vanilla_option_pricing package¶

vanilla_option_pricing.calibration module¶

class vanilla_option_pricing.calibration.ModelCalibration(options: List[vanilla_option_pricing.option.VanillaOption])¶

Bases: object

Calibrate option pricing models with prices of listed options

Parameters:	options – a collection of `VanillaOption`

DEFAULT_PARAMETER_LOWER_BOUND = 0.0001¶

calibrate_model(model: vanilla_option_pricing.option_pricing.OptionPricingModel, method: str = None, options: Dict[KT, VT] = None, bounds: Union[str, Sequence[Tuple[float, float]]] = 'default') → Tuple[scipy.optimize.optimize.OptimizeResult, vanilla_option_pricing.option_pricing.OptionPricingModel]¶

Tune model parameters and returns a tuned model. The algorithm tries to minimize the squared difference between the prices of listed options and the prices predicted by the model: the parameters of the model are the optimization variables. The numerical optimization is performed by minimize() in the scipy package.

Parameters:	model – the model to calibrate method – see `minimize()` options – see `minimize()` bounds – the bounds to apply to parameters. If none is specified, then the `DEFAULT_PARAMETER_LOWER_BOUND` is applied for all the parameters. Otherwise, a list of tuples (lower_bound, upper_bound) for each parameter shall be specified.
Returns:	a tuple (res, model), where res is the result of `minimize()`, while model a calibrated model

vanilla_option_pricing.models module¶

class vanilla_option_pricing.models.GeometricBrownianMotion(s: float)¶

Bases: vanilla_option_pricing.option_pricing.OptionPricingModel

The celebrated Geometric Brownian Motion model

Parameters:	s – the volatility, must be non-negative

name = 'Geometric Brownian Motion'¶

parameters¶: Model parameters, as a tuple of real numbers, in the order (s, ).

variance(t: float) → float¶

The variance of the model output at a given time instant

Parameters:	t – the time when the variance is evaluated
Returns:	the variance at time t

class vanilla_option_pricing.models.LogMeanRevertingToGeneralisedWienerProcess(p_0: numpy.array, l: float, s_x: float, s_y: float)¶

Bases: vanilla_option_pricing.option_pricing.OptionPricingModel

The Log Mean-Reverting To Generalised Wiener Process model. It is a two-factor, mean reverting model, where the long-term behaviour is given by a Geometric Brownian motion, while the short-term mean-reverting tendency is modelled by an Ornstein-Uhlenbeck process.

Parameters:	p_0 – the initial variance, that is the variance of the state at time t=0. Must be a 2x2 numpy array l – the strength of mean-reversion, must be non-negative s_x – volatility of the long-term process, must be non-negative s_y – volatility of the short-term process, must be non-negative

name = 'Log Mean-Reverting To Generalised Wiener Process'¶

parameters¶: Model parameters, as a tuple of real numbers, in the order l, s_x, s_y.

variance(t: float) → float¶

The variance of the model output at a given time instant

Parameters:	t – the time when the variance is evaluated
Returns:	the variance at time t

class vanilla_option_pricing.models.NumericalLogMeanRevertingToGeneralisedWienerProcess(p_0: numpy.array, l: float, s_x: float, s_y: float)¶

Bases: vanilla_option_pricing.option_pricing.OptionPricingModel

This model relies on the same stochastic process as LogMeanRevertingToGeneralisedWienerProcess, but uses a numerical procedures based on a matrix exponential instead of the analytical formulas to compute the variance. As this approach is considerably slower, it is strongly suggested to adopt LogMeanRevertingToGeneralisedWienerProcess instead, using this class only for benchmarking

Parameters:	p_0 – the initial variance, that is the variance of the state at time t=0. Must be a 2x2 numpy array, symmetric and positive semidefinite l – the strength of mean-reversion, must be non-negative s_x – volatility of the long-term process, must be non-negative s_y – volatility of the short-term process, must be non-negative

name = 'Numerical Log Mean-Reverting To Generalised Wiener Process'¶

parameters¶: Model parameters, as a tuple of real numbers, in the order l, s_x, s_y.

variance(t: float) → float¶

The variance of the model output at a given time instant

Parameters:	t – the time when the variance is evaluated
Returns:	the variance at time t

class vanilla_option_pricing.models.NumericalModel(A: numpy.array, B: numpy.array, p_0: numpy.array)¶

Bases: object

A general-purpose linear stochastic system. All the parameters must be matrices (as Numpy arrays) of suitable dimensions.

Parameters:	A – the dynamic matrix A of the system B – the input matrix B of the system p_0 – the initial variance, that is the variance of the state at time t=0, must be symmetric and positive semidefinite

variance(t: float) → float¶

The variance of the model output at a given time instant

Parameters:	t – the time when the variance is evaluated
Returns:	the variance at time t

class vanilla_option_pricing.models.OrnsteinUhlenbeck(p_0: float, l: float, s: float)¶

Bases: vanilla_option_pricing.option_pricing.OptionPricingModel

The single-factor, mean-reverting Ornstein-Uhlenbeck process.

