(本文由公众号越声攻略(yslc688)整理，仅供参考，不构成操作建议。如自行操作，注意仓位控制和风险自负。)
股市有一句谚语，量在价先，即天量见天价，地量见地价;底部放量，闭眼买进。讲的就是成交量对股价的影响，因为成交量是价格形成的基础，基本决定了当前股价以及后面股价的走势，所以老股民最常看的就是成交量了，今天我就给大家讲讲怎样用成交量选股，在所有的成交量选股策略中，其中有3个策略是用的最多的。
一阳穿三线
受限于A股的交易规则，T日买进的股票，当天是无法卖出的，因此在T+1日的时候，无论涨跌，不管是赚了多少，还是亏了多少，也不管T+2日是否还会继续涨，一定要择机选择卖出，要坚决完成卖出操作，不能犹豫，不能恋战，否则会形成不好的习惯！
判断圆弧底的条件是市场从急跌变为缓跌,再由缓涨变为跳空暴涨,跳空暴涨为圆弧底完成的标志。
在这一波反弹行情中，有些庄股出现了低位放量的现象，但股价却并没有上涨，而是进行了平台式的整理。随着反弹行情的结束，这部分庄股也走出了破位之势。应该说，这部分庄股的出货迹象还是十分明显的，只不过是这种出货手法具有相当的欺骗性，具体表现在低位与放量两个方面。
这样，T日计划买入的股票标的可以定在1-3只，多了会就存在选择困难。
1. 概述
将股票在短期内的量价走势分类为量价背离与量价同向，并通过量价相关性来衡量量价走势的背离与同向程度
按照量价因子选股的月度多空收益在1%以上，得到了很显著的alpha
纯多头组合在六年回测中年化收益达到22.4%，信息比率达到2.22
量价因子等权叠加了反转因子后，六年回测年化收益达到26.0%，信息比率达到2.55
2. 量价因子构建
def getVolPriceCorrAll(universe, begin, end, window, file_name): # 计算各股票历史区间window天窗口移动的量价相关系数 # 拿取上海证券交易所日历 cal_dates = DataAPI.TradeCalGet(exchangeCD=u"XSHG", beginDate=begin, endDate=end).sort('calendarDate') cal_dates = cal_dates[cal_dates['isOpen']==1] all_dates = cal_dates['calendarDate'].values.tolist() # 工作日列表 print str(window) + ' days Price-Volume-Corr will be calculated for ' + str(len(universe)) + ' stocks:' count = 0 secs_time = 0 start_time = time.time() ret_data = pd.DataFrame() # 保存计算出来的收益率数据 ret_data.to_csv(file_name) N = 10 for i in range(len(universe)/N+1): sub_univ = universe[i*N:(i+1)*N] if len(sub_univ) == 0: continue data = DataAPI.MktEqudAdjGet(secID=sub_univ, beginDate=begin, endDate=end, field='secID,tradeDate,turnoverRate,preClosePrice,closePrice') # 拿取数据 for stk in sub_univ: # 对每一只股票分别计算历史window天前望收益率 tmp_ret_data = data[data.secID==stk].sort('tradeDate') corr_data = range(len(tmp_ret_data)) for i in range(window-1, len(tmp_ret_data)): x = tmp_ret_data['turnoverRate'].values[i-window+1:i+1] y = tmp_ret_data['closePrice'].values[i-window+1:i+1] corr_data[i] = st.spearmanr(x, y)[0] # 计算前向收益率 tmp_ret_data['corr'] = corr_data tmp_ret_data = tmp_ret_data[['tradeDate','corr']] tmp_ret_data.columns = ['tradeDate', stk] ret_data = pd.read_csv(file_name) if ret_data.empty: ret_data = tmp_ret_data else: ret_data = ret_data[ret_data.columns[1:]] ret_data = ret_data.merge(tmp_ret_data, on='tradeDate', how='outer') ret_data = ret_data.sort('tradeDate') ret_data.to_csv(file_name) # 打印进度部分 count += 1 if count > 0 and count % 2 == 0: finish_time = time.time() print count*N, print ' ' + str(np.round((finish_time-start_time) - secs_time, 0)) + ' seconds elapsed.' secs_time = (finish_time-start_time) return ret_data def getBackwardReturnsAll(universe, begin, end, window, file_name): # 计算各股票历史区间回报率，过去window天的收益率 print str(window) + ' days backward returns will be calculated for ' + str(len(universe)) + ' stocks:' count = 0 secs_time = 0 start_time = time.time() N = 50 ret_data = pd.DataFrame() for stk in universe: data = DataAPI.MktEqudAdjGet(secID=stk, beginDate=begin, endDate=end, field='secID,tradeDate,closePrice') # 拿取数据 tmp_ret_data = data.sort('tradeDate') # 计算历史窗口收益率 tmp_ret_data['forwardReturns'] = tmp_ret_data['closePrice'] / tmp_ret_data['closePrice'].shift(window) - 1.0 tmp_ret_data = tmp_ret_data[['tradeDate','forwardReturns']] tmp_ret_data.columns = ['tradeDate', stk] if ret_data.empty: ret_data = tmp_ret_data else: ret_data = ret_data.merge(tmp_ret_data, on='tradeDate', how='outer') # 打印进度部分 count += 1 if count > 0 and count % N == 0: finish_time = time.time() print count, print ' ' + str(np.round((finish_time-start_time) - secs_time, 0)) + ' seconds elapsed.' secs_time = (finish_time-start_time) ret_data.to_csv(file_name) return ret_data def getForwardReturnsAll(universe, begin, end, window, file_name): # 计算各股票历史区间前瞻回报率，未来window天的收益率 print str(window) + ' days forward returns will be calculated for ' + str(len(universe)) + ' stocks:' count = 0 secs_time = 0 start_time = time.time() N = 50 ret_data = pd.DataFrame() for stk in universe: data = DataAPI.MktEqudAdjGet(secID=stk, beginDate=begin, endDate=end, field='secID,tradeDate,closePrice') # 拿取数据 tmp_ret_data = data.sort('tradeDate') # 计算历史窗口前瞻收益率 tmp_ret_data['forwardReturns'] = tmp_ret_data['closePrice'].shift(-window) / tmp_ret_data['closePrice'] - 1.0 tmp_ret_data = tmp_ret_data[['tradeDate','forwardReturns']] tmp_ret_data.columns = ['tradeDate', stk] if ret_data.empty: ret_data = tmp_ret_data else: ret_data = ret_data.merge(tmp_ret_data, on='tradeDate', how='outer') # 打印进度部分 count += 1 if count > 0 and count % N == 0: finish_time = time.time() print count, print ' ' + str(np.round((finish_time-start_time) - secs_time, 0)) + ' seconds elapsed.' secs_time = (finish_time-start_time) ret_data.to_csv(file_name) return ret_data def getMarketValueAll(universe, begin, end, file_name): # 获取股票历史每日市值 print 'MarketValue will