Modeling financial instruments using bid and ask prices

• US 8,090,644 B2
• Filed: 04/06/2009
• Issued: 01/03/2012
• Est. Priority Date: 10/31/2005
• Status: Expired due to Fees

• Current Assignee: Penson Worldwide, Inc.
• Original Assignee: Penson Worldwide, Inc.
• Inventors: Ferguson, Ralph Bruce, Cheung, Liam, Boyd, Ronald Scott
• Primary Examiner(s): Patel, Jagdish
• Application Number:
  US12/418,951
• Publication Number:
  US 20090192949A1
• Time in Patent Office: 1,002 Days
• Field of Search: 705/35, 705/37, 705/39
• US Class Current: 705/37
• CPC Class Codes: G06Q 20/10 specially adapted for electronic funds transfer [EFT] systems; specially adapted for home banking systems
• CPC Class Codes: G06Q 40/00 Finance; Insurance; Tax strategies; Processing of corporate or income taxes
• CPC Class Codes: G06Q 40/04 Trading; Exchange, e.g. stocks, commodities, derivatives or currency exchange
• CPC Class Codes: G06Q 40/06 Asset management; Financial planning or analysis

Abstract

A method for modeling an investment significant parameter of a financial instrument, using a computer. At least one series of historical bid prices of the financial instrument or historical ask prices of the financial instrument is provided. A financial model type that has at least one variable parameter is selected. The variable parameter(s) of the selected financial model type is initialized. The series of historical bid prices and/or historical ask prices is applied to the initialized financial model type to estimate the variable parameter(s). The resulting model of the financial instrument may be used to predict future values of the investment significant parameter of the financial instrument. These predicted future values may be used to determine whether to perform automated trades of the financial instrument.

Claims

1. A method for modeling an investment significant parameter of a financial instrument, using a computer, the method comprising the steps of:

a) providing at least one series of historical bid prices of the financial instrument or historical ask prices of the financial instrument as training data;

b) selecting a financial model type having at least one variable parameter for matching the selected financial model type to the behavior of the financial instrument;

c) initializing the at least one variable parameter of the selected financial model type; and

d) estimating the at least one variable parameter by applying the training data to the initialized financial model type with the computer to model the investment significant parameter of the financial instrument, wherein step (d) includes the steps of;

d1) calculating a series of historical values of the investment significant parameter based on the at least one series of historical bid prices or historical ask prices provided in step (a);

d2) using the financial model type selected in step (b) to calculate a predicted investment significant parameter, the calculation based on a set of historical quotes having a predetermined quantity of consecutive historical quotes of one of the at least one series of historical bid prices or historical ask prices provided in step (a);

d3) repeating step (d2) for each set of historical quotes to calculate a plurality of predicted investment significant parameters;

d4) comparing the plurality of predicted investment significant parameters to the series of historical values of the investment significant parameter;

d5) varying the at least one variable parameter of the selected financial model type based on differences between the plurality of predicted investment significant parameters and the series of historical values of the investment significant parameter; and

d6) repeating steps (d1), (d2), (d3), (d4), (d5), and (d6) until the plurality of predicted investment significant parameters calculated in steps (d2) and (d3) substantially correspond to the series of historical values of the investment significant parameter calculated in step (d1).

2. The method according to claim 1, wherein the financial instrument is a publicly traded financial instrument.

3. The method according to claim 1, wherein the financial instrument is one of a stock, a bond, a commodity, a currency, an equity, a derivative, or a future.

4. The method according to claim 1, wherein step (a) includes providing at least one complete consecutive series of:

the historical bid prices of the financial instrument spanning a predetermined period of time;

the historical ask prices of the financial instrument spanning the predetermined period of time;

a predetermined number of the historical bid prices of the starting from a predetermined time;

or

the predetermined number of the historical ask prices of the starting from the predetermined time.

5. The method according to claim 1, wherein:

step (a) includes providing at least one of;

a time series of the historical bid prices of the financial instrument, including corresponding bid times;

or

a time series of the historical ask prices of the financial instrument, including corresponding ask times; and

step (d) includes applying the at least one of the time series of historical bid prices or the time series of historical ask prices to the initialized financial model type to estimate the at least one variable parameter.

