On Nimb, a theme is a labelled group of markets about the same underlying story, such as "Bitcoin & ETH" or "US midterms". Themes are drawn as dashed lines between markets and carry no money. They tell you where else a story trades; they do not show that those markets move together.
Themes help market reading by putting a single market back in its story. This page explains how they are built, how they differ from factors in finance, and, with the numbers from Nimb's own tests, what they can and cannot tell you.
What is a theme on Nimb?
A theme is a labelled group of markets about the same underlying story, such as Bitcoin & ETH or US midterms. Each market belongs to at most one theme. On the map, when you hover over or select a market, dashed lines join it to the other markets of its theme and the theme's name appears above them. Dashed lines are themes, never money: they have no width, no colour for Yes or No and no dollar value.
In the snapshot of 2026-11-07, 145 of the 179 events on the map belonged to one of 25 themes. The largest were Football (16 events), Iran (13), Stocks & IPOs (12) and Bitcoin & ETH (10). The other 34 matched no theme.
Is a theme a factor in the finance sense?
No. In finance, a factor is a common driver of returns that is estimated from price data, such as the market, interest rates or the oil price, and each asset gets a measured sensitivity to it (a beta). A factor is a statistical claim: these prices move together, by this much. A theme is a topical claim: these markets are about the same thing.
| Theme (Nimb today) | Factor (finance sense) | |
|---|---|---|
| Built from | Market titles | Price history |
| Says | "Same story" | "Moves with this driver, by this much" |
| Has a size or beta | No | Yes |
| Needs testing against noise | Only for correct labels | Yes, always |
Some parts of the app use the word "factor" for themes. In these docs, "theme" is the topical group and "factor" is reserved for a tested statistical driver.
How are themes built?
Each event's title and its first outcomes are checked against a reviewed list of 25 patterns, one per theme, such as words naming Bitcoin, Ethereum or MicroStrategy for "Bitcoin & ETH". The first pattern that matches wins. Sports patterns are checked first, so team names and "vs." never send a match into Politics. An event that matches nothing is "not part of a theme yet".
This is simple and transparent, and it has known failure modes. Nimb's own tests, run on the snapshot of 2026-11-07 with 228 events, found about 13 events mislabelled by first-match rules. Examples: a Japan 10-year "Bond" market landed in Film & TV, a GTA 6 "launch" market in Token launches, and a New Zealand election and a Press Secretary market in US midterms. The patterns have since been tightened (for instance, Film & TV now needs "Bond actor", and Press Secretary sits in US courts & cabinet), but any word list will mislabel some titles.
Can a theme tell me markets will move together?
Not on today's data. Nimb's tests checked whether markets in each theme moved together more than random groups of markets would:
| Test | Result |
|---|---|
| Themes tested | 28 |
| Themes whose markets moved together more than a random-basket baseline | 6 of 28 |
| Same test after cleaning returns (no forward-filled prices, no kick-off jumps) | 1 of 20 (Bitcoin & ETH) |
| Stability of each market's sensitivity to its theme, first half vs second half of the month (median correlation) | r = 0.32 |
Three reasons explain most of it. Thirty days of 12-hour price steps is too little history: with two markets, the first principal component explains at least 50% of their joint moves by construction, so "explains 60%" can be pure noise. Many prediction-market prices sit still for hours and then jump on news, and those jumps dominate the variance. And testing 28 themes at once without correcting for multiple tests produces some apparent winners by chance. A data-driven alternative, clustering markets by their price correlations, found about as many clusters as shuffled data did: 30 days of 12-hour data cannot support data-driven factors.
Nimb will not show "explains X%" for a theme unless the figure is statistically significant against a noise baseline. The current map shows no such figure.
What can a theme tell you?
- Where else the story trades. If you are reading a ceasefire market, the Iran theme shows the other markets that price the same conflict: dates, leaders, deals.
- Whether related prices are consistent. Markets in one theme often have logical links: a "by March" market should not be priced above a "by June" market on the same event. A theme puts them side by side so you can check.
- How much money a story draws across all its markets, by comparing circle sizes.
- Who holds across a story. Focus on a theme's markets and the wallet lines show which holders have positions in several of them.
What can a theme not tell you?
- That its markets will move together, or by how much.
- That holding several markets of one theme is diversified, or that holding opposite sides is a hedge. Shared topic is not shared risk.
- That a market outside a theme is unrelated. "Not part of a theme yet" means no pattern matched its title, nothing more.
- Anything about the right price.
For statistically tested links between pairs of markets, see moved together, moved opposite. Known gaps are collected in known limitations.
What would it take to call a theme a factor?
More and cleaner evidence than 30 days of prices can give: trade-based price history over months rather than weeks, without forward-filled prices; a test against a noise baseline with a correction for testing many themes at once; and measured sensitivity to outside prices such as Bitcoin, Treasury yields, oil or the S&P 500. Until a theme passes tests like these, Nimb treats it as a topic, not a driver.
How do I use themes on the map?
Open "Themes" in the dock. The panel reads "Themes · tick the ones you want on the map" and lists every theme with the number of markets it has in the categories you are showing. Tick or untick themes to filter the map; "All" and "None" select everything or nothing. The button then shows how many themes are on, for example "Themes · 3". The search bar also finds themes by name.
In a market panel, "How it connects" names the theme ("Part of Iran, a group of 13 markets on the same topic. Dashed lines on the map join them.") or says "This market is not part of a theme yet." Themes do not change which markets are on the map; that follows the rules in market selection. Definitions of theme, factor and correlation are in the glossary.
From reading to execution
Use a theme to see the whole story, then read each market on its own before you use the order ticket. Related markets are not a substitute for reading the one you trade. More on the map in how to read the Nimb map.
Last reviewed 2026-11-07 · Data from Polymarket · Not financial advice · Nimb is not affiliated with Polymarket.