This data quality wasn't available to retail investors before.
In theory, you could look everything up on Kalshi yourself. In practice, the data sits scattered across thousands of individual markets, with no comparison, no weighting, no link to your portfolio. This page explains where our data comes from, how we prepare it — and why the result has a different quality than any DIY research.
Official Kalshi data license
We source our market data directly from Kalshi, the prediction exchange operated under the oversight of the US regulator CFTC. We hold an official Kalshi data license.
That's not a legal formality. It means three things:
Completeness
We see the entire market universe, not the slice a public website happens to display.
Reliability
No scrapes that break with every layout change. No data gaps because a request got blocked.
Legal certainty for you
You receive licensed data from a provider entitled to use it. With services that scrape market data without a license, that's not a given.
Why Kalshi, not Polymarket
Kalshi and Polymarket together account for the overwhelming majority of the prediction market. For an analysis service aimed at stock analysis, Kalshi is the sturdier foundation.
| Kalshi | Polymarket | |
|---|---|---|
| Oversight | Under oversight of the US federal regulator CFTC | Originally set up offshore, crypto-native |
| Settlement currency | US dollar | Stablecoin |
| Market structure | Exchange-like, usable institutionally | Crypto trading environment |
| Contract definition | Formally specified, clear resolution criteria | Less consistent |
| Data access | Licensable | No comparable commercial licensing |
What follows from that: when a contract is formally defined and the resolution criteria are unambiguous, the quote is interpretable. You know exactly what the 74 percent refers to. With inconsistently worded contracts, you don't — and a number whose reference point is unclear isn't fit as an analysis foundation.
Kalshi settling in US dollars also removes an additional source of error: the quote contains no cryptocurrency volatility that would first need to be stripped out.
Filtering: a rulebook, not AI
Every day we screen over 10,000 markets. Ten make it into the report. Over 99.9 percent fall out – and that selection is exactly the work you're paying for.
How markets get excluded
Sports, entertainment, celebrity, curiosities: out, regardless of trading volume. Markets without a relevant move: out, regardless of topic. Contracts mapping the same event with a different threshold: scored together, so one topic doesn't take up five slots. And at most two markets per topic channel.
How markets get assigned
Every remaining market is assigned to a topic area — monetary policy, inflation and the labor market, commodities, geopolitics, fiscal policy and regulation, elections, technology and AI, companies and deals, crypto. The assignment follows fixed rules based on contract characteristics, not a language model's judgment.
Why this is the better way
A language model can't guarantee that. The same question asked twice can yield two different answers, and neither can be traced back.
If a market shows up in the report, there's a rule that explains why. If one is missing, the same applies.
A rulebook can misclassify a market. It can't invent a market that doesn't exist, or claim a connection that isn't there.
For data your analysis rests on, that's not a technical detail. Without reproducibility, you never know whether a change in the report comes from the markets or from the system.
The scoring
The biggest percentage move isn't the strongest signal. If selection were sorted purely by move size, contracts expiring on a routine schedule would regularly top the list — they swing mechanically toward 0 or 100 percent once a question resolves, without any new expectation behind it.
Our score therefore weighs five factors against each other. What fits into five lines here is the plain-language summary – behind it lie half a year of development and a rulebook with many thresholds, weightings, and exceptions:
over 24 hours and seven days.
how directly the event carries over to listed companies.
how much is traded.
whether the move carries new information or just reflects the approaching expiry.
whether there is news on the event right now.
The loudest market doesn't win — the most meaningful one does.
Where AI contributes — and where it doesn't
We say this openly because it helps you judge us.
Data retrieval, categorization, move detection, consolidation of related contracts, scoring, selecting markets for the report. All rule-based and reproducible.
The report's text: summary, research for each market, the other side, and equity transmission. The report states this explicitly.
What that means: a verifiable rulebook decides which markets you see and in what order. The research and the chain of effect to the stock are drafted with AI assistance and – like any AI-assisted text – can contain errors. That's why we give a source for every statement.
Most providers in this space let a model handle the entire process, from selection to text, without saying so. We separate the layers and name them.
What you'd have to do yourself
The comparison is meant seriously. Nothing about our work is impossible in principle — it's just not doable on a daily basis.
| Work step | DIY research | With Tradejectory |
|---|---|---|
| Find every relevant market | Scan thousands of individual markets, no filter for stock relevance | done |
| Detect moves | Note prior-day values and compare manually | done |
| Merge related contracts | Reconcile multiple thresholds per event by hand | done |
| Weigh relevance | Gut feeling | score |
| Check the contract's rules | Read every contract's fine print | done |
| Research news and data | Dozens of sources per market | done, with sources |
| Find counterarguments | Rarely done | done |
| Find affected stocks and ETFs | Derive the chain of effect yourself | done, with the weakest link |
| Prepare a readable result | – | finished report |
| Daily time spent | Hours – without the full picture | 5 minutes |
And even then, you'd have no comparison to the prior day, no consistency over weeks, and no certainty you hadn't missed anything. An AI subscription won't do it for you either: a chatbot doesn't see every market, doesn't compare against the prior day, and won't return the same selection twice for the same question. It took us half a year of development before our system could do that reliably.
Why now: the market behind the data
Prediction markets were a niche phenomenon for a long time. That has changed — which is why this data is fit for analysis today.
In December 2025, Kalshi and Polymarket together traded nearly $12 billion, over 400 percent more than the same month a year earlier.
Rising open interest and tighter bid-ask spreads mean order-book depth is approaching the microstructure of traditional derivatives markets. Prices now reflect a broad range of opinion instead of a small speculative fringe.
Hedge funds and asset managers use prediction data to refine probabilistic models — especially around Federal Reserve decisions. Macro research teams integrate event probabilities into their scenario analyses.
“Analysts are increasingly able to trace how shifts in prediction-market odds translate into expected spread movements, equity factor weightings, or sector rotations.”
That's exactly what our market analysis delivers. The difference: institutional houses employ teams for this. You get it as a finished report.
Brokers like Robinhood, Coinbase, and Webull have added event contracts to their offerings. The infrastructure is moving into everyday finance.
Prediction quotes don't replace an established metric. They add a dimension implied option volatility, futures curves, and survey-based expectations don't provide: a continuously traded probability for a concretely defined event.
Source: Demetrios Berdousis, “Prediction Markets on the Rise: How Financial Markets Are Adapting,” IBISWorld Analyst Insights, April 8, 2026. Volume figures there cite Piper Sandler.
An analysis service, not a betting operator
We get asked this, so we answer it fully here.
Tradejectory doesn't broker contracts, execute trades, or accept money for bets. You don't need a Kalshi account. You never stake a cent with us.
We analyze publicly traded, licensed quotes as a data source — the way a research house analyzes options prices or futures curves. What you get from us is a daily report, online and as PDF, and an analysis interface.
We also deliberately avoid everything prediction-market platforms use to drive engagement: no leaderboards, no streaks, no gamification, no social feeds. Those are mechanics for increasing trading activity. Our product is meant to save you time, not increase your screen time.
Independent and sober
No price targets. No hot tips. No hype.
We stay away from price-target marketing and the signals scene. We deliver data and context — the analysis belongs to you.
Behind Tradejectory is a team of analysts and developers who deliberately stay behind the product. We'd rather convince you with the daily report than with faces.
Judge the data quality for yourself.
Read the sample report online or try the first week free. Cancel anytime, even during the trial week.