be calculated for ' + str(len(universe)) + ' stocks:' count = 0 secs_time = 0 start_time = time.time() N = 50 ret_data = pd.DataFrame() for stk in universe: data = DataAPI.MktEqudAdjGet(secID=stk, beginDate=begin, endDate=end, field='secID,tradeDate,marketValue') # 拿取数据 tmp_ret_data = data.sort('tradeDate') # 市值部分 tmp_ret_data = tmp_ret_data[['tradeDate','marketValue']] tmp_ret_data.columns = ['tradeDate', stk] if ret_data.empty: ret_data = tmp_ret_data else: ret_data = ret_data.merge(tmp_ret_data, on='tradeDate', how='outer') # 打印进度部分 count += 1 if count > 0 and count % N == 0: finish_time = time.time() print count, print ' ' + str(np.round((finish_time-start_time) - secs_time, 0)) + ' seconds elapsed.' secs_time = (finish_time-start_time) ret_data.to_csv(file_name) return ret_data def getWindowMeanTurnoverRateAll(universe, begin, end, window, file_name): # 获取股票历史滚动窗口平均换手率 print 'WindowMeanTurnoverRate will be calculated for ' + str(len(universe)) + ' stocks:' count = 0 secs_time = 0 start_time = time.time() N = 100 ret_data = pd.DataFrame() for stk in universe: data = DataAPI.MktEqudAdjGet(secID=stk, beginDate=begin, endDate=end, field='secID,tradeDate,turnoverRate') # 拿取数据 tmp_ret_data = data.sort('tradeDate') # 市值部分 tmp_ret_data['windowMeanTurnoverRate'] = pd.rolling_mean(tmp_ret_data['turnoverRate'], window=window) tmp_ret_data = tmp_ret_data[['tradeDate','windowMeanTurnoverRate']] tmp_ret_data.columns = ['tradeDate', stk] if ret_data.empty: ret_data = tmp_ret_data else: ret_data = ret_data.merge(tmp_ret_data, on='tradeDate', how='outer') # 打印进度部分 count += 1 if count > 0 and count % N == 0: finish_time = time.time() print count, print ' ' + str(np.round((finish_time-start_time) - secs_time, 0)) + ' seconds elapsed.' secs_time = (finish_time-start_time) ret_data.to_csv(file_name) return ret_data
begin_date = '20060101' # 开始日期 end_date = '20160802' # 结束日期 universe = set_universe('A') # 股票池 universe = universe[0:50] # 计算速度缓慢，仅以部分股票池作为sample # ----------- 计算量价相关系数部分 ---------------- window_corr = 15 print '=======================' start_time = time.time() forward_returns_data = getVolPriceCorrAll(universe=universe, begin=begin_date, end=end_date, window=window_corr, file_name='VolPriceCorr_W15_FullA_sample.csv') finish_time = time.time() print '' print str(finish_time-start_time) + ' seconds elapsed in total.' # ----------- 计算股票历史窗口（一个月）收益率部分 ---------------- window_return = 20 print '=======================' start_time = time.time() forward_returns_data = getBackwardReturnsAll(universe=universe, begin=begin_date, end=end_date, window=window_return, file_name='BackwardReturns_W20_FullA_Sample.csv') finish_time = time.time() print '' print str(finish_time-start_time) + ' seconds elapsed in total.' # ----------- 计算股票历史窗口（三个月）收益率部分 ---------------- window_return = 60 print '=======================' start_time = time.time() forward_returns_data = getBackwardReturnsAll(universe=universe, begin=begin_date, end=end_date, window=window_return, file_name='BackwardReturns_W60_FullA_Sample.csv') finish_time = time.time() print '' print str(finish_time-start_time) + ' seconds elapsed in total.' # ----------- 计算股票前瞻收益率部分 ---------------- window_return = 20 print '=======================' start_time = time.time() forward_returns_data = getForwardReturnsAll(universe=universe, begin=begin_date, end=end_date, window=window_return, file_name='ForwardReturns_W20_FullA_Sample.csv') finish_time = time.time() print '' print str(finish_time-start_time) + ' seconds elapsed in total.' # ----------- 计算股票历史市值部分 ---------------- print '=======================' start_time = time.time() forward_returns_data = getMarketValueAll(universe=universe, begin=begin_date, end=end_date, file_name='MarketValues_FullA_Sample.csv') finish_time = time.time() print '' print str(finish_time-start_time) + ' seconds elapsed in total.' # ----------- 计算历史月度日均换手率部分 ---------------- window = 20 print '=======================' start_time = time.time() forward_returns_data = getWindowMeanTurnoverRateAll(universe=universe, begin=begin_date, end=end_date, window=window, file_name='TurnoverRateWindowMean_W20_FullA_Sample.csv') finish_time = time.time() print '' print str(finish_time-start_time) + ' seconds elapsed in total.'3. 量价因子截面特征
3.2 量价相关因子截面特征
