Systems and methods for electronic trade order routing
The present application generally relates to electronic trading systems, and more specifically to systems and methods for electronic trade order routing. A routing algorithm may generate an execution style recommendation for incoming IG corporate orders. For example, the style recommendation may be shown as a new column in the trading application dashboard that suggests a course of action to traders for each order. In one implementation, automatic decision making may occur based on the style recommendation.
1 . A method of routing an incoming trade order at a routing server, the method comprising:
training a machine learning based classification model using historical patterns of trade costs from a training dataset of historical trade data to predict an execution style based on a trade cost;
in response to receiving, at a communication interface of the routing server, information of the incoming trade order,
transforming, by a processor at the routing server, the received information of the incoming trade order into a vector of attributes;
generating, by the trained machine learning based classification model, a predicted likelihood that the incoming trade order is to be executed under a specific execution style from a set of execution styles based on the vector of attributes;
determining, from the set of execution styles, a recommended execution style based on an estimate of trade cost associated with each execution style and the predicted likelihood corresponding to the specific execution style; and
routing, by the routing server, an electronic message comprising the incoming trade order to a destination trade execution system based on the recommended execution style, causing the incoming trade order to be executed at the destination trade execution system under the recommended execution style.
2 . The method of claim 1 , wherein the set of execution styles include an auto-execution style, a request for quote style, and a direct voice-activated style.
3 . The method of claim 1 , wherein the vector of attributes includes any combination of:
a side value,
an original trade size,
a price change in dollar terms in response to change in spread by a single basis point,
a wallet spread,
a coupon rate,
an amount issued, and
a rating.
4 . The method of claim 1 , wherein the predicted likelihood takes a form of a probability that is generated by the machine learning based classification model taking an input of the vector of attributes,
wherein the machine learning based classification model is pretrained on historical trade execution data.
5 . The method of claim 1 , wherein the trade cost associated with an execution style is determined based on an estimate of an implementation shortfall metric for the incoming trade order being executed under the respective execution style subject to a threshold requirement placed on the predicted likelihood.
6 . The method of claim 5 , wherein the estimate of the implementation shortfall metric is generated by a gradient boosting regression model corresponding to the respective execution style.
7 . The method of claim 5 , wherein the estimate of the implementation shortfall metric is generated in a form of a conditional mean implementation shortfall for the incoming trade order, and in a form of a prediction confidence interval having a lower bound and an upper bound.
8 . The method of claim 1 , further comprising:
filtering execution styles that correspond to respective probabilities no greater than a propensity threshold.
9 . The method of claim 1 , further comprising:
providing, via a user interface, the recommended execution style prior to transmitting the electronic message to the trading platform; or
automatically transmitting the electronic message to the destination trade execution system upon determining the recommended execution style.
10 . A system of routing an incoming trade order at a routing server, the system comprising:
a communication interface that receives information of the incoming trade order;
a memory storing a plurality of processor-executable instructions; and
a processor reading from the memory and executing the plurality of processor-executable instructions to perform operations comprising:
training a machine learning based classification model using historical patterns of trade costs from a training dataset of historical trade data to predict an execution style based on a trade cost;
transforming, by a processor at the routing server, the received information of the incoming trade order into a vector of attributes;
generating, by the trained machine learning based classification model trained using historical patterns of trade costs, a predicted likelihood that the incoming trade order is to be executed under a specific execution style from a set of execution styles based on the vector of attributes;
determining, from the set of execution styles, a recommended execution style based on an estimate of trade cost associated with each execution style and the predicted likelihood corresponding to the specific execution style; and
routing, by the routing server, an electronic message comprising the incoming trade order to a destination trade execution system based on the recommended execution style, causing the incoming trade order to be executed at the destination trade execution system under the recommended execution style.
11 . The system of claim 10 , wherein the set of execution styles include an auto-execution style, a request for quote style, and a direct voice-activated style.
12 . The system of claim 10 , wherein the vector of attributes includes any combination of:
a side value,
an original trade size,
a price change in dollar terms in response to change in spread by a single basis point,
a wallet spread,
a coupon rate,
an amount issued, and
a rating.
13 . The system of claim 10 , wherein the predicted likelihood takes a form of a probability that is generated by the neural network machine learning based classification model taking an input of the vector of attributes,
wherein the machine learning based classification model is pretrained on historical trade execution data.
14 . The system of claim 10 , wherein the trade cost associated with an execution style is determined based on an estimate of an implementation shortfall metric for the incoming trade order being executed under the respective execution style subject to a threshold requirement placed on the predicted likelihood.
15 . The system of claim 14 , wherein the estimate of the implementation shortfall metric is generated by a gradient boosting regression model corresponding to the respective execution style.
16 . The system of claim 14 , wherein the estimate of the implementation shortfall metric is generated in a form of a conditional mean implementation shortfall for the incoming trade order, and in a form of a prediction confidence interval having a lower bound and an upper bound.
17 . A non-transitory storage processor-readable medium storing a plurality of processor-executed instructions for routing an incoming trade order at a routing server, the instructions being executed by a processor to perform operations comprising:
training a machine learning based classification model using historical patterns of trade costs from a training dataset of historical trade data to predict an execution style based on a trade cost;
in response to receiving, at a communication interface of the routing server, information of the incoming trade order
transforming, by a processor at the routing server, the received information of the incoming trade order into a vector of attributes;
generating, by the trained machine learning based classification model, a predicted likelihood that the incoming trade order is to be executed under a specific execution style from a set of execution styles based on the vector of attributes;
determining, from the set of execution styles, a recommended execution style based on an estimate of trade cost associated with each execution style and the predicted likelihood corresponding to the specific execution style; and
routing, by the routing server, an electronic message comprising the incoming trade order to a destination trade execution system based on the recommended execution style, causing the incoming trade order to be executed at the destination trade execution system under the recommended execution style.
18 . The non-transitory storage processor-readable medium of claim 17 , wherein the set of execution styles include an auto-execution style, a request for quote style, and a direct voice-activated style.