication, compensation terms, and a matching on-chain or financial record. A ticker, anonymous cluster, or apology alone does not qualify. | 25%About 25 in 100 · untested judgment | — |
Mark “Yes” only on documentary evidence. Mark “No” after the stated deadline and a documented search. Unresolved or unsupported resolutions do not enter the score. New evidence can correct an outcome without rewriting the original probability. This browser-local ledger is editable, not an immutable third-party forecast registry.
The source list below says what each record supports and what it does not. “Primary” means an original official record. “Reported” means we are relying on a publisher’s account. “Unverified” means we have not established the claim. A coin’s name alone does not prove who launched or controls it.
| Record / status | Evidence, observation & remaining uncertainty |
|---|---|
| L00 primary identity | SEC charging release identifies DeAndre Cortez Way (Soulja Boy) ↗Event: 2023-03-22 · Checked: 2026-11-08 · primaryDeAndre Cortez Way is identified by the SEC as Soulja Boy.Limit: Identity only; allegations are distinct from findings. |
| L01 reported audience | Chartmetric public artist profile ↗Event: 2026-11-08 · Checked: 2026-11-08 · secondaryPublic search snapshot reports approximately 7.1M Instagram and 4.9M X followers. The supplied baseline includes 3.6M TikTok.Limit: No time-stamped platform-native capture; TikTok exact 3.6M not verified. Followers overlap and do not equal reach, voters, or buyers. |
| L02 reported signal | Soulja Boy Hints at Running for President in 2028 ↗Event: 2026-10-10/2026-10-12 · Checked: 2026-11-08 · secondaryOctober 2026 coverage describes an exploratory Instagram post discussing economic opportunity, criminal justice reform, education, and technology/AI.Limit: Sources differ on Oct10 vs Oct11. Exploratory signal is not proof of registered candidacy, a full campaign, or monetization intent. He is not the first rapper to seek the presidency. |
| L03 unverified fec-baseline | FEC candidate search starting point; supplied absence claimEvent: 2026-11-08 · Checked: 2026-11-08 · user-suppliedUser-supplied baseline asserts zero Form1 and zero Form2 filings 25+ days after the post.Limit: No complete filing audit was performed. Web-index searches for Soulja Boy/DeAndre Way found no relevant hit but cannot prove absence. Need all name variants, committee names, raw filings, image records, query parameters, time and pagination. No filings does not prove less than $5k raised or spent. |
| L04 primary fec-rule | FEC candidate registration ↗Event: 2026-11-08 · Checked: 2026-11-08 · primaryCandidate registration generally follows contributions received or expenditures made in excess of $5,000; Form2 within15 days and designated principal committee Form1 within10 days.Limit: Actual campaign activity and testing-the-waters rules matter; zero filings is not regulatory permission. More than $5k, rather than at least $5k. Personal financial disclosure has separate timing and rules. |
| L05 primary fec-waters | FEC testing the waters ↗Event: 2026-11-08 · Checked: 2026-11-08 · primaryGenuine feasibility testing may avoid registration even above $5k until candidacy/campaigning begins; funds count retroactively if candidacy follows.Limit: Testing-the-waters funds remain subject to applicable contribution limits and prohibitions. Public campaigning can end the exemption. |
| L06 reported crypto-ball | Reuters report on the inauguration-weekend Crypto Ball ↗Event: 2025-01-17 · Checked: 2026-11-08 · secondaryReuters reported Soulja Boy in the entertainment lineup at the January17 2025 Washington Crypto Ball.Limit: Event affiliation does not show who paid the artist or establish a PAC payment, endorsement contract, political commitment, or protection from enforcement. |
| L07 reported quote | AllHipHop report of Instagram Live response ↗Event: 2025-01-19 · Checked: 2026-11-08 · secondaryReported short excerpt: They paid me a bag. Coverage describes him emphasizing payment when defending the appearance.Limit: The longer user-supplied quote is a cleaned composite/ellipsis excerpt, not a certified verbatim transcript. Coverage also reports he said Trump was absent and later felt misled about politics. Payment amount and payer not independently established. |
| L08 reported promotions | NFTgators coverage of ZachXBT research ↗Event: 2023-04-19 · Checked: 2026-11-08 · secondaryZachXBT was reported to have counted73 crypto promotions and16 NFT drops and estimated more than $730k from promotional fees.Limit: Counts concern promotions, not necessarily73 unique projects. Attributed investigator estimates and allegations, not audited receipts or adjudicated fraud findings. Repeated reporting is not independent corroboration. |
| L09 reported rate-card | Benzinga report on promotional tweet with compensation terms ↗Event: 2021-05 · Checked: 2026-11-08 · secondaryA deleted promotional tweet reportedly exposed a contingent $24k fee if the project raised $240k.Limit: Reported screenshot/deleted tweet; not direct proof that the contingent fee was paid. A contingent deal is distinct from the separate $10k/X and $12k/Instagram rate-card estimate. |