# 量价相关性历史表现 n_quantile = 10 # 和海通研报一样，统计十分位数 cols_mean = ['meanQ'+str(i+1) for i in range(n_quantile)] cols = cols_mean corr_means = pd.DataFrame(index=corr_data.index, columns=cols) # 计算相关系数分组平均值 for dt in corr_means.index: qt_mean_results = [] # 相关系数去掉nan和绝对值大于1的 tmp_corr = corr_data.ix[dt].dropna() tmp_corr = tmp_corr[(tmp_corr<=1.0) & (tmp_corr>=-1.0)] pct_quantiles = 1.0/n_quantile for i in range(n_quantile): down = tmp_corr.quantile(pct_quantiles*i) up = tmp_corr.quantile(pct_quantiles*(i+1)) mean_tmp = tmp_corr[(tmp_corr<=up) & (tmp_corr>=down)].mean() qt_mean_results.append(mean_tmp) corr_means.ix[dt] = qt_mean_results # corr_means是对历史每一天，求量价相关系数在各个十分位里面的平均值 corr_means.tail()
# 量价相关性历史表现作图 fig = plt.figure(figsize=(16, 6)) ax1 = fig.add_subplot(111) lns1 = ax1.plot(corr_means.index, corr_means.meanQ1, label='Q1') lns2 = ax1.plot(corr_means.index, corr_means.meanQ5, label='Q5') lns3 = ax1.plot(corr_means.index, corr_means.meanQ10, label='Q10') lns = lns1+lns2+lns3 labs = [l.get_label() for l in lns] ax1.legend(lns, labs, bbox_to_anchor=[0.5, 0.1], loc='', ncol=3, mode="", borderaxespad=0., fontsize=12) ax1.set_ylabel(u'量价相关系数', fontproperties=font, fontsize=16) ax1.set_xlabel(u'日期', fontproperties=font, fontsize=16) ax1.set_title(u"量价相关性历史表现", fontproperties=font, fontsize=16) ax1.grid()3.3 量价因子的预测能力初探
-0.04 median of IC: -0.0477574767849 the number of IC(all, plus, minus): (2572, 778, 1760)
# ‘过去十五天量价相关系数’和‘之后20天收益’的秩相关系数作图 fig = plt.figure(figsize=(16, 6)) ax1 = fig.add_subplot(111) lns1 = ax1.plot(ic_data.index, ic_data.IC, label='IC') lns = lns1 labs = [l.get_label() for l in lns] ax1.legend(lns, labs, bbox_to_anchor=[0.5, 0.1], loc='', ncol=3, mode="", borderaxespad=0., fontsize=12) ax1.set_ylabel(u'相关系数', fontproperties=font, fontsize=16) ax1.set_xlabel(u'日期', fontproperties=font, fontsize=16) ax1.set_title(u"量价因子和未来20日收益之间的秩相关系数", fontproperties=font, fontsize=16) ax1.grid()4. 量价因子历史回测概述
n_quantile = 10 # 和海通研报一样，统计十分位数 cols_mean = [i+1 for i in range(n_quantile)] cols = cols_mean excess_returns_means = pd.DataFrame(index=corr_data.index, columns=cols) # 计算相关系数分组的超额收益平均值 for dt in excess_returns_means.index: qt_mean_results = [] # 相关系数去掉nan和绝对值大于1的 tmp_corr = corr_data.ix[dt].dropna() tmp_corr = tmp_corr[(tmp_corr<=1.0) & (tmp_corr>=-1.0)] tmp_return = forward_20d_return_data.ix[dt].dropna() tmp_return_mean = tmp_return.mean() pct_quantiles = 1.0/n_quantile for i in range(n_quantile): down = tmp_corr.quantile(pct_quantiles*i) up = tmp_corr.quantile(pct_quantiles*(i+1)) i_quantile_index = tmp_corr[(tmp_corr<=up) & (tmp_corr>=down)].index mean_tmp = tmp_return[i_quantile_index].mean() - tmp_return_mean qt_mean_results.append(mean_tmp) excess_returns_means.ix[dt] = qt_mean_results excess_returns_means.dropna(inplace=True) excess_returns_means.tail()
上表计算结果，给出了2006年开始，每天进行量价因子十分位选股后，每个分组内股票在未来一个月相对于市场平均收益的超额收益均值
注意：十分位分组中，量价因子由小到大排序，即第一组为量价因子最小的组
下图展示，量价因子十分位选股后，在未来一个月各个分组的超额收益，可以发现：因子多空收益明显，且因子空头收益更强
fig = plt.figure(figsize=(12, 6)) ax1 = fig.add_subplot(111) excess_returns_means_dist = excess_returns_means.mean() # lns1 = ax1.plot(excess_returns_means_dist.index, excess_returns_means_dist.values, '--o', label='IC') excess_dist_plus = excess_returns_means_dist[excess_returns_means_dist>0] excess_dist_minus = excess_returns_means_dist[excess_returns_means_dist<0] lns2 = ax1.bar(excess_dist_plus.index, excess_dist_plus.values, align='center', color='r', width=0.35) lns3 = ax1.bar(excess_dist_minus.index, excess_dist_minus.values, align='center', color='g', width=0.35) ax1.set_xlim(left=0.5, right=len(excess_returns_means_dist)+0.5) ax1.set_ylim(-0.01, 0.004) ax1.set_ylabel(u'超额收益', fontproperties=font, fontsize=16) ax1.set_xlabel(u'十分位分组', fontproperties=font, fontsize=16) ax1.set_xticks(excess_returns_means_dist.index) ax1.set_xticklabels([int(x) for x in ax1.get_xticks()], fontproperties=font, fontsize=14) ax1.set_yticklabels([str(x*100)+'0%' for x in ax1.get_yticks()], fontproperties=font, fontsize=14) ax1.set_title(u"量价相关性选股因子超额收益", fontproperties=font, fontsize=16) ax1.grid()4.2 量价因子选股的市值分布特征
n_quantile = 10 # 和海通研报一样，统计十分位数 cols_mean = [i+1 for i in range(n_quantile)] cols = cols_mean mkt_value_means = pd.DataFrame(index=corr_data.index, columns=cols) # 计算相关系数分组的超额收益平均值 for dt in mkt_value_means.index: qt_mean_results = [] # 相关系数去掉nan和绝对值大于1的 tmp_corr = corr_data.ix[dt].dropna() tmp_corr = tmp_corr[(tmp_corr<=1.0) & (tmp_corr>=-1.0)] tmp_mkt_value = mkt_value_data.ix[dt].dropna() tmp_mkt_value = tmp_mkt_value.rank()/len(tmp_mkt_value) pct_quantiles = 1.0/n_quantile for i in range(n_quantile): down = tmp_corr.quantile(pct_quantiles*i) up = tmp_corr.quantile(pct_quantiles*(i+1)) i_quantile_index = tmp_corr[(tmp_corr<=up) & (tmp_corr>=down)].index mean_tmp = tmp_mkt_value[i_quantile_index].mean() qt_mean_results.append(mean_tmp) mkt_value_means.ix[dt] = qt_mean_results mkt_value_means.dropna(inplace=True) mkt_value_means.tail()
上表计算结果，给出了2006年开始，每天进行量价因子十分位选股后，每个分组内股票的市值百分位均值
下图展示，量价因子十分位选股后，各个分组的市值百分位历史均值：量价因子有略微的大市值暴露，与市值因子负相关
