The integrity build: how prediction markets assembled a surveillance stack in 18 months
In less than two years, US prediction markets have put together the kind of policing infrastructure that took traditional exchanges decades to construct. What stock markets slowly layered on after wave upon wave of scandal-monitoring engines, insider screens, league integrity deals, encrypted watchlists, whistleblower channels-has been compressed into an eighteen‑month sprint, under regulatory glare and rapid growth.
The result is a full integrity “stack” with proper names and defined responsibilities: Poirot, ProhiBet, HALO, league partnerships, academic forensics, and, most recently, employment disclosures for traders in sensitive markets. The industry that has long sold itself on radical openness is now quietly building a sophisticated machine for keeping people out.
From century-long drift to eighteen-month sprint
Legacy securities markets did not begin with modern surveillance. For more than a century, the New York Stock Exchange operated with informal norms, rudimentary rules, and almost no systematic detection of insider trading or manipulation. What exists today-pattern-detection engines, cross‑venue surveillance, automated alerts for spoofing and layering-was accumulated piecemeal, case by case, after scandals and crises forced each new layer.
Prediction markets have flipped that order. Instead of waiting for a defining meltdown, they are constructing their police force in advance, and in public. Platforms have done this while their volumes have been growing fast enough that controls implemented one quarter are strained by the time the next one arrives. Oversight is not just catching up; it is being refactored in real time.
Naming the new police: Poirot, HALO, ProhiBet, and more
Kalshi has been the most transparent about how its integrity stack works, and that blueprint effectively defines the emerging norm for the US prediction market sector.
At the core is Poirot, Kalshi’s in‑house surveillance engine. It continuously scans every order and trade, flagging unusual timing, suspiciously consistent win rates, and signs of coordinated accounts. The workflow follows a three‑stage sequence familiar from equity exchanges: detect, investigate, enforce. That framing is deliberate; the platform wants regulators to recognize something that looks like the infrastructure of a national securities exchange, not a casual betting app.
Around Poirot sit three external layers provided by specialized vendors and partners:
– Solidus Labs and HALO: Solidus, originally built for crypto trading venues, provides HALO, an AI‑driven surveillance platform tuned to detect wash trading, spoofing, layering, and other manipulations typical of thin electronic order books. The choice is revealing: prediction markets see their core risks not as classic sportsbook fixing, but as exchange‑style microstructure abuse.
– IC360 and ProhiBet: IC360, a Las Vegas-based integrity firm, services major professional leagues, the NCAA, regulators, and sportsbooks. Its sports‑specific role includes ProhiBet, an encrypted database of individuals barred from betting: athletes, coaches, referees, team staff, and certain league employees. Platforms can query this list to block accounts without exposing the underlying identities, a crucial privacy and liability shield.
– Wharton Forensic Analytics Lab: An academic partner contributes statistical methods honed on insider trading and accounting fraud. This gives platforms advanced detection techniques and, equally important, reputational insulation-the ability to say that their methods have been stress‑tested by independent academics, not just built in‑house.
On top of this technical architecture, platforms have added front‑end controls: whistleblower buttons on market pages that let users flag suspicious activity, internal escalation paths, and policy teams that decide when flagged patterns cross the threshold for enforcement.
The March 2026 pivot: from after‑the‑fact policing to pre‑trade blocking
For most of the industry’s life, the rules around insiders were written as prohibitions but enforced after the fact. Traders suspected of being athletes, league officials, campaign staffers, or other banned participants could be investigated post‑trade; accounts might be frozen, profits clawed back, and regulators notified-but only once the activity had already hit the order book.
In March 2026, that logic flipped. After months of work with IC360 and direct integration with at least one major league (the NHL), Kalshi moved to preemptive screening. Instead of investigating suspicious trades later, known insiders are now blocked before a single order reaches the market.
The change came in two major areas:
– Sports markets: Athletes, referees, league and team employees, and other covered individuals are now screened using ProhiBet and league‑supplied lists. When they attempt to trade on related markets, the system denies the order at the gate.
– Political markets: Earlier bans on sitting elected officials were extended to cover candidates trading on their own campaigns-an important, if obvious, step in reducing the appearance of “betting on inside information” about one’s own race.