| L10 primary sec-charge | SEC March2023 crypto touting charges ↗Event: 2023-03-22 · Checked: 2026-11-08 · primarySEC charged Way and other celebrities for undisclosed compensation for TRX and/or BTT promotion.Limit: Historical charge, not proof of every alleged act. Way did not join six celebrity settlements announced in that release. The pending claim was voluntarily dismissed in March2026. |
| L11 primary sec-dismissal | SEC dismissal of pending Section17(b) claim against Way ↗Event: 2026-03-05 · Checked: 2026-11-08 · primarySEC announced a voluntary dismissal of its pending Section17(b) claim against Way as part of the broader resolution.Limit: SEC expressly calls dismissal discretionary; it creates no general safe harbor and establishes no causal relationship to his Crypto Ball appearance or politics. |
| L12 primary sec-notice | Notice of Voluntary Dismissal as to Defendant Way, docket97 ↗Event: 2026-03-05 · Checked: 2026-11-08 · primaryThe SEC notice voluntarily dismisses claims alleged against DeAndre Cortez Way under Rule41(a)(1)(A)(i).Limit: This document does not state that participation in political activity immunizes paid token promotion. |
| L13 primary sec-statute | 15USC77q(b), Section17(b) ↗Event: 2026-11-08 · Checked: 2026-11-08 · primaryThe paid securities publicity provision concerns receipt or prospective receipt of issuer/underwriter/dealer consideration and disclosure of that consideration and its amount.Limit: Applicability depends on the asset and facts; disclosure alone does not cure other securities-law issues. First Amendment political speech is no blanket exemption. |
| L14 reported base-token | Base token launch and apology coverage ↗Event: 2025-12-15 · Checked: 2026-11-08 · secondaryDecember2025 reporting describes a Base SOULJABOY token and an X apology about previous paid promotions and diligence.Limit: Primary X links were blocked/empty on retrieval. No independently verified creator-linked contract registry or current economics. Reported launch date was the preceding Friday. |
| L15 reported solana-holdings | LBank Soulja Coin SBOY holdings table ↗Event: 2026-11-08 · Checked: 2026-11-08 · secondaryFetched Solana SBOY holder table listed61.92% for shortened address5Q544f...pge4j1; search snapshot showed63.01%.Limit: No block height, full mint, wallet attribution or cluster method verified. Could reflect pool, custody or other structure. A top token account is not an identified insider cluster, and supply share is not cohort share or realized cash. User61.7% kept only as scenario assumption. |
| L16 reported ballot-aggregate | Ballot access overview with historical nationwide estimate ↗Event: 2016 · Checked: 2026-11-08 · secondaryBallotpedia cites an estimate above860k signatures for an independent presidential candidate nationwide in2016.Limit: Cannot turn this into a universal800k-plus2028 statutory requirement. Route, party status, states, deadlines and changing laws matter; many paths differ. |
| L17 primary ballot-texas | Texas Secretary of State presidential ballot guidance ↗Event: 2026-11-08 · Checked: 2026-11-08 · primaryTexas guidance currently lists113,887 signatures for an independent presidential candidate in2028 with other restrictions and electors requirements.Limit: Texas-specific official guidance; does not establish50-state total. |
| L18 primary brand | JoeyVictorino current public theme stylesheet ↗Event: 2026-11-08 · Checked: 2026-11-08 · primaryCurrent site uses ivory#F7F3EA, forest#1B3A30, muted#4F6459, gold#A58D65 and thin lines. Fonts stylesheet identifies Instrument Serif, Instrument Sans and JetBrains Mono.Limit: Inspected public HTML/CSS through curl. Root brand remains applicable even if dossier chooses dark counterpart. |
| L19 unverified baseline-audience | User-supplied baseline · audienceEvent: 2026-11-08 · Checked: 2026-11-08 · user-suppliedTarget Entity: DeAndre Cortez Way ("Soulja Boy Tell 'Em"), operating across "The Soulja Boy Net" (7.1M Instagram, 4.9M X, 3.6M TikTok followers).Limit: Audience numbers supplied by user; partial corroboration in audience record, no unique-reach inference. |
| L20 unverified baseline-signal | User-supplied baseline · signalEvent: 2026-11-08 · Checked: 2026-11-08 · user-suppliedThe 2028 Exploratory Signal (Oct 10–12, 2026): Announced 2028 U.S. Presidential run on Instagram citing four pillars: economic opportunity, criminal justice reform, education, and the future of technology/AI.Limit: Exploratory signal reported; formal candidacy not established. See signal record. |
| L21 unverified baseline-fec | User-supplied baseline · fecEvent: 2026-11-08 · Checked: 2026-11-08 · user-suppliedThe FEC TOCTOU Gap (Nov 2026 Audit): Zero Form 1 (Statement of Organization) and zero Form 2 (Statement of Candidacy) filings in the FEC database 25+ days post-announcement. He remains strictly below—or unrecorded against—the $5,000 federal campaign finance statutory trigger.Limit: Supplied assertion. No completed database audit or spending/receipts verification. No actual FEC query result is claimed. Absence of filings cannot prove below-threshold status. |