fig = plt.figure(figsize=(12, 6)) ax1 = fig.add_subplot(111) ax2 = ax1.twinx() mkt_value_means_dist = mkt_value_means.mean() lns1 = ax1.bar(mkt_value_means_dist.index, mkt_value_means_dist.values, align='center', width=0.35) lns2 = ax2.plot(excess_returns_means_dist.index, excess_returns_means_dist.values, 'o-r') ax1.legend(lns1, ['market value(left axis)'], loc=2, fontsize=12) ax2.legend(lns2, ['excess return(right axis)'], fontsize=12) ax1.set_ylim(0.4, 0.6) ax2.set_ylim(-0.01, 0.004) ax1.set_xlim(left=0.5, right=len(mkt_value_means_dist)+0.5) ax1.set_ylabel(u'市值百分位数', fontproperties=font, fontsize=16) ax2.set_ylabel(u'超额收益', fontproperties=font, fontsize=16) ax1.set_xlabel(u'十分位分组', fontproperties=font, fontsize=16) ax1.set_xticks(mkt_value_means_dist.index) ax1.set_xticklabels([int(x) for x in ax1.get_xticks()], fontproperties=font, fontsize=14) ax1.set_yticklabels([str(x*100)+'%' for x in ax1.get_yticks()], fontproperties=font, fontsize=14) ax2.set_yticklabels([str(x*100)+'0%' for x in ax2.get_yticks()], fontproperties=font, fontsize=14) ax1.set_title(u"量价相关性选股因子市值分布特征", fontproperties=font, fontsize=16) ax1.grid()
4.3 量价因子选股的换手率分布特征
n_quantile = 10 # 和海通研报一样，统计十分位数 cols_mean = [i+1 for i in range(n_quantile)] cols = cols_mean turnover_rate_means = pd.DataFrame(index=corr_data.index, columns=cols) # 计算相关系数分组的超额收益平均值 for dt in turnover_rate_means.index: qt_mean_results = [] # 相关系数去掉nan和绝对值大于1的 tmp_corr = corr_data.ix[dt].dropna() tmp_corr = tmp_corr[(tmp_corr<=1.0) & (tmp_corr>=-1.0)] tmp_turnover_rate = turnover_rate_data.ix[dt].dropna() pct_quantiles = 1.0/n_quantile for i in range(n_quantile): down = tmp_corr.quantile(pct_quantiles*i) up = tmp_corr.quantile(pct_quantiles*(i+1)) i_quantile_index = tmp_corr[(tmp_corr<=up) & (tmp_corr>=down)].index mean_tmp = tmp_turnover_rate[i_quantile_index].mean() qt_mean_results.append(mean_tmp) turnover_rate_means.ix[dt] = qt_mean_results turnover_rate_means.dropna(inplace=True) turnover_rate_means.tail()
上表计算结果，给出了2006年开始，每天进行量价因子十分位选股后，每个分组内股票的前一个月日均换手率的均值
下图展示，量价因子十分位选股后，各个分组的1个月日均换手率均值：量价因子对于低换手率有一定风险暴露，换手率随组别上升而逐渐升高
fig = plt.figure(figsize=(12, 6)) ax1 = fig.add_subplot(111) ax2 = ax1.twinx() turnover_rate_means_dist = turnover_rate_means.mean() lns1 = ax1.bar(turnover_rate_means_dist.index, turnover_rate_means_dist.values, align='center', width=0.35) lns2 = ax2.plot(excess_returns_means_dist.index, excess_returns_means_dist.values, 'o-r') ax1.legend(lns1, ['turnover rate(left axis)'], loc=2, fontsize=12) ax2.legend(lns2, ['excess return(right axis)'], fontsize=12) ax1.set_ylim(0, 0.05) ax2.set_ylim(-0.01, 0.004) ax1.set_xlim(left=0.5, right=len(turnover_rate_means_dist)+0.5) ax1.set_ylabel(u'换手率', fontproperties=font, fontsize=16) ax2.set_ylabel(u'超额收益', fontproperties=font, fontsize=16) ax1.set_xlabel(u'十分位分组', fontproperties=font, fontsize=16) ax1.set_xticks(turnover_rate_means_dist.index) ax1.set_xticklabels([int(x) for x in ax1.get_xticks()], fontproperties=font, fontsize=14) ax1.set_yticklabels([str(x*100)+'%' for x in ax1.get_yticks()], fontproperties=font, fontsize=14) ax2.set_yticklabels([str(x*100)+'0%' for x in ax2.get_yticks()], fontproperties=font, fontsize=14) ax1.set_title(u"量价相关性选股因子换手率分布特征", fontproperties=font, fontsize=16) ax1.grid()
4.4 量价因子选股的一个月反转分布特征
n_quantile = 10 # 和海通研报一样，统计十分位数 cols_mean = [i+1 for i in range(n_quantile)] cols = cols_mean hist_returns_means = pd.DataFrame(index=corr_data.index, columns=cols) # 计算相关系数分组的超额收益平均值 for dt in hist_returns_means.index: qt_mean_results = [] # 相关系数去掉nan和绝对值大于1的 tmp_corr = corr_data.ix[dt].dropna() tmp_corr = tmp_corr[(tmp_corr<=1.0) & (tmp_corr>=-1.0)] tmp_return = backward_20d_return_data.ix[dt].dropna() tmp_return_mean = tmp_return.mean() pct_quantiles = 1.0/n_quantile for i in range(n_quantile): down = tmp_corr.quantile(pct_quantiles*i) up = tmp_corr.quantile(pct_quantiles*(i+1)) i_quantile_index = tmp_corr[(tmp_corr<=up) & (tmp_corr>=down)].index mean_tmp = tmp_return[i_quantile_index].mean() - tmp_return_mean qt_mean_results.append(mean_tmp) hist_returns_means.ix[dt] = qt_mean_results hist_returns_means.dropna(inplace=True) hist_returns_means.tail()
上表计算结果，给出了2006年开始，每天进行量价因子十分位选股后，每个分组内股票的前一个月超额涨幅(超出市场平均值)的均值
下图展示，量价因子十分位选股后，各个分组的前一个月超额涨幅均值：量价因子对于一个月反转因子有一定风险暴露（多头组合即第一组中的股票前一个月平均跑输市场）
