The first backtesting program written in Python

The first backtesting program written in Python

2016-08-06

1 defSavfig (Figureobj, fn_prefix1='Backtest8', fn_prefix2='_1_'):2     Importdatetime3fmt='%y_%m_%d_%h_%m_%s'4now =Datetime.datetime.now ()5Fname_savfig = fn_prefix1 + fn_prefix2 + now.strftime (FMT) +'. PNG'6Figureobj.savefig (Fname_savfig, facecolor=Fig.get_facecolor ())7 8 9 defBacktest8 (OHLC=OHLC, sd=1.0, n_short=2, n_long=20, f_savfig=False):TenU" " One Double Moving Averages strategy backtesting function A Signature:backtest8 (OHLC=OHLC, sd=1.0, n_short=2, n_long=20, F_savfig=false) - param:: - OHLC-DOHLCVA data, dataframe structure of the the sd-ma1/ma2 > SD trigger long-buy fast moving average/slow moving average threshold - f_savefig-flag for saving matplot output figures -      -      +      -     " " +     Importmatplotlib A     #import Seaborn as SNS at     #sns.set_style (' white ') -      -Myfontprops =Matplotlib.font_manager. Fontproperties ( -Fname='C:/windows/fonts/msyh.ttf')#Microsoft Ya-Black -                          -Mashort =PD. Series.rolling (OHLC. C, N_short). Mean () inMalong =PD. Series.rolling (OHLC. C, N_long). Mean () -  to      +Fig=plt.figure ()#Create new figure -Ohlc. C.plot (Grid=true, figsize= (8,4)) theMashort.plot (label='MA'+str (n_short)) *Malong.plot (grid=true,label='MA'+str (n_long)) $ #Ohlc.iloc[:,[0,1,2,3]].plot (Grid=false, figsize= (8,4))Panax Notoginseng #Ohlc.iloc[:,[0,1,2,3]].plot (grid=true,figsize= (8,4)) -Plt.legend (loc=' Best') thePlt.title (s=u'Historical share price', fontproperties=myfontprops) +     ifF_savfig: ASavfig (Fig,'Backtest8','_0_') the          + #sd=1.0 -Regime = Np.where (Mashort/malong > SD, 1, 0) $Regime = PD. Series (Regime, index=Mashort.index) $     Print('regime Length =%s'%regime.size) -          -Fig=plt.figure ()#Create new figure theRegime[:].plot (lw=1.5, ylim= ( -0.1, 1.1), figsize= (8,4), Title=u'regime') -     ifF_savfig:WuyiSavfig (Fig,'Backtest8','_1_') the          -Fig=plt.figure ()#Create new figure WuRegime[-100:].plot (lw=1.5, ylim= ( -0.1, 1.1), figsize= (8,4), Title=u'regime') -     ifF_savfig: AboutSavfig (Fig,'Backtest8','_2_') $      -      -      -PP_RATIO_BNH = Np.log (OHLC. C/OHLC. C.shift (1) ) APp_ratio_strategy = Regime.shift (1) *PP_RATIO_BNH +     #In the end, we sum up the daily yield and get the final cumulative rate of return . the     #(This is because we are using an exponential yield, so it is reasonable to accumulate the daily earnings), -     #This cumulative process can also be done easily with Dataframe's built-in function Cumsum: $NORM_RETURN_BNH =pp_ratio_bnh. Cumsum (). Apply (NP.EXP) theNorm_return_strategy =pp_ratio_strategy.cumsum (). Apply (NP.EXP) the      theFig=plt.figure ()#Create a new figure theNorm_return_strategy. Plot (lw=1.5, figsize= (8,4), Label=u'Strategy') -Norm_return_bnh. Plot (lw=1.5, label=u'BnH') in      thePlt.legend (loc=' Best') thePlt.title (s=u'strategic yield vs. historical price', fontproperties=myfontprops) About     ifF_savfig: theSavfig (Fig,'Backtest8','_3_') the      the     assert(Regime.index = = OHLC. C.index). All () ==true#' signal index not equals price index ' +     #assert is used to determine whether a statement is true or FALSE, and if false it will trigger a assertionerror error, an expression that prompts the developer for an error -     returnNorm_return_strategy, N_short, N_long, SD

Result diagram: There are four, mainly for the purpose of quality control.

1. Historical price

2. Trading Signals

3. A subset of the 2nd, zoom in to see clearly, technical indicators timing model details (how to trigger trading signals)

4. Rate of return on the strategy

Follow-up additions:

1. Encapsulation into class

2. Add a Performance strategy indicator: a whole bunch of stuff.

3. Adding optimizations

4. Perfect drawing program, intelligently Select input (data_obj, param, **kwargs)

The first backtesting program written in Python