6. The method according to claim 1, wherein step (a) includes providing at least one of:

the historical bid prices of the financial instrument, including corresponding bid sizes;

or

the historical ask prices of the financial instrument, including corresponding ask sizes.

7. The method according to claim 1, wherein step (a) includes the steps of:

a1) providing a time series of the historical bid prices of the financial instrument, including corresponding bid times, and a time series of the historical ask prices of the financial instrument, including corresponding ask times;

a2) calculating a time series of a spread between the historical bid prices and the historical ask prices of the financial instrument from the time series of the historical bid prices and the time series of the historical ask prices;

a3) identifying outlier bids in the time series of the historical bid prices using the time series of the spread;

a4) removing the identified outliers bids and corresponding bid times from the time series of the historical bid prices;

a5) identifying outlier asks in the time series of the historical ask prices using the time series of the spread; and

a6) removing the identified outliers asks and corresponding ask times from the time series of the historical ask prices.

8. The method according to claim 1, wherein step (a) includes the steps of:

a1) providing a time series of the historical bid prices of the financial instrument, including corresponding bid times, and a time series of the historical ask prices of the financial instrument, including corresponding ask times; and

a2) calculating a time series of a spread between the historical bid prices and the historical ask prices of the financial instrument from the time series of the historical bid prices and the time series of the historical ask prices; and

step (d) includes applying the time series of the historical bid prices, the time series of the historical ask prices, and the time series of the spread between the historical bid prices and the historical ask prices to the initialized financial model type to estimate the at least one variable parameter.

9. The method according to claim 1, wherein step (b) includes selecting the financial model type to be one of:

an expert system model;

a linear analytic model;

a non-linear analytic model;

a chaotic model;

a Markov-chain Monte Carlo model;

a wavelet transformation model;

a regression model;

a fractal model;

a support vector machine model;

or a Bayesian model.

10. The method according to claim 1, wherein the investment significant parameter is one of:

a future trade price;

a future bid price;

a future ask price;

a future spread;

a fair market value (FMV);

an expected profit;

a change in trade price between two times;

a change in bid price between two times;

a change in ask price between two times;

a change in spread between two times;

a change in FMV between two times;

a change in profit between two times;

a rate of change of trade price;

a rate of change of bid price;

a rate of change of ask price;

a rate of change of spread;

a rate of change of FMV;

a rate of change of profit;

a prediction of winners and losers;

or a buy/sell instruction.

11. A method for modeling an investment significant parameter of a financial instrument, using a computer, the method comprising the steps of:

a) providing at least one series of historical bid prices of the financial instrument or historical ask prices of the financial instrument as training data;

b) selecting a financial model type having at least one variable parameter;

c) initializing the at least one variable parameter of the selected financial model type; and

d) estimating the at least one variable parameter by applying the training data to the initialized financial model type with the computer to model the investment significant parameters of the financial instrument, wherein the investment significant parameter is at least one of a future bid price or a future ask price; and

step (d) includes the steps of;

d1) using the financial model type selected in step (b) to calculate the at least one of a predicted bid price or a predicted ask price, the calculation based on a set of historical quotes having a predetermined quantity of consecutive historical quotes of one of the at least one series of historical bid prices or historical ask prices provided in step (a);

d2) repeating step (d1) for each set of historical quotes to calculate at least one of a plurality of predicted bid prices or a plurality of predicted ask prices;

d3) comparing the at least one of the plurality of predicted bid prices or the plurality of predicted ask prices to the at least one series of historical bid prices or historical ask prices;

d4) varying the at least one variable parameter of the selected financial model type based on differences between the least one of the plurality of predicted bid prices or the plurality of predicted ask prices and the at least one series of historical bid prices or historical ask prices; and

d5) repeating steps (d1), (d2), (d3), (d4), and (d5) until the at least one of the plurality of predicted bid prices or the plurality of predicted ask prices calculated in step (d1) and (d2) substantially correspond to the at least one series of historical bid prices or historical ask prices.