fig = plt.figure(figsize=(12, 6)) ax1 = fig.add_subplot(111) ax2 = ax1.twinx() hist_returns_means_dist = hist_returns_means.mean() lns1 = ax1.bar(hist_returns_means_dist.index, hist_returns_means_dist.values, align='center', width=0.35) lns2 = ax2.plot(excess_returns_means_dist.index, excess_returns_means_dist.values, 'o-r') ax1.legend(lns1, ['20 day return(left axis)'], loc=2, fontsize=12) ax2.legend(lns2, ['excess return(right axis)'], fontsize=12) ax1.set_ylim(-0.03, 0.07) ax2.set_ylim(-0.01, 0.004) ax1.set_xlim(left=0.5, right=len(hist_returns_means_dist)+0.5) ax1.set_ylabel(u'历史一个月收益率', fontproperties=font, fontsize=16) ax2.set_ylabel(u'超额收益', fontproperties=font, fontsize=16) ax1.set_xlabel(u'十分位分组', fontproperties=font, fontsize=16) ax1.set_xticks(hist_returns_means_dist.index) ax1.set_xticklabels([int(x) for x in ax1.get_xticks()], fontproperties=font, fontsize=14) ax1.set_yticklabels([str(x*100)+'%' for x in ax1.get_yticks()], fontproperties=font, fontsize=14) ax2.set_yticklabels([str(x*100)+'0%' for x in ax2.get_yticks()], fontproperties=font, fontsize=14) ax1.set_title(u"量价相关性选股因子一个月历史收益率（一个月反转因子）分布特征", fontproperties=font, fontsize=16) ax1.grid()4.5 量价因子选股的三个月反转分布特征
n_quantile = 10 # 和海通研报一样，统计十分位数 cols_mean = [i+1 for i in range(n_quantile)] cols = cols_mean hist_returns_means = pd.DataFrame(index=corr_data.index, columns=cols) # 计算相关系数分组的超额收益平均值 for dt in hist_returns_means.index: qt_mean_results = [] # 相关系数去掉nan和绝对值大于1的 tmp_corr = corr_data.ix[dt].dropna() tmp_corr = tmp_corr[(tmp_corr<=1.0) & (tmp_corr>=-1.0)] tmp_return = backward_60d_return_data.ix[dt].dropna() tmp_return_mean = tmp_return.mean() pct_quantiles = 1.0/n_quantile for i in range(n_quantile): down = tmp_corr.quantile(pct_quantiles*i) up = tmp_corr.quantile(pct_quantiles*(i+1)) i_quantile_index = tmp_corr[(tmp_corr<=up) & (tmp_corr>=down)].index mean_tmp = tmp_return[i_quantile_index].mean() - tmp_return_mean qt_mean_results.append(mean_tmp) hist_returns_means.ix[dt] = qt_mean_results hist_returns_means.dropna(inplace=True) hist_returns_means.tail()
上表计算结果，给出了2006年开始，每天进行量价因子十分位选股后，每个分组内股票的前三个月超额涨幅(超出市场平均值)的均值
下图展示，量价因子十分位选股后，各个分组的前三个月超额涨幅均值：股票分组在三个月涨幅上的分布并未呈现出明显的单调性，仅呈现出“两头高，中间低”的特点
fig = plt.figure(figsize=(12, 6)) ax1 = fig.add_subplot(111) ax2 = ax1.twinx() hist_returns_means_dist = hist_returns_means.mean() lns1 = ax1.bar(hist_returns_means_dist.index, hist_returns_means_dist.values, align='center', width=0.35) lns2 = ax2.plot(excess_returns_means_dist.index, excess_returns_means_dist.values, 'o-r') ax1.legend(lns1, ['60 day return(left axis)'], loc=2, fontsize=12) ax2.legend(lns2, ['excess return(right axis)'], fontsize=12) ax1.set_ylim(-0.02, 0.04) ax2.set_ylim(-0.01, 0.004) ax1.set_xlim(left=0.5, right=len(hist_returns_means_dist)+0.5) ax1.set_ylabel(u'历史三个月收益率', fontproperties=font, fontsize=16) ax2.set_ylabel(u'超额收益', fontproperties=font, fontsize=16) ax1.set_xlabel(u'十分位分组', fontproperties=font, fontsize=16) ax1.set_xticks(hist_returns_means_dist.index) ax1.set_xticklabels([int(x) for x in ax1.get_xticks()], fontproperties=font, fontsize=14) ax1.set_yticklabels([str(x*100)+'%' for x in ax1.get_yticks()], fontproperties=font, fontsize=14) ax2.set_yticklabels([str(x*100)+'0%' for x in ax2.get_yticks()], fontproperties=font, fontsize=14) ax1.set_title(u"量价相关性选股因子三个月历史收益率（三个月反转因子）分布特征", fontproperties=font, fontsize=16) ax1.grid()
5. 量价因子历史回测净值表现
回测时段为2010年1月1日至2016年8月1日
股票池为A股全部股票
组合每15个交易日调仓，交易费率设为双边万分之二
调仓时，涨停、停牌不买入，跌停、停牌不卖出；
每月底调仓时，选择股票池中量价因子最小的20%的股票；
5.1 量价因子最小20%股票
start = '2010-01-01' # 回测起始时间 end = '2016-08-01' # 回测结束时间 benchmark = 'ZZ500' # 策略参考标准 universe = set_universe('A') # 证券池，支持股票和基金 capital_base = 10000000 # 起始资金 freq = 'd' # 策略类型，'d'表示日间策略使用日线回测 refresh_rate = 15 # 调仓频率，表示执行handle_data的时间间隔 corr_data = pd.read_csv('VolPriceCorr_W15_FullA.csv') # 读取量价因子数据 corr_data = corr_data[corr_data.columns[1:]].set_index('tradeDate') corr_dates = corr_data.index.values quantile_five = 1 # 选取股票的量价因子五分位数，1表示选取股票池中因子最小的10%的股票 commission = Commission(0.0002,0.0002) # 交易费率设为双边万分之二 def initialize(account): # 初始化虚拟账户状态 pass def handle_data(account): # 每个交易日的买入卖出指令 pre_date = account.previous_date.strftime("%Y-%m-%d") if pre_date not in corr_dates: # 只在计算过量价因子的交易日调仓 return # 拿取调仓日前一个交易日的量价因子，并按照相应十分位选择股票 pre_corr = corr_data.ix[pre_date] pre_corr = pre_corr.dropna() pre_corr = pre_corr[(pre_corr<=1.0) & (pre_corr>=-1.0)] pre_corr_min = pre_corr.quantile((quantile_five-1)*0.2) pre_corr_max = pre_corr.quantile(quantile_five*0.2) my_univ = pre_corr[pre_corr>=pre_corr_min][pre_corr<pre_corr_max].index.values # 调仓逻辑 univ = [x for x in my_univ if x in account.universe] # 不在股票池中的，清仓 for stk in account.valid_secpos: if stk not in univ: order_to(stk, 0) # 在目标股票池中的，等权买入 for stk in univ: order_pct_to(stk, 1.1/len(univ))
bt_all = {} # 用来保存三个策略运行结果：量价因子，20日反转因子，量价因子与20日反转因子等权重叠加 bt_all['corr'] = bt # 保存量价因子回测结果5.2 一个月反转因子最小（近一个月涨幅最低的）20%股票