Parameters:	p_0 – the initial variance, that is the variance of the state at time t=0, must be positive semidefinite and symmetric l – the strength of the mean-reversion, must be non-negative s – the volatility, must be non-negative

name = 'Ornstein-Uhlenbeck'¶

parameters¶: Model parameters, as a list of real numbers, in the order [l, s].

variance(t: float) → float¶

The variance of the model output at a given time instant

Parameters:	t – the time when the variance is evaluated
Returns:	the variance at time t

vanilla_option_pricing.option module¶

class vanilla_option_pricing.option.VanillaOption(instrument: str, option_type: str, date: datetime.datetime, price: float, strike: float, spot: float, maturity: datetime.datetime, dividend=0)¶

Bases: object

A European vanilla option. All the prices must share the same currency.

Parameters:	instrument – name of the underlying option_type – type of the option (c for call, p for put) date – the date when the option is traded price – option price strike – option strike price spot – spot price of the underlying maturity – the maturity date dividend – underlying dividend - if any, expressed as a decimal number

DAYS_IN_YEAR = 365.2425¶

implied_volatility_of_undiscounted_price¶: The implied volatility of the option, considering an undiscounted price. Returns zero if the implied volatility is negative.

to_dict()¶

Returns:	all the fields of the object in a dictionary

years_to_maturity¶: The years remaining to option maturity, as a decimal number.

vanilla_option_pricing.option.check_option_type(option_type: str)¶: A utility function to check the validity of the type of an option. Raises a ValueError if the type is invalid. :param option_type: the type of the option: valid types are “c” for call and “p” for put

vanilla_option_pricing.option.option_list_to_pandas_dataframe(options: List[vanilla_option_pricing.option.VanillaOption]) → pandas.core.frame.DataFrame¶

A utility function to convert a list of VanillaOption to a pandas dataframe.

Parameters:	options – a list of `VanillaOption`
Returns:	a pandas dataframe, containing option data

vanilla_option_pricing.option.pandas_dataframe_to_option_list(data_frame: pandas.core.frame.DataFrame) → List[vanilla_option_pricing.option.VanillaOption]¶

A utility function to convert a pandas dataframe to a list of VanillaOption. For this function to work, the dataframe columns should be named as the parameters of VanillaOption’s constructor.

Parameters:	data_frame – a pandas dataframe, containing option data
Returns:	a list of `VanillaOption`

vanilla_option_pricing.option_pricing module¶

class vanilla_option_pricing.option_pricing.OptionPricingModel¶

Bases: abc.ABC

A model which can be used to price European vanilla options.

parameters¶: The model parameters, returned as a list of real numbers.

price_black(option_type: str, spot: float, strike: float, years_to_maturity: float) → float¶

Finds the no-arbitrage price of a European Vanilla option. Price is computed using the Black formulae, but the variance of the underlying is extracted from this model.

Parameters:	option_type – the type of the option (c for call, p for put) spot – the spot price of the underlying strike – the option strike price years_to_maturity – the years remaining before maturity - as a decimal number
Returns:	the no-arbitrage price of the option

price_black_scholes_merton(option_type: str, spot: float, strike: float, years_to_maturity: float, risk_free_rate: float, dividend: float = 0) → float¶

Finds the no-arbitrage price of a European Vanilla option. The price is computed using the Black-Scholes-Merton framework, but the variance of the underlying is extracted from this model.

Parameters:	option_type – the type of the option (c for call, p for put) spot – the spot price of the underlying strike – the option strike price years_to_maturity – the years remaining before maturity - as a decimal number risk_free_rate – the risk-free interest rate dividend – the dividend paid by the underlying - as a decimal number
Returns:	the no-arbitrage price of the option

price_option_black(option: vanilla_option_pricing.option.VanillaOption) → float¶

Same as price_black(), but the details of the vanilla option are provided by a VanillaOption object.

Parameters:	option – a `VanillaOption`
Returns:	the no-arbitrage price of the option

price_option_black_scholes_merton(option: vanilla_option_pricing.option.VanillaOption, risk_free_rate: float) → float¶

Same as price_black_scholes_merton(), but the details of the vanilla option are provided by a VanillaOption object.

Parameters:	option – a `VanillaOption` risk_free_rate – the risk-free interest rate
Returns:	the no-arbitrage price of the option

standard_deviation(t: float) → float¶

The standard deviation of the model output at a given instant, that is the squared root of the variance() at the same instant

Parameters:	t – the time when the standard deviation is evaluated
Returns:	the standard deviation at time t

variance(t: float) → float¶

The variance of the model output at a given time.

Parameters:	t – the time when the variance is evaluated
Returns:	the variance at time t

volatility(t: float) → float¶

The volatility of the model output at a certain instant, that is the standard_deviation() divided by the squared root of the time

Parameters:	t – the time when the volatility is evaluated
Returns:	the volatility at time t