| L22 unverified baseline-quote | User-supplied baseline · quoteEvent: 2026-11-08 · Checked: 2026-11-08 · user-suppliedThe Trump-Crypto Revealed Preference (Jan 2025): Performed at the inaugural Crypto Ball in Washington, D.C. ahead of Donald Trump's inauguration. On Instagram Live, flashing stacks of $100 bills, he stated verbatim: "They paid me a bag... Obama ain't never put no money in my pockets. Kamala ain't never put no money in my pockets. Trump put money in my pockets... It was the Crypto Ball for crypto, and I'll go again."Limit: Supplied excerpt preserved exactly. Published transcript has editing/profanity differences, so do not call it independently certified verbatim. Payment focus reported; payer, fee, formal endorsement and SEC causation not established. |
| L23 unverified baseline-promotions | User-supplied baseline · promotionsEvent: 2026-11-08 · Checked: 2026-11-08 · user-suppliedSEC Enforcement & On-Chain Provenance (2021–2026): Documented by ZachXBT promoting 73 crypto projects and 16 NFT drops (~$730,000 estimated take), including accidentally tweeting his own promotional rate-card copy ($24,000 fee if a project raised $240,000). Charged by the SEC in March 2023 (alongside Justin Sun) under Section 17(b) of the Securities Act for undisclosed touting of TRX and BTT.Limit: Supplied assertion; corroboration is attributed research counting promotions (not necessarily unique projects), estimated take and contingent fee. Historical charge is verified, but pending claim voluntarily dismissed March5 2026; see sec-dismissal and sec-notice. |
| L24 unverified baseline-tokens | User-supplied baseline · tokensEvent: 2026-11-08 · Checked: 2026-11-08 · user-suppliedRe-entered Web3 in Dec 2025–2026 after a public X apology, launching $SOULJABOY on Coinbase's Base L2 and SBOY (Soulja Coin) on Solana, where on-chain cluster telemetry flags >61.7% of supply in a single top wallet cluster.Limit: Supplied assertion. Base launch/apology reported; Solana identity and creator attribution not independently established. Exchange top token-account table is not an insider-cluster analysis. Full mint, block height, wallets, supply denominator, LP/custody labels and clustering method missing. |
| L25 primary fec-bounded-query | OpenFEC bounded candidate-index query audit, 8 November 2026 ↗Event: 2026-11-08 · Checked: 2026-11-08 · primaryFour 2028 presidential candidate-index name queries (SOULJA; DEANDRE CORTEZ WAY; DEANDRE WAY; WAY, DEANDRE) returned HTTP200, exact count0 and no results. Broader presidential DEANDRE and all-office SOULJA queries also returned count0. An all-cycle presidential WAY query returned62 entries on one page; none matched DeAndre Cortez Way. A positive control for TRUMP, presidential office,2024 returned candidateP80001571.Limit: A bounded candidate-index search, not a complete audit of raw Form1/Form2 filings, committee aliases, unindexed images, ingestion delays or campaign money. Name matching and publication lag can miss records. The live query link may change after collection. No inference of less than $5,000 raised/spent, regulatory permission, or certainty about a 2028 outcome. Raw response snapshots retained in the working research directory; the public summary should avoid publishing unrelated candidates' personal fields. |
Think of each record as a numbered receipt. We give each receipt a digital fingerprint, then combine the fingerprints into one final fingerprint. Change one word and it no longer matches. This can catch changed bytes; it cannot make a false claim true. The full SHA-256 tree and proof checker follow.
Every displayed ledger entry is committed in display order. Leaves hash 0x00 || UTF8(canonicalJSON(entry)); parent nodes hash 0x01 || left || right. Recursively split at the largest power of two smaller than the subtree size. No duplicate-last padding. Canonical JSON sorts object keys lexicographically, preserves array order, and has no whitespace (dataset: strings, numbers and null). Specification §2.1 ↗
A hash proves consistency with these bytes. It does not authenticate a source, establish truth, timestamp publication, or show a public append-only transparency service. No signed tree head or externally witnessed consistency proof is claimed.
Select a node above.
Selected leaf proofs recompute the root from canonical leaf bytes and ordered sibling hashes. Internal nodes reveal their ranges and child commitments. The source links open original publishers; raw remote page snapshots are not bundled.
Authenticated filings, disclosed contracts, independently reconciled payments, state-certified ballot records, and issuer-verified chain addresses would change the analysis. A proper experiment could estimate conversion or forecast calibration. Until then, the appropriate output is a bounded scenario—not a mathematical verdict on a person.