Trading Technologies International, Inc.

Contextual searches may be performed to identify subsets of search results from an organized dataset. Contextual searching may return search results based on a relevance of the results to a query or search terms. Characters may be input by users at a search field that may be used to define context filters for limiting the search results of the organized dataset. Triggering events may be received that may trigger the definition of context filters that may be used to obtain the subsets of search results from the organized dataset. The triggering event may be a delimiter in the character string, for example. The subset of search results may include a subset of tradeable objects that may be traded at an electronic exchange. The subset of search results may continue to be limited based on additional context filters defined from the search input at the search field.

Trading tools may be manipulated by allowing assembly, disassembly, and configuration of different trading tools. The assembly of different trading tools may generate combined trading tools. The combined trading tools may enable interoperability between the trading tools therein. Customization of the combined trading tools may be enabled by enabling a user to rearrange predefined areas displayed within the trading tools within the combined trading tools. The customization of the combined trading tools may enable customization of different functionality within the combined trading tools. The trading tools within a combined trading tool may include a tab tool for enabling multiple instances of another trading tool to which the tab tool is attached. Each instance of a trading tool may be associated with a respective tab for enabling a user to toggle between the multiple instances of the trading tool.

Certain embodiments provide systems, apparatus, and methods to analyze incoming data messages and create market information constructs. An example method includes receiving a data message including an instruction to initiate a market event. The example method includes evaluating the instruction to determine whether it is associated with two or more causally linked market events. The example method also includes classifying the instruction based on the evaluating as part of a sequence of causally linked market events or as a single market event. The example method includes queuing the sequence of causally linked market events. The example method further includes detecting an end of the sequence of causally linked market events. The example method includes constructing a logically reduced market data message construct descriptive of the one or more market events represented by the queued sequence of causally linked events.

Various techniques are described for using machine-learning artificial intelligence to improve how trading data can be processed to detect improper trading behaviors such as trade spoofing. In an example embodiment, semi-supervised machine learning is applied to positively labeled and unlabeled training data to develop a classification model that distinguishes between trading behavior likely to qualify as trade spoofing and trading behavior not likely to qualify as trade spoofing. Also, clustering techniques can be employed to segment larger sets of training data and trading data into bursts of trading activities that are to be assessed for potential trade spoofing status.

Certain embodiments reduce the risks of traditionally programmed algorithms such as syntax errors, unclear logic, and the need for a non-trader programmer to develop the algorithm as specified by a trader by reducing or eliminating the writing of programming code by a user. Certain embodiments provide building block buttons and an algorithm area to define an algorithm. Certain embodiments provide live evaluation of an expression as the algorithm is being defined. Certain embodiments provide a design canvas area and blocks for designing an algorithm. Certain embodiments provide live feedback for blocks as the algorithm is being designed. Certain embodiments provide for initiating placement of an order to be managed by a selected user-defined trading algorithm from a value axis and for displaying working orders being managed by different user-defined trading algorithms on the value axis. Certain embodiments provide a ranking tool.

Methods and systems to prevent adverse exchange limit effects are disclosed. An example method of message management includes tracking, at a gateway, a transaction count of trade messages communicated to an exchange via an exchange link. The example method includes defining a first transaction limit based on the transaction count, and calculating a second transaction limit as a function of the first transaction limit, wherein the second transaction limit is to correspond to a reserve capacity associated with the exchange link. The example method includes assigning a message priority to each of a plurality of trade messages to communicate to the exchange. The example method includes communicating trade messages of the plurality of trade messages having a first message priority to the exchange when the transaction count is greater than the second transaction limit, and delaying or rejecting trade messages of the plurality of trade messages having a second message priority.

Systems and methods for optimizing order placement in an electronic trading environment are provided. Global queue holder orders are placed in a tradeable object to reserve quantity for traders. Traders may submit order requests for quantity at price levels at which a global queue holder order is holding quantity. When a trader submits an order for a quantity at a price level at which quantity is being held, a gateway compares the order to the global queue holder order information stored in a database. If the trader's order coincides with a price level at which quantity is held, then the gateway virtually transfers ownership of that quantity to the trader, who then assumes the advantageous queue position. When the quantity associated with the order is filled, the fill information is sent to the gateway, which to the database to determine which trader is associated with the filled order quantity.

Methods, system and articles of manufacture are disclosed including configurable order entry, matching, coordination and market data intervals. An example method to decouple order entry and matching of contra-side orders includes receiving a trade order during an order entry interval, wherein the order entry interval defines a first period. The example method also includes starting a matching interval at the expiration of the order entry interval, wherein the matching interval includes a second period and wherein the first period is contiguous to the second period. The example method further includes matching the trade order, during the matching interval, with one or more opposing tradable object trade orders received during the order entry interval, the trade order having a same price or same quantity as the one or more opposing trade orders.

A trade order may be submitted to an electronic exchange and updated, according to pay-up parameter values, in an attempt to receive a complete fill within a period of time. The trade order may comprise an order price and an order quantity when submitted to the electronic exchange. The pay-up parameters may include a pay-up interval, a pay-up amount, a pay-up counter, or a combination thereof. The price of the trade order may continue to be updated after the expiration of pay-up interval in an attempt to receive a complete fill of the trade order. The price of the trade order may be increased or decreased by the pay-up amount. The trade order may continue to be updated until a complete fill is received, the pay-up counter expires, and/or an ending time expires for which the pay-up parameters may be implemented.

Various techniques are described for using machine-learning artificial intelligence to improve how trading data can be processed to detect improper trading behaviors such as trade spoofing. In an example embodiment, semi-supervised machine learning is applied to positively labeled and unlabeled training data to develop a classification model that distinguishes between trading behavior likely to qualify as trade spoofing and trading behavior not likely to qualify as trade spoofing. Also, clustering techniques can be employed to segment larger sets of training data and trading data into bursts of trading activities that are to be assessed for potential trade spoofing status.