Crowding

Crowding is the central idea in intenthybrid AI. It is a measure of how one-sided and how levered the positioning in a market has become. A crowded book is a fragile book, because everyone leaning the same way means there is little left to push the trade further and a large pool of forced sellers if it reverses.

What the crowding index represents

intenthybrid AI combines positioning signals into a single crowding index per market, expressed as a level that moves as positioning shifts. A higher reading means the crowd is more concentrated on one side and more exposed to a flush. The index is directional: a book can be crowded long or crowded short.

In a walkthrough you might see a market read as crowding at 82 percent, long-skewed, with positive funding. That illustrative reading says three things at once: the crowd is heavily on the long side, that concentration is high, and longs are paying funding to hold, which tends to mark a stretched position.

Why a single index

Each underlying signal is incomplete on its own. Funding hints at skew but not size. Open interest hints at size but not direction. Liquidation flow shows what is already breaking but not how much fuel remains. Folding them into one crowding index gives you a single number to watch per market, and a consistent way to compare one market against another.

How crowded is too crowded

There is no universal threshold. What counts as dangerously crowded depends on the market, its typical positioning, and how fast it moves. This is one reason the risk read matters: the model frames the crowding level in context rather than handing you a bare number, and you set your own thresholds for alerts and agents based on the markets you actually trade.

Crowding over time

Crowding is most useful as a moving picture. A market that has been climbing in crowding for hours while funding stays positive is building risk, even if price is calm. The crowding and liquidations manager is designed to surface exactly this, sorted by where a squeeze is forming.