Research
Cross-venue pricing, implied probability, and how prices behave as an event approaches settlement.
Cogitate is a research lab for prediction markets. We measure how markets price uncertain events, test those prices against what happened, and build on what holds.
Forecasts were drawn by hand, wired down a line, shouted across a pit. What is new is the speed, and the chance to measure forecasts properly.




We study prediction markets and event trading: how markets put a price on whether something will happen, and whether that price holds. Every claim we make is checked against market data. Three kinds of work run side by side.
Cross-venue pricing, implied probability, and how prices behave as an event approaches settlement.
Data capture, matching and replay built to the millisecond, so a test sees the market as a trader would have.
What survives testing becomes trading systems and tools. Two are in development and will be announced here.
A gap between two prices for the same event looks like an edge. Many vanish once you pay to reach them. Switch the costs on and watch.
Each venue takes its cut on entry and exit.
Prices move while an order travels. We test in 100 ms steps.
Two venues can word the same event differently, or settle it on different sources.
Does the gap hold on history the rule never saw?
Will the S&P 500 close higher this week?An illustrative event, priced on two markets.
Each block is one cent of the quoted 7¢ gap. Figures are illustrative, not results.
Ideas that fail are recorded as dead ends, not quietly dropped. That record is most of the work.
| Event | Market A | Market B | Net gap | Verdict |
|---|---|---|---|---|
| S&P 500 closes higher this week | 0.62 | 0.55 | 3¢ | Holds |
| EUR/USD above 1.12 on Friday at 10am | 0.41 | 0.44 | 0¢ | Dead end |
| US CPI above 3% year on year | 0.27 | 0.22 | 2¢ | Holds |
| Gold settles higher this month | 0.73 | 0.66 | 0¢ | Dead end |
| Nasdaq-100 falls 2% or more this week | 0.18 | 0.24 | 4¢ | Holds |
Illustrative rows. Real research results are not published.
Four steps, in this order.
Capture prices and order books from the venues we study, with timestamps we can trust.
Pair the same event across venues, rule by rule, so two prices really compare.
Replay history with fees, latency and settlement, then test on data the idea never saw.
Turn what survives into systems and products, and keep measuring them live.
If one of these is you, we would like to hear from you.
Market structure, pricing and liquidity research, and integrations.
Joint research on event pricing and execution.
Feeds, history and reference data we can test against.
People who like being proven wrong by data. We are hiring.
Every message is read by the research team, not a sales inbox.
paul@cogitate.ai