Underneath this is a much more robust identity regime. Traders must supply names, addresses, and government identification before they can participate at meaningful scale. The vision of anonymous, frictionless political and economic wagering has been replaced by something much closer to a regulated brokerage onboarding flow.
The quiet cost: exclusion, friction, and liquidity
All of this integrity infrastructure carries a cost that platforms are reluctant to quantify publicly: every new layer of screening and surveillance introduces points of friction, and friction is the enemy of liquidity.
– When identity checks become stricter, casual participants drop off.
– When prohibited‑person lists cast a wider net, some legitimate but hard‑to‑verify users get caught in the dragnet.
– When suspicious accounts are frozen or limited, volumes in sensitive markets thin out, making them easier-not harder-to manipulate with relatively small orders.
This is the paradox prediction markets now face. Their credibility depends on keeping out people who have both privileged information and direct influence over outcomes. But their forecasting value depends on maximizing participation, especially from sophisticated traders who often sit close to the events being priced. Each integrity upgrade is, structurally, an exclusion upgrade.
The industry’s bet is that long‑term survival requires absorbing that short‑term cost. In a political climate where prediction markets are under active congressional scrutiny, platforms are signaling that they would rather sacrifice some liquidity than invite the kind of scandal that could feed a de facto ban.
How older markets were policed-and why that precedent matters
There is a precedent for this trade‑off. Stock and derivatives markets did not volunteer surveillance out of altruism. They built it because each major scandal-insider trading rings, manipulated IPOs, flash crashes-threatened the legitimacy of the entire asset class. Regulators responded with enforcement actions and more demanding rulebooks, and exchanges that could not demonstrate credible internal controls lost trust and order flow.
Prediction markets are trying to skip the scandal‑then‑crackdown phase and jump straight to “we are already policed.” The reference points are clear:
– Proprietary engines like Poirot echo the trader surveillance systems of large stock and futures exchanges.
– Vendors such as Solidus Labs mirror the crypto ecosystem’s pivot from unmonitored venues to compliance‑focused platforms.
– League partnerships via IC360 and ProhiBet parallel the integrity arrangements that emerged in the US sportsbook boom.
In practice, this means the vendors themselves are now writing the standard. When multiple platforms rely on the same surveillance providers, those providers’ definitions of suspicious behavior, insider status, and acceptable risk become the de facto rules of the sector.
The trade nobody wants to price
The title “trade nobody wants to price” captures the core tension: how much forecasting accuracy and market depth is the sector willing to sacrifice in exchange for political and regulatory safety?
Platforms rarely address it openly, but several uncomfortable realities follow from the integrity stack:
– Some of the best‑informed participants-players, staff, campaign insiders-are being systematically removed from participation. Markets will, by design, be pricing outcomes with less direct inside input.
– Cleanup operations are expensive. Running an in‑house engine, paying external vendors, supporting legal review, and investigating edge cases costs real money, which has to come from somewhere: wider spreads, higher fees, or constrained product offerings.
– Aggressive surveillance discourages the “gray‑zone” sophisticated trader who sits near sensitive information but is not technically an insider. These traders may self‑select out, further thinning the top layer of informed liquidity.
Yet the alternative-letting insiders trade freely and hoping scandals do not break publicly-is barely viable in a climate where every loss by a heavily favored team or surprise election outcome invites questions about integrity.
Where the volume goes when barriers rise
Raising barriers in one venue never eliminates demand; it merely pushes it somewhere else. As prediction markets add identity checks, prohibited‑person filters, and heavier monitoring, three displacement paths become more likely:
1. Unregulated offshore platforms
Participants who value anonymity and unrestricted access can migrate to offshore sites with minimal or no surveillance. These venues may attract precisely the traders US platforms are trying to exclude, concentrating risk where regulators have the least reach.
2. Private and informal markets
In tightly connected communities-political staff, sports insiders, industry specialists-private side bets, syndicates, and chat‑room agreements can partially replicate prediction‑market exposure without any formal venue. These markets are opaque and unmeasured, but they siphon off some of the most informed volume.
