Podemos buscar nuevos arboles de decisión intenetando mejorar nuestro sistema.
df['variable4']=df.variable3*df.decision4
total aciertos 3060
===============
total fallos 2758
===============
total 5818
===============
% aciertos 52
===============
suma aciertos 739.84
===============
suma fallos -609.53
===============
ratio 1.21378767247
===============
sharpe 1.18013497034
===============
<matplotlib.figure.Figure at 0x7f04562e0ad0>
95% de no perder mas de
********************************
-0.52
Mediana de la distribucion, trata que sea positiva
********************************
0.0
media de la distribucion
********************************
0.0313057057057
maxima ganancia mensual
********************************
4.54
maxima perdida mensual
********************************
-3.03
desviacion tipica
********************************
0.425890375859
dos veces la desviacion tipica
********************************
-0.820475046013
tres veces la desviacion tipica
********************************
-1.24636542187
Algo lo mejora y por tanto lo mantenemos. Hasta aquí tenemos un sharpe de 1,18.
y esta es la comparativa con la anterior
<matplotlib.figure.Figure at 0x7f0457522410>
total aciertos 3143 ===> 3060
===============
total fallos 2844 ===> 2758
===============
total 5987 ===> 5818
===============
% aciertos 52 ===> 52
===============
suma aciertos 753.24 ===> 739.84
===============
suma fallos -625.64 ===> -609.53
===============
ratio 1.20395115402 ===> 1.21378767247
===============
auma total 127.6 ===> 130.31
===============
===============^^^^^^^^^^^^^^==================
95% de no perder mas de
________________________
-3.73 -3.573
Mediana de la distribucion, trata que sea positiva
________________________
8.515 9.01
media de la distribucion
________________________
0.674732222625 10.0049414389
maxima ganancia mensual
________________________
41.0 41.01
maxima perdida mensual
________________________
-6.87 -7.96
desviacion tipica
________________________
2.04254495842 2.03880284317
dos veces la desviacion tipica
________________________
-3.41035769422 -3.3883438247
tres veces la desviacion tipica
________________________
-19.9050599688 -19.6231802808
ratio sharpe anualizado
________________________
1.14591169875 1.18013497034
Vamos a itener mejorarlo solo cuando una media corta este por debajo de la cotización.
df['variable6']=df.variable4*df.decision6
total aciertos 1928
===============
total fallos 1658
===============
total 3586
===============
% aciertos 53
===============
suma aciertos 469.92
===============
suma fallos -344.33
===============
ratio 1.36473731595
===============
sharpe 1.50162593117
===============
<matplotlib.figure.Figure at 0x7f04561e6f10>
95% de no perder mas de
********************************
-0.34
Mediana de la distribucion, trata que sea positiva
********************************
0.0
media de la distribucion
********************************
0.0301717717718
maxima ganancia mensual
********************************
4.54
maxima perdida mensual
********************************
-3.03
desviacion tipica
********************************
0.337865921194
dos veces la desviacion tipica
********************************
-0.645560070617
tres veces la desviacion tipica
********************************
-0.983425991812
este filtro tambien le mejora,llegando a un sharpe de 1,50
| Hour |
2000 |
2001 |
2002 |
2003 |
2004 |
2005 |
2006 |
2007 |
2008 |
2009 |
2010 |
2011 |
2012 |
2013 |
2014 |
2015 |
2016 |
2017 |
All |
| 0 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 5 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 6 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 7 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 8 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 9 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 10 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 11 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 12 |
|
|
|
|
|
|
|
0,35 |
9,88 |
-0,2 |
0,13 |
2,44 |
6,54 |
1,95 |
-0,76 |
2,75 |
0,96 |
-1,27 |
22,77 |
| 13 |
|
|
|
|
|
|
|
7,98 |
4,29 |
7,42 |
1,91 |
3,56 |
3,48 |
0,67 |
1,05 |
-2,29 |
0,71 |
0,14 |
28,92 |
| 14 |
|
|
|
|
|
|
|
9,91 |
13,58 |
0,19 |
1,51 |
5,34 |
4,15 |
0,52 |
4,07 |
-2,02 |
-1,16 |
-0,82 |
35,27 |
| 15 |
|
|
|
|
|
|
|
-0,91 |
5,76 |
-1,33 |
2,67 |
4,62 |
0,2 |
-0,22 |
-0,09 |
-0,19 |
1,35 |
-0,23 |
11,63 |
| 16 |
|
|
|
|
|
|
|
5,31 |
3,42 |
1,14 |
0,93 |
5,6 |
1,32 |
3,64 |
1,77 |
-1,42 |
4,06 |
1,23 |
27 |
| 17 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 19 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 20 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 21 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 22 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 23 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| All |
|
|
|
|
|
|
|
22,64 |
36,93 |
7,22 |
7,15 |
21,56 |
15,69 |
6,56 |
6,04 |
-3,17 |
5,92 |
-0,95 |
125,59 |
Y podemos compararlo con la serie anterior
total aciertos 3060 ===> 1928
===============
total fallos 2758 ===> 1658
===============
total 5818 ===> 3586
===============
% aciertos 52 ===> 53
===============
suma aciertos 739.84 ===> 469.92
===============
suma fallos -609.53 ===> -344.33
===============
ratio 1.21378767247 ===> 1.36473731595
===============
auma total 130.31 ===> 125.59
===============
===============^^^^^^^^^^^^^^==================
95% de no perder mas de
________________________
-3.573 -0.207
Mediana de la distribucion, trata que sea positiva
________________________
9.01 12.93
media de la distribucion
________________________
0.689261861644 16.2353352399
maxima ganancia mensual
________________________
41.01 54.36
maxima perdida mensual
________________________
-7.96 -5.14
desviacion tipica
________________________
2.03880284317 1.73279646843
dos veces la desviacion tipica
________________________
-3.3883438247 -2.80217255778
tres veces la desviacion tipica
________________________
-19.6231802808 -25.5175462538
ratio sharpe anualizado
________________________
1.18013497034 1.50162593117
0 Comentarios