Granted Patent B2

US 8,660,925 · App. 13/423,768 · Granted Feb 25, 2014

Abstract

Claims (54)

1. A system for constructing an order book from displayed market data for a tradable asset, said system comprising one or more computer processors configured to:

measure an effective spread of the tradable asset from displayed market data;

measure a mid-quote volatility of the tradable asset from displayed market data;

measure additions between best bid and ask of the tradable asset from displayed market data;

measure additions less cancellations of the tradable asset from displayed market data;

calculate, with the one or more computer processors, a probability of a hidden order for the tradable asset as a function of the measured effective spread, the mid-quote volatility, additions between best bid and ask, and additions less cancellations;

calculate, with the one or more computer processors, a hidden order volume between the best bid and ask;

calculate, with the one or more computer processors a hidden order price; and

construct an order book for the tradable asset that includes displayed order volume and hidden order volume based on said calculated probability of a hidden order, said calculated hidden order volume, and said calculated hidden order price.

2. The system in accordance with claim 1 wherein the probability of a hidden order is also a function of the time of day.

3. A system for constructing an order book from displayed market data for a tradable asset said system comprising:

means for measuring an effective spread of the tradable asset from displayed market data;

means for measuring a mid-quote volatility of the tradable asset from displayed market data;

means for measuring additions between best bid and ask of the tradable asset from displayed market data;

means for measuring additions less cancellations of the tradable asset from displayed market data;

means for calculating. with one or more computer processors, a probability of a hidden order for the tradable asset as a function of the measured effective spread, the mid-quote volatility, additions between best bid and ask, and additions less cancellations;

means for calculating, with one or more computer processors, a hidden order volume between the best bid and ask; and

means for calculating, with one or more computer processors, a hidden order price; and

means for constructing an order book for the tradable asset that includes displayed order volume and hidden order volume based on said calculated probability of a hidden order, said calculated hidden order volume, and said calculated hidden order price.

4. The system according to claim 3 , further comprising means for merging the calculated hidden order volume and hidden order price with a displayed order book.

5. A system for creating a model for calculating a probability and a characteristic of a hidden order for a tradable asset, said system comprising one or more computer processors configured to:

access a plurality of electronic, level-2 trading messages from a trading forum for a predefined period of time, each message including information about one or more orders for tradable assets or executed trades for tradable assets, said order information including identification of a tradable asset, a price, and a quantity;

identify executed trades from said messages;

classify a trade from the identified trades as displayed if the trade can be matched to orders in said messages;

classify a trade as hidden where said trade cannot be matched to orders in said messages;

determine a side of each order corresponding to a trade classified as hidden;

calculate, with the one or more computer processors, a hidden trade volume and a hidden trade location for tradable assets based upon said classifying steps and said determining step;

group each tradable asset in the plurality of tradable assets into one of a plurality of liquidity groups based upon said each tradable asset's median trade volume over a pre-determined liquidity period;

calculate, with the one or more computer processors, for each tradable asset in the plurality of tradable assets at least one market condition; and

calculate, with the one or more computer processors, for a liquidity group a coefficient associating the at least one market condition with at least one of said hidden trade volume and said hidden trade location.

6. The system according to claim 5 wherein the pre-determined liquidity period is a 21-day period coinciding with a first 21-days of the plurality of trading messages.

7. The system according to claim 5 , wherein the market condition comprises at least one of the an effective spread, a mid-quote volatility, additions between best bid and ask, average first level depth, order placements, order cancellations, and additions less cancellations over a pre-determined trading horizon.

8. The system according to claim 5 wherein the number of liquidity groups is 11.

9. The system of claim 5 wherein the real-time trading messages are obtained from ARCA Comstock L1 and L2 feeds.

10. The system of claim 5 wherein the trading messages are obtained from a direct exchange L2 feed.

11. The system of claim 5 wherein at least one coefficient x is standardized as X (standard) by its corresponding mean and standard deviation over a pre-determined prior standardization period.

12. The system of claim 11 wherein the pre-determined prior standardization period is the prior three months.

13. The system of claim 12 , wherein the standardized coefficient X (standard) is computed using the formula

X

(

standard

)

=

x

-

x

_

σ

⁡

(

x

)

where x is the mean over the pre-determined prior standardization period and σ(x) is the standard deviation of x over the pre-determined prior standardization period.

14. The system of claim 5 , wherein the one or more computer processors are further configured to estimate a McFadden's LRI to approximate a pseudo R 2 for assessing the goodness of fit of a coefficient.

Assignments (6)