Orc Wing Model
Orc Wing Model
Orc Wing Model
This section covers the volatility calculations and interpolation of volatilities for Orc volatility
models.
Wing model
Volatility surface
The volatility surface in Orc is defined as a set of volatility skews. Each volatility skew is
assigned to a certain expiration date. Depending on the preference setting, these expiration
dates can be set as fixed or can shift with time, keeping the interval to expiration date constant
for each of the volatility skews. This setting is selected on the Preferences panel and is per
client.
Each volatility skew is defined by its skew settings and current forward price (F) for its
expiration date..
The skew swimmingness rate (ssr) defines the skew movement along the strike scale with
respect to changes in the ATM forward price. The 0% ssr makes the skew fixed, that is, not
sensitive to movements in market prices. When ssr is set to 100%, the central skew point
follows the ATM forward price. The ssr setting can be defined for each expiration date
independently.
Note The calculated volatility is limited to the same range as the volatility reference parameter.
These parameters together with the "current forward price" (F) define the volatility curve for a
given expiration date. The volatility curve consists of six parts: two parabolic wings with a
beginning at the central point, two smoothing ranges behind each wing, and two constant levels
on each side of the skew.
The calculation of the volatility for a given strike can be described as follows.
Firstly, we calculate the current forward price as follows:
• If SSR = 100, F equals the ATM forward.
• If SSR = 0, F equals the reference price for the volatility entry.
• If SSR is somewhere
-ssr
in% 1between,
---&
– -ssr i.e. it is a fraction of swimming skew,
--- # $
then F = ! Atm " 100 ' ! Ref " 100
where SSR is any number between 0% and 100% and a zero value corresponds to a fixed
skew and 100% corresponds to a swimming skew.
Next, we calculate the current volatility and the current slope at the central skew point. These
values may differ from the reference volatility and the reference slope is SSR, SCR and VCR
are set.
• Current volatility,
vcr ' ssr ' ! Atm – Ref "
vc = % vr – ------------------------------ &
# Ref $
• Current slope,
scr ' ssr ' ! Atm – Ref "
sc = % sr – ------------------------------&
# Ref $
dc ! 1 + dsm " + x:
2
vol ! x " = vc + dc ! 2 + dsm " ! sc ( 2 " + ! 1 + dsm "pc ' dc
uc ! 1 + usm " * x:
2
vol ! x " = vc + uc ! 2 + usm " ! sc ( 2 " + ! 1 + usm "cc ' uc
Fitting
From a users point of view there is no difference between fitting and interpolation in this model.
The actual curve fitting in this model only amounts to saving the strike prices and the mean
values of all implied volatilities provided to the models fitting algorithm. When fitting is done
from the Volatility Manager or from a Trading window, the points that are to be stored on the
curve are calculated and a curve is calculated using the cubic spline interpolation.
Interpolation
Interpolation is performed with cubic spline polynomials. Given the strike prices X 0/ X 1/ X 2/ ./ X N
and the volatilities v0/ v1/ v 2/ ./ vN at these strike prices, the curve is approximated by cubic
polynomials on each interval (X i/ X i + 1 0 , i.e. for a strike price X 1 (X i/ X i + 1 0 the volatility at X is
give by
2 3
P i ! X " = a i + b i ! X – X i " + c i ! X – X i " + d i ! X – Xi "
The coefficients ai/ bi/ ci and di/ i = 0/ 1/ 2/ ./ N – 1 , are computed requiring that the resulting
polynomials satisfies C2 regularity at the internal nodes, i.e. that for each i = 0/ 1/ 2/ ./ N – 2
Pi ! Xi + 1 " = Pi + 1 ! Xi + 1 "
P i 2 ! X i + 1 " = P2 i + 1 ! X i + 1 "
P3 i ! X i + 1 " = P3i + 1 ! X i + 1 "