CODEWITHSUNDEEP
python-3.xnumpymatplotlib
SiXUlm
SiXUlm

Reputation: 159

The graph produced by IPython is not correct

I have 2 files GBM_simulation.py and test_GBM.py.

GBM_simulation.py

import numpy as np

# Return path as numpy array of length and step predefined

# simulate exact GBM

def GBM(n, dt, s0, r, sigma):
    path = np.zeros(n)
    s_old = s0
    for i in range(n):
        s_new = s_old*np.exp((r-sigma**2/2)*dt + sigma*(dt)**(1/2)*np.random.normal())
        path[i] = s_new
        s_old = s_new
        return path

# simulate Euler scheme GBM:

def Euler_GBM(n, dt, s0, r, sigma):
    path = np.zeros(n)
    s_old = s0
    for i in range(n):
        s_new = s_old*(1 + r*dt + sigma*(dt)**(1/2)*np.random.normal())
        path[i] = s_new
        s_old = s_new
        return path

test_GBM.py

import matplotlib.pyplot as plt
import GBM_simulation as sim

s0 = 2
n = 500
r = 1
sigma = 1
dt = 0.001

exact = sim.GBM(n, dt, s0, r, sigma)
euler = sim.Euler_GBM(n, dt, s0, r, sigma)

# Now plot
plt.plot(exact, label = 'Exact')
plt.plot(euler, label = 'Euler')
plt.legend(loc = 'best')
plt.show()

However, when I run test_GBM.py on IPython console, the graph I get doesn't look correct at all. I tested on Jupyter Notebook (with some minor modifications) and it worked properly. Why does this happen?

enter image description here

Edit: this is the graph I got on Jupyter notebook

enter image description here

Upvotes: 0

Views: 88

Answers (1)

Mr. T
Mr. T

Reputation: 12410

I can't believe that I didn't notice this immediately. Indentation matters - your return statement was inside the loop. The correct function definition is of course:

import numpy as np

# Return path as numpy array of length and step predefined

# simulate exact GBM

def GBM(n, dt, s0, r, sigma):
    path = np.zeros(n)
    s_old = s0
    for i in range(n):
        s_new = s_old*np.exp((r-sigma**2/2)*dt + sigma*(dt)**(1/2)*np.random.normal())
        path[i] = s_new
        s_old = s_new
    return path

# simulate Euler scheme GBM:

def Euler_GBM(n, dt, s0, r, sigma):
    path = np.zeros(n)
    s_old = s0
    for i in range(n):
        s_new = s_old*(1 + r*dt + sigma*(dt)**(1/2)*np.random.normal())
        path[i] = s_new
        s_old = s_new
    return path

For future reference my debugging procedure - I printed out the values for each single data point in the function loop to see, why the values drop immediately to zero. Et voilà - only one data point was calculated, so something must have been wrong with your loop definition.

Upvotes: 1

Related Questions