Risk_Budgeting/hist_sampling_main.py at main · Marco112358/Risk_Budgeting · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
from api_testing import cg_pull
import pandas as pd
import functools as ft
import numpy as np
import functions as fn
import matplotlib.pyplot as plt


token_nms = ['btc', 'eth', 'atom']
col_nms = ['datetime'] + token_nms
st_dollars = 10000
fee_pct = 0.003
tgt_wghts = [0.33333333, 0.33333333, 0.33333333]
st_price = 100
sample_len = 30
no_samples = 12
no_trials = 100
prc_st_list = [st_price] * len(token_nms)

# API Pulls from Coin Gecko
btc = cg_pull('bitcoin', 'usd', 'max', 'daily')
atom = cg_pull('cosmos', 'usd', 'max', 'daily')
# kuji = cg_pull('kujira', 'usd', 'max', 'daily')
# usdc = cg_pull('usd-coin', 'usd', 'max', 'daily')
eth = cg_pull('ethereum', 'usd', 'max', 'daily')
# sol = cg_pull('solana', 'usd', 'max', 'daily')
# ada = cg_pull('cardano', 'usd', 'max', 'daily')
dfs = [btc, eth, atom]  # this needs to be changed if u change the tokens allocated to

# Implied Parameters
timeperiod = sample_len * no_samples
prices = ft.reduce(lambda left, right: pd.merge(left, right, on='datetime'), dfs)
prices.columns = col_nms
prices.set_index('datetime', inplace=True)
rtns = np.divide(prices.iloc[1:], prices.iloc[:-1]) - 1

# Historical Sample
hist_prc_samp, hist_rtn_samp = fn.hist_sample(rtns, sample_len, no_samples, prc_st_list)


# Compare 30D Rebal to No Rebalance on 500 Different Start Dates
rebal_freq3 = 30
no_rebal = timeperiod
n_days = sample_len * no_samples
index_list = np.arange(0, n_days)
diff_final = pd.DataFrame(index=hist_prc_samp.index, columns=index_list)

for i in index_list:
    prices_trim = hist_prc_samp.loc[hist_prc_samp.index >= i]
    fwd_time = prices_trim.shape[0]
    # 30D Rebal
    tkns_final3, fees3 = fn.rebal_by_period(fwd_time, rebal_freq3, prices_trim, st_dollars, tgt_wghts, fee_pct)
    port_val3 = pd.DataFrame(data=np.sum(tkns_final3 * prices_trim, axis=1), index=prices_trim.index,
                             columns=['Portfolio Values 30D Rebal'])
    # No Rebal
    tkns_final10, fees10 = fn.rebal_by_period(fwd_time, no_rebal, prices_trim, st_dollars, tgt_wghts, fee_pct)
    port_val10 = pd.DataFrame(data=np.sum(tkns_final10 * prices_trim, axis=1), index=prices_trim.index,
                              columns=['Portfolio Values No Rebalance'])
    # Create the dataset of Rebal / No Rebal (value > 1 means Rebal better on that date
    mrg = pd.concat([port_val3, port_val10], axis=1)
    diff = mrg.iloc[:, 0] / mrg.iloc[:, 1]
    diff_final[i] = diff
diff_final.fillna(value=1, inplace=True)

diff_final.columns = index_list

fig = plt.figure(figsize=(8, 6))
ax = fig.add_subplot()
ax.plot(diff_final.index, diff_final.loc[:, index_list], color='b', linewidth=0.05)
ax.grid()
fig.tight_layout()
plt.show()

last_diff = diff_final.tail(1)
pct_above_one = last_diff[last_diff > 1].count(axis=1) / last_diff.shape[1]
print('Percentage of Rebalanced Portfolios that end Greater Than No Rebalanced Portfolios starting every day is '
      + str(pct_above_one * 100) + '%')


# Loop Through No Trials
rebal_freq3 = 30
port_nms = ['Portfolio Values 30D Rebal', 'Portfolio Values No Rebalance']
no_rebal = timeperiod
diff_final = pd.DataFrame(index=np.arange(0, timeperiod), columns=np.arange(0, no_trials))
pv_finals = pd.DataFrame(index=np.arange(0, no_trials), columns=port_nms)
for i in np.arange(0, no_trials):
    hist_prc_samp, hist_rtn_samp = fn.hist_sample(rtns, sample_len, no_samples, prc_st_list)
    # 30D Rebal
    tkns_final3, fees3 = fn.rebal_by_period(timeperiod, rebal_freq3, hist_prc_samp, st_dollars, tgt_wghts, fee_pct)
    port_val3 = pd.DataFrame(data=np.sum(tkns_final3 * hist_prc_samp, axis=1), index=hist_prc_samp.index,
                             columns=[port_nms[0]])
    # No Rebal
    tkns_final10, fees10 = fn.rebal_by_period(timeperiod, no_rebal, hist_prc_samp, st_dollars, tgt_wghts, fee_pct)
    port_val10 = pd.DataFrame(data=np.sum(tkns_final10 * hist_prc_samp, axis=1), index=hist_prc_samp.index,
                              columns=[port_nms[1]])
    # Create the dataset of Rebal / No Rebal (value > 1 means Rebal better on that date
    mrg = pd.concat([port_val3, port_val10], axis=1)
    diff = mrg.iloc[:, 0] / mrg.iloc[:, 1]
    diff_final[i] = diff
    pv_finals.iloc[i, :] = mrg.tail(1)

# Add Absolute Difference between Rebalanced and No Rebalance
pv_finals.loc[:, 'Difference'] = pv_finals.iloc[:, 0] - pv_finals.iloc[:, 1]
port_nms.append('Difference')

# Plot the Relative difference (rebal / no rebal) across all trials for each day
fig = plt.figure(figsize=(8, 6))
ax = fig.add_subplot()
ax.plot(diff_final.index, diff_final.loc[:, np.arange(0, no_trials)], color='b', linewidth=0.05)
ax.grid()
fig.tight_layout()
plt.show()

# Plot the Final Value of Each Portfolio (rebal and no rebal), and the Absolute Difference across all Trials
fig = plt.figure(figsize=(8, 6))
ax = fig.add_subplot()
ax.plot(pv_finals.index, pv_finals, label=port_nms)
plt.legend(loc='lower left')
ax.grid()
fig.tight_layout()
plt.show()

# Print the % of Rebal Portfolios that end > no Rebal Portfolios
last_diff = diff_final.tail(1)
pct_above_one = last_diff[last_diff > 1].count(axis=1) / last_diff.shape[1]
print('Percentage of Rebalanced Portfolios that end Greater Than No Rebalanced Portfolios starting every day is '
      + str(pct_above_one * 100) + '%')

# Print the Average and Median Absolute Difference Between Rebal and No Rebal Portfolios
ave_diff = np.average(pv_finals.loc[:, 'Difference'])
med_diff = np.median(pv_finals.loc[:, 'Difference'])
print('Average Difference between Rebalanced Portfolios versus No Rebalanced Portfolios: $'
      + str(round(ave_diff)))
print('Median Difference between Rebalanced Portfolios versus No Rebalanced Portfolios: $'
      + str(round(med_diff)))