start = '2010-01-01' # 回测起始时间 end = '2016-08-01' # 回测结束时间 benchmark = 'ZZ500' # 策略参考标准 universe = set_universe('A') # 证券池，支持股票和基金 capital_base = 10000000 # 起始资金 freq = 'd' # 策略类型，'d'表示日间策略使用日线回测 refresh_rate = 15 # 调仓频率，表示执行handle_data的时间间隔 revs_data = pd.read_csv('BackwardReturns_W20_FullA.csv') # 读取反转因子数据 revs_data = revs_data[revs_data.columns[1:]].set_index('tradeDate') revs_dates = revs_data.index.values quantile_five = 1 # 选取股票的20日反转因子的五分位数，1表示选取股票池中因子最小的20%的股票 commission = Commission(0.0002,0.0002) # 交易费率设为双边万分之二 def initialize(account): # 初始化虚拟账户状态 pass def handle_data(account): # 每个交易日的买入卖出指令 pre_date = account.previous_date.strftime("%Y-%m-%d") if pre_date not in revs_dates: # 只在计算过反转因子的交易日调仓 return # 拿取调仓日前一个交易日的反转因子，并按照相应十分位选择股票 pre_revs = revs_data.ix[pre_date] pre_revs = pre_revs.dropna() pre_revs_min = pre_revs.quantile((quantile_five-1)*0.2) pre_revs_max = pre_revs.quantile(quantile_five*0.2) my_univ = pre_revs[pre_revs>=pre_revs_min][pre_revs<pre_revs_max].index.values # 调仓逻辑 univ = [x for x in my_univ if x in account.universe] # 不在股票池中的，清仓 for stk in account.valid_secpos: if stk not in univ: order_to(stk, 0) # 在目标股票池中的，等权买入 for stk in univ: order_pct_to(stk, 1.1/len(univ))
bt_all['revs'] = bt # 保存一个月反转因子回测结果5.3 量价因子叠加反转因子选股
量价因子和反转因子分别标准化，之后相加生成叠加因子，选叠加因子最小的20%股票
start = '2010-01-01' # 回测起始时间 end = '2016-08-01' # 回测结束时间 benchmark = 'ZZ500' # 策略参考标准 universe = set_universe('A') # 证券池，支持股票和基金 capital_base = 10000000 # 起始资金 freq = 'd' # 策略类型，'d'表示日间策略使用日线回测 refresh_rate = 15 # 调仓频率，表示执行handle_data的时间间隔 corr_data = pd.read_csv('VolPriceCorr_W15_FullA.csv') # 读取量价因子数据 corr_data = corr_data[corr_data.columns[1:]].set_index('tradeDate') corr_dates = corr_data.index.values revs_data = pd.read_csv('BackwardReturns_W20_FullA.csv') # 读取反转因子数据 revs_data = revs_data[revs_data.columns[1:]].set_index('tradeDate') quantile_five = 1 # 选取股票的因子五分位数，1表示选取股票池中因子最小的20%的股票 commission = Commission(0.0002,0.0002) # 交易费率设为双边万分之二 def initialize(account): # 初始化虚拟账户状态 pass def handle_data(account): # 每个交易日的买入卖出指令 pre_date = account.previous_date.strftime("%Y-%m-%d") if pre_date not in corr_dates: # 只在计算过量价因子的交易日调仓 return # 拿取调仓日前一个交易日的量价因子和反转因子，并按照相应分位选择股票 pre_corr = corr_data.ix[pre_date] pre_corr = pre_corr[(pre_corr<=1.0) & (pre_corr>=-1.0)] pre_revs = revs_data.ix[pre_date] # 量价因子和反转因子只做简单的等权叠加 pre_data = pd.Series(standardize(pre_corr.to_dict())) + pd.Series(standardize(pre_revs.to_dict())) # 因子标准化使用了uqer的函数standardize pre_data = pre_data.dropna() pre_data_min = pre_data.quantile((quantile_five-1)*0.2) pre_data_max = pre_data.quantile(quantile_five*0.2) my_univ = pre_data[pre_data>=pre_data_min][pre_data<pre_data_max].index.values # 调仓逻辑 univ = [x for x in my_univ if x in account.universe] # 不在股票池中的，清仓 for stk in account.valid_secpos: if stk not in univ: order_to(stk, 0) # 在目标股票池中的，等权买入 for stk in univ: order_pct_to(stk, 1.1/len(univ))
bt_all['corr + revs'] = bt5.4 上述三个组合对比
fig = plt.figure(figsize=(10,8)) fig.set_tight_layout(True) ax1 = fig.add_subplot(211) ax2 = fig.add_subplot(212) ax1.grid() ax2.grid() for qt in ['corr','revs','corr + revs']: bt = results[qt]['bt'] data = bt[[u'tradeDate',u'portfolio_value',u'benchmark_return']] data['portfolio_return'] = data.portfolio_value/data.portfolio_value.shift(1) - 1.0 # 总头寸每日回报率 data['portfolio_return'].ix[0] = data['portfolio_value'].ix[0]/ 10000000.0 - 1.0 data['excess_return'] = data.portfolio_return - data.benchmark_return # 总头寸每日超额回报率 data['excess'] = data.excess_return + 1.0 data['excess'] = data.excess.cumprod() # 总头寸对冲指数后的净值序列 data['portfolio'] = data.portfolio_return + 1.0 data['portfolio'] = data.portfolio.cumprod() # 总头寸不对冲时的净值序列 data['benchmark'] = data.benchmark_return + 1.0 data['benchmark'] = data.benchmark.cumprod() # benchmark的净值序列 results[qt]['hedged_max_drawdown'] = max([1 - v/max(1, max(data['excess'][:i+1])) for i,v in enumerate(data['excess'])]) # 对冲后净值最大回撤 results[qt]['hedged_volatility'] = np.std(data['excess_return'])*np.sqrt(252) results[qt]['hedged_annualized_return'] = (data['excess'].values[-1])**(252.0/len(data['excess'])) - 1.0 # data[['portfolio','benchmark','excess']].plot(figsize=(12,8)) # ax.plot(data[['portfolio','benchmark','excess']], label=str(qt)) ax1.plot(data['tradeDate'], data[['portfolio']], label=str(qt)) ax2.plot(data['tradeDate'], data[['excess']], label=str(qt)) ax1.legend(loc=0, fontsize=12) ax2.legend(loc=0, fontsize=12) ax1.set_ylabel(u"净值", fontproperties=font, fontsize=16) ax2.set_ylabel(u"对冲净值", fontproperties=font, fontsize=16) ax1.set_title(u"量价因子和反转因子选股能力对比 - 净值走势", fontproperties=font, fontsize=16) ax2.set_title(u"量价因子和反转因子选股能力对比 - 对冲中证500指数后净值走势", fontproperties=font, fontsize=16)
蓝色曲线为量价因子，绿色为反转因子，红色为量价因子叠加反转因子
量价因子的漫长的熊市中走势稳健，并一直打败反转因子
反转因子在15年之后表现出色
量价因子叠加反转因子，能起到意想不到的叠加效果
5.5 量价因子选股 —— 不同五分位数组合回测走势比较
