Advanced model with high accuracy predicts the levels of pending orders accumulation.
Model widget appearance.
One of the features of the derivatives market is the ability to open leveraged or margin positions. Any such margin position will have a liquidation price. As soon as the price of the underlying asset reaches the level of the liquidation price, the exchange will forcibly close the trader's position at a loss in order to prevent the client's balance from going into the negative zone.
Single liquidations are not of any interest, as their impact on the market is insignificant. However, the accumulation of such orders can significantly affect the price and provoke a jump in volatility. The accumulation of pending orders is otherwise called a margin zone.
Ordinary users do not have access to information about them - data of such a high level is available only to exchanges (and usually to some affiliated persons conducting trading activities related to cryptocurrency assets of the exchange). Modern quantitative models in the field of market microstructure allow calculating margin zones with a fairly high accuracy.
Based on the data analyzed using a dynamic neural network, we identified regularities in the placement of margin zones taking into account the specific behavior of a particular asset and built an appropriate model that depicts clusters of pending orders.
An important part of our miscalculation algorithm is data obsolescence, which allows us to level out "hung" clusters.
To add the "Predict Liquidations" tool to your workspace, first click on the "Widget menu" button in the platform header. Then click on the button of the same name in the "Microstructure" section.
The widget is a histogram in which:
Predicted location and volume of pending orders for SHORT positions.
Predicted location and volume of pending orders for LONG positions.
In the lower part of the tool there is a scale of the volume of forecasted pending orders in dollars.