Please,I'v 2 pricers for pricing vanilla options with Heston model :and the 2 models gives me differents results.Wich one is the good? or the more robust?? And what does that mean : in Heston93.xls : Integral ( fct(k),k=0..64 ) with 8 subintervals of length = 8 Thanks by advance

Hi,I compraed the heston93_pdf to my algorithm: If rho = 0 it seems that my algorithm returns the same results, for rho different from 0 I obtain different prices. However, your worksheet seems to apply Gauss Legendre algorithm to compute the integral while I am using Gauss Laguerre (150 nodes). Don't know wthere my prices are better or yours. Pls let me know when you learn more about it.

Hallo Mucki,it is the same code as in free code for Heston. Would you mind to give your figures and look at examples there thus having all at one place?

Here's my excel sheet. The DLL file contains the code for the heston formula and must be copied into the same directory.

Sorry this hasn't worked out the first time

Up to discounting it is the same (which is missing as "mib" pointed out in the other thread).The only thing might be numerical problems with high order Laguerre: one comes throughthe oscillating integral (see below) and the other through significant digits and roots forthe Laguerre polynomial: for order 64 the 'last' root is around 250 (i do not know it for a degree of 150) where it islikely that you get rounding errors _even if you have the exact root_ (which itself may bea problem). Otherwise stated: high order does not mean high exactness, it might result inthe converse. As i have only 15 digits in Excel i dicided to take a lower order.For highly oscillating integrands your nodes may be placed not that well w r t periodics,and for the following input the DLL gives a price of 29,9999999832 - below 30:Stock 100Strike 70Current Variance 0,04Long term Variance 0,04Mean reversion speed 1,15Volvol 0,39Interest Rate 0Time to expiry 0,03Correlation -0,64Market Price of Vol Risk 0Practical that is not a deep problem, but i think you will have exacter results takinglower order (which can adopt to the periodics) and do repeated finite integrationsand if you like Gauss Laguerre do one at the end.