# 可编辑部分与 strategy 模式一样，其余部分按本例代码编写即可 # -----------回测参数部分开始，可编辑------------ start = '2010-01-01' # 回测起始时间 end = '2016-08-01' # 回测结束时间 benchmark = 'ZZ500' # 策略参考标准 universe = set_universe('A') # 证券池，支持股票和基金 capital_base = 10000000 # 起始资金 freq = 'd' # 策略类型，'d'表示日间策略使用日线回测 refresh_rate = 15 # 调仓频率，表示执行handle_data的时间间隔 corr_data = pd.read_csv('VolPriceCorr_W15_FullA.csv') # 读取量价因子数据 corr_data = corr_data[corr_data.columns[1:]].set_index('tradeDate') corr_dates = corr_data.index.values # ---------------回测参数部分结束---------------- # 把回测参数封装到 SimulationParameters 中，供 quick_backtest 使用 sim_params = quartz.SimulationParameters(start, end, benchmark, universe, capital_base) # 获取回测行情数据 idxmap, data = quartz.get_daily_data(sim_params) # 运行结果 results_corr = {} # 调整参数(选取股票的量价因子五分位数)，进行快速回测 for quantile_five in range(1, 6): # ---------------策略逻辑部分---------------- commission = Commission(0.0002,0.0002) # 交易费率设为双边万分之二 def initialize(account): # 初始化虚拟账户状态 pass def handle_data(account): # 每个交易日的买入卖出指令 pre_date = account.previous_date.strftime("%Y-%m-%d") if pre_date not in corr_dates: # 只在计算过量价因子的交易日调仓 return # 拿取调仓日前一个交易日的量价因子，并按照相应十分位选择股票 pre_corr = corr_data.ix[pre_date] pre_corr = pre_corr.dropna() pre_corr = pre_corr[(pre_corr<=1.0) & (pre_corr>=-1.0)] pre_corr_min = pre_corr.quantile((quantile_five-1)*0.2) pre_corr_max = pre_corr.quantile(quantile_five*0.2) my_univ = pre_corr[pre_corr>=pre_corr_min][pre_corr<pre_corr_max].index.values # 调仓逻辑 univ = [x for x in my_univ if x in account.universe] # 不在股票池中的，清仓 for stk in account.valid_secpos: if stk not in univ: order_to(stk, 0) # 在目标股票池中的，等权买入 for stk in univ: order_pct_to(stk, 1.1/len(univ)) # ---------------策略逻辑部分结束---------------- # 把回测逻辑封装到 TradingStrategy 中，供 quick_backtest 使用 strategy = quartz.TradingStrategy(initialize, handle_data) # 回测部分 bt, acct = quartz.quick_backtest(sim_params, strategy, idxmap, data, refresh_rate=refresh_rate, commission=commission) # 对于回测的结果，可以通过 perf_parse 函数计算风险指标 perf = quartz.perf_parse(bt, acct) # 保存运行结果 tmp = {} tmp['bt'] = bt tmp['annualized_return'] = perf['annualized_return'] tmp['volatility'] = perf['volatility'] tmp['max_drawdown'] = perf['max_drawdown'] tmp['alpha'] = perf['alpha'] tmp['beta'] = perf['beta'] tmp['sharpe'] = perf['sharpe'] tmp['information_ratio'] = perf['information_ratio'] results_corr[quantile_five] = tmp print str(quantile_five), print 'done'
fig = plt.figure(figsize=(10,8)) fig.set_tight_layout(True) ax1 = fig.add_subplot(211) ax2 = fig.add_subplot(212) ax1.grid() ax2.grid() for qt in results_corr: bt = results_corr[qt]['bt'] data = bt[[u'tradeDate',u'portfolio_value',u'benchmark_return']] data['portfolio_return'] = data.portfolio_value/data.portfolio_value.shift(1) - 1.0 # 总头寸每日回报率 data['portfolio_return'].ix[0] = data['portfolio_value'].ix[0]/ 10000000.0 - 1.0 data['excess_return'] = data.portfolio_return - data.benchmark_return # 总头寸每日超额回报率 data['excess'] = data.excess_return + 1.0 data['excess'] = data.excess.cumprod() # 总头寸对冲指数后的净值序列 data['portfolio'] = data.portfolio_return + 1.0 data['portfolio'] = data.portfolio.cumprod() # 总头寸不对冲时的净值序列 data['benchmark'] = data.benchmark_return + 1.0 data['benchmark'] = data.benchmark.cumprod() # benchmark的净值序列 results_corr[qt]['hedged_max_drawdown'] = max([1 - v/max(1, max(data['excess'][:i+1])) for i,v in enumerate(data['excess'])]) # 对冲后净值最大回撤 results_corr[qt]['hedged_volatility'] = np.std(data['excess_return'])*np.sqrt(252) results_corr[qt]['hedged_annualized_return'] = (data['excess'].values[-1])**(252.0/len(data['excess'])) - 1.0 # data[['portfolio','benchmark','excess']].plot(figsize=(12,8)) # ax.plot(data[['portfolio','benchmark','excess']], label=str(qt)) ax1.plot(data['tradeDate'], data[['portfolio']], label=str(qt)) ax2.plot(data['tradeDate'], data[['excess']], label=str(qt)) ax1.legend(loc=0, fontsize=12) ax2.legend(loc=0, fontsize=12) ax1.set_ylabel(u"净值", fontproperties=font, fontsize=16) ax2.set_ylabel(u"对冲净值", fontproperties=font, fontsize=16) ax1.set_title(u"量价因子 - 不同五分位数分组选股净值走势", fontproperties=font, fontsize=16) ax2.set_title(u"量价因子 - 不同五分位数分组选股对冲中证500指数后净值走势", fontproperties=font, fontsize=16)
# results 转换为 DataFrame import pandas results_pd = pandas.DataFrame(results_corr).T.sort_index() results_pd = results_pd[[u'alpha', u'beta', u'information_ratio', u'sharpe', u'annualized_return', u'max_drawdown', u'volatility', u'hedged_annualized_return', u'hedged_max_drawdown', u'hedged_volatility']] for col in results_pd.columns: results_pd[col] = [np.round(x, 3) for x in results_pd[col]] cols = [(u'风险指标', u'Alpha'), (u'风险指标', u'Beta'), (u'风险指标', u'信息比率'), (u'风险指标', u'夏普比率'), (u'纯股票多头时', u'年化收益'), (u'纯股票多头时', u'最大回撤'), (u'纯股票多头时', u'收益波动率'), (u'对冲后', u'年化收益'), (u'对冲后', u'最大回撤'), (u'对冲后', u'收益波动率')] results_pd.columns = pd.MultiIndex.from_tuples(cols) results_pd.index.name = u'五分位组别' results_pd5.6 量价因子叠加反转因子选股 —— 不同五分位数组合回测走势比较
量价因子和反转因子分别标准化，之后直接等权相加生成叠加因子
# 可编辑部分与 strategy 模式一样，其余部分按本例代码编写即可 # -----------回测参数部分开始，可编辑------------ start = '2010-01-01' # 回测起始时间 end = '2016-08-01' # 回测结束时间 benchmark = 'ZZ500' # 策略参考标准 universe = set_universe('A') # 证券池，支持股票和基金 capital_base = 10000000 # 起始资金 freq = 'd' # 策略类型，'d'表示日间策略使用日线回测 refresh_rate = 15 # 调仓频率，表示执行handle_data的时间间隔 corr_data = pd.read_csv('VolPriceCorr_W15_FullA.csv') # 读取量价因子数据 corr_data = corr_data[corr_data.columns[1:]].set_index('tradeDate') corr_dates = corr_data.index.values revs_data = pd.read_csv('BackwardReturns_W20_FullA.csv') # 读取反转因子数据 revs_data = revs_data[revs_data.columns[1:]].set_index('tradeDate') # ---------------回测参数部分结束---------------- # 把回测参数封装到 SimulationParameters 中，供 quick_backtest 使用 sim_params = quartz.SimulationParameters(start, end, benchmark, universe, capital_base) # 获取回测行情数据 idxmap, data = quartz.get_daily_data(sim_params) # 运行结果 results_corrPlusRevs = {} # 调整参数(选取股票的因子五分位数)，进行快速回测 for quantile_five in range(1, 6): # ---------------策略逻辑部分---------------- commission = Commission(0.0002,0.0002) # 交易费率设为双边万分之二 def initialize(account): # 初始化虚拟账户状态 pass def handle_data(account): # 每个交易日的买入卖出指令 pre_date = account.previous_date.strftime("%Y-%m-%d") if pre_date not in corr_dates: # 只在计算过量价因子的交易日调仓 return # 拿取调仓日前一个交易日的量价因子和反转因子，并按照相应分位选择股票 pre_corr = corr_data.ix[pre_date] pre_corr = pre_corr[(pre_corr<=1.0) & (pre_corr>=-1.0)] pre_revs = revs_data.ix[pre_date] # 量价因子和反转因子只做简单的等权叠加 pre_data = pd.Series(standardize(pre_corr.to_dict())) + pd.Series(standardize(pre_revs.to_dict())) pre_data = pre_data.dropna() pre_data_min = pre_data.quantile((quantile_five-1)*0.2) pre_data_max = pre_data.quantile(quantile_five*0.2) my_univ = pre_data[pre_data>=pre_data_min][pre_data<pre_data_max].index.values # 调仓逻辑 univ = [x for x in my_univ if x in account.universe] # 不在股票池中的，清仓 for stk in account.valid_secpos: if stk not in univ: order_to(stk, 0) # 在目标股票池中的，等权买入 for stk in univ: order_pct_to(stk, 1.1/len(univ)) # ---------------策略逻辑部分结束---------------- # 把回测逻辑封装到 TradingStrategy 中，供 quick_backtest 使用 strategy = quartz.TradingStrategy(initialize, handle_data) # 回测部分 bt, acct = quartz.quick_backtest(sim_params, strategy, idxmap, data, refresh_rate=refresh_rate, commission=commission) # 对于回测的结果，可以通过 perf_parse 函数计算风险指标 perf = quartz.perf_parse(bt, acct) # 保存运行结果 tmp = {} tmp['bt'] = bt tmp['annualized_return'] = perf['annualized_return'] tmp['volatility'] = perf['volatility'] tmp['max_drawdown'] = perf['max_drawdown'] tmp['alpha'] = perf['alpha'] tmp['beta'] = perf['beta'] tmp['sharpe'] = perf['sharpe'] tmp['information_ratio'] = perf['information_ratio'] results_corrPlusRevs[quantile_five] = tmp print str(quantile_five), print 'done'
fig = plt.figure(figsize=(10,8)) fig.set_tight_layout(True) ax1 = fig.add_subplot(211) ax2 = fig.add_subplot(212) ax1.grid() ax2.grid() for qt in results_corrPlusRevs: bt = results_corrPlusRevs[qt]['bt'] data = bt[[u'tradeDate',u'portfolio_value',u'benchmark_return']] data['portfolio_return'] = data.portfolio_value/data.portfolio_value.shift(1) - 1.0 # 总头寸每日回报率 data['portfolio_return'].ix[0] = data['portfolio_value'].ix[0]/ 10000000.0 - 1.0 data['excess_return'] = data.portfolio_return - data.benchmark_return # 总头寸每日超额回报率 data['excess'] = data.excess_return + 1.0 data['excess'] = data.excess.cumprod() # 总头寸对冲指数后的净值序列 data['portfolio'] = data.portfolio_return + 1.0 data['portfolio'] = data.portfolio.cumprod() # 总头寸不对冲时的净值序列 data['benchmark'] = data.benchmark_return + 1.0 data['benchmark'] = data.benchmark.cumprod() # benchmark的净值序列 results_corrPlusRevs[qt]['hedged_max_drawdown'] = max([1 - v/max(1, max(data['excess'][:i+1])) for i,v in enumerate(data['excess'])]) # 对冲后净值最大回撤 results_corrPlusRevs[qt]['hedged_volatility'] = np.std(data['excess_return'])*np.sqrt(252) results_corrPlusRevs[qt]['hedged_annualized_return'] = (data['excess'].values[-1])**(252.0/len(data['excess'])) - 1.0 # data[['portfolio','benchmark','excess']].plot(figsize=(12,8)) # ax.plot(data[['portfolio','benchmark','excess']], label=str(qt)) ax1.plot(data['tradeDate'], data[['portfolio']], label=str(qt)) ax2.plot(data['tradeDate'], data[['excess']], label=str(qt)) ax1.legend(loc=0, fontsize=12) ax2.legend(loc=0, fontsize=12) ax1.set_ylabel(u"净值", fontproperties=font, fontsize=16) ax2.set_ylabel(u"对冲净值", fontproperties=font, fontsize=16) ax1.set_title(u"量价因子与反转因子等权叠加选股 - 不同五分位数分组选股净值走势", fontproperties=font, fontsize=16) ax2.set_title(u"量价因子与反转因子等权叠加选股 - 不同五分位数分组选股对冲中证500指数后净值走势", fontproperties=font, fontsize=16)
# results