[R-SIG-Finance] Pricing option using Explicit Finite Differencemethod

[Panov, Evgeny]

Dear Joseph,

It seems like you need to price a European option.

Doing it R with finite-difference method will be very inefficient (will probably take more than one second per option), because R is inefficient with "for" loops (unless you can represent these multiplications that you are doing within the loop as one or two a huge matrix multiplications).

However, if it is just a European option, you can do everything in closed-form (Black-Scholes for lognormal RND or Black-Scholes with moment-matching for non-lognormal with discrete dividends will do the job), and it will take less than a second in R. 

So maybe instead of debugging finite-differences in R, you may consider not using them at all (if you want to do finite-differences, it can be done much more efficiently in c++, .net or java).

Best regards,

Gene

-----Original Message-----
From: r-sig-finance-bounces at stat.math.ethz.ch
[mailto:r-sig-finance-bounces at stat.math.ethz.ch]On Behalf Of Joseph
Khalil
Sent: Tuesday, July 17, 2007 9:56 AM
To: r-sig-finance at stat.math.ethz.ch
Subject: Re: [R-SIG-Finance] Pricing option using Explicit Finite
Differencemethod


I am a new user of R and new to this mailing list.         I am using 
the book "Implementing Derivatives Models" by Les Clewlow and Chris
Strickland, and I am trying to simulate

the example on pricing options using the explicit finite difference
method in  chapter 3. I have tried the following to price plain European calls  but I am getting errors.  





K<-100      #Strike price

T<-1        #Time of maturity

t<-0        #Current time

sigma<-0.2  #Volatility

r<-0.06     #Interest rate

div<-0.03   # Dividend yield

S<-100      #Central stock price

dx<-0.02     #Space step around central asset price (logarithmic measurement exp(dx))

Nj<-30      #Number of space steps

N<-500      #Number of time steps until T



dt<-T/N # Time step

nu<-r-div-0.5*sigma^2

pu<-0.5*dt*((sigma/dx)^2+nu/dx)      # Probability for go up

pm<-1-dt*(sigma/dx)^2-r*dt           # Probability for stay the same

pd<-0.5*dt*((sigma/dx)^2-nu/dx)      # Probability for go down

# Initialise matrices

St<-matrix(data=0,nrow=2*Nj+1,ncol=1)

Call<-matrix(data=0,nrow=2*Nj+1,ncol=N+1)



#Compute independent values

for (j in (-Nj:Nj)) {

St[Nj-j+1,1]<-S*exp(j*dx)                 # Initialise asset prices at maturity

Call[Nj-j+1,N+1]<-max(0,St[Nj-j+1,1]-K) }  # Initialise option prices at maturity



#Compute dependent option values

    for (i in (N:1)) {       

# Explicit compution of option values (backwards)

        for (j in (-Nj+1:Nj-1))  { 

           
Call[j+Nj+1,i]<-as.matrix(pu*(as.matrix(Call[j+Nj,i+1])))+as.matrix(pm*Call[j+Nj+1,i+1])+as.matrix(pd*Call[j+Nj+2,i+1])


                                  }    

Call[1,i]<-Call[2,i]+St[1,1]-St[2,1]  # Upper boundary condition

Call[2*Nj+1,i]<-Call[2*Nj,i]          #Lower boundary condition

}

----- Original Message ----
From: Joseph Khalil <joedtka at yahoo.com>
To: r-sig-finance at stat.math.ethz.ch
Sent: Monday, July 16, 2007 10:01:27 PM
Subject: [R-SIG-Finance] Pricing option using Explicit Finite Difference method

Hi,

Does anyone have R code to calculate option prices (European, plain vanilla) using Explicit Finite
Difference method?   

Thanks,
Joe

_______________________________________________
R-SIG-Finance at stat.math.ethz.ch mailing list
https://stat.ethz.ch/mailman/listinfo/r-sig-finance
-- Subscriber-posting only. 
-- If you want to post, subscribe first.

_______________________________________________
R-SIG-Finance at stat.math.ethz.ch mailing list
https://stat.ethz.ch/mailman/listinfo/r-sig-finance
-- Subscriber-posting only. 
-- If you want to post, subscribe first.