3. Shifting into adjacent assets
Instead of betting directly in a prediction market, insiders can seek exposure through correlated assets: stocks, tokens, or derivatives whose prices move with the event. Surveillance in those markets is mature but differently structured, and the detection of event‑linked trading remains uneven.
For prediction markets, this raises a strategic dilemma. Heavy policing raises their institutional credibility but can erode their edge as uniquely informative, broad‑based aggregators of belief. The more volume leaks out to darker venues, the less representative and robust their prices become.
Are surveillance vendors becoming the standard?
Because the sector is small and under scrutiny, there is strong pressure to converge on a recognizable, auditable pattern of controls. That effectively hands a shaping role to the integrity vendors:
– If Solidus HALO decides that a particular trading pattern constitutes manipulation, that definition does not just guide one platform; it influences enforcement norms across the industry.
– If ProhiBet expands its categories of prohibited persons-say, to cover broader classes of team affiliates or contractors-that expansion will ripple through every platform integrated with it.
– If academic partners establish specific thresholds for statistical suspicion, those thresholds may become sticky, even when edge cases suggest nuance.
This vendor‑driven standardization has upsides: regulators get predictable practices, and platforms can demonstrate adherence to third‑party benchmarks. The downside is that the industry’s rules can be shaped by a small number of private companies and labs whose incentives and risk tolerances are not always aligned with maximizing market openness or innovation.
Does more surveillance actually make prediction markets better?
“Better” depends on what prediction markets are supposed to maximize:
– If the goal is regulatory viability and political acceptability, then yes: more surveillance, clearer exclusions, and faster enforcement all make the markets safer in the eyes of policymakers. They reduce the chance that a single scandal leads to sweeping restrictions.
– If the goal is pure forecasting accuracy, the answer is more ambiguous. Barring insiders removes some of the most informed signals. Strict identity requirements deter marginal participants. Excessive fear of enforcement can make traders underreact to valid information that looks “too good,” muting price moves.
– If the goal is social legitimacy, integrity systems arguably help. The perception that markets are “fair” and not obviously tilted by insiders is central to public acceptance, especially in sensitive domains like elections and professional sports.
The emerging consensus in the industry seems to be that prediction markets cannot survive without a strong integrity posture, even if that means sacrificing some amount of theoretical accuracy. They are willing to trade a measure of informational purity for the right to exist at scale.
What participants should take from this
For traders and observers, several takeaways are clear:
– Expect more KYC, not less. Identity verification is now foundational to how platforms enforce prohibited‑person lists and insider screens. The days of casual, low‑friction accounts on regulated US venues are over.
– Assume that patterns, not just profits, are watched. High win rates, uncanny timing, or coordinated account behavior will eventually be flagged, even if no explicit rule has been broken yet.
– Understand that rules are evolving. As vendors refine their tools and regulators react to new edge cases, definitions of “insider,” “manipulation,” and “high‑risk market” will continue to change.
– Recognize the political layer. Congressional investigations and regulatory debates are not background noise; they are central drivers of why the integrity stack looks the way it does. Platforms are building not only for today’s rules, but for the scrutiny they expect tomorrow.
What to watch next
The integrity build is not finished. Several developments will determine how far and how fast the policing of prediction markets evolves:
– Whether more leagues and political bodies formalize direct data‑sharing or integrity partnerships.
– How courts and regulators classify different types of markets, and whether additional surveillance becomes a precondition for licensure.
– Whether new entrants try a deliberately lighter‑touch model and, if so, whether they gain or lose market share.
– How offshore and informal markets respond to the tightening of US‑based venues, and whether they create parallel ecosystems that challenge the regulated ones for informational primacy.
Prediction markets began as experiments in open information aggregation, flirting with the idea that anyone, anywhere, could put a price on the future. In the US, that vision is being re‑engineered into something more constrained: a tightly monitored, identity‑verified, rule‑bound environment that looks much less like a betting pool and much more like a conventional financial exchange.
The integrity stack-Poirot, ProhiBet, HALO, academic forensics, whistleblower portals, employer disclosures-is the architecture of that transformation. It is the new police force of a market that is still deciding how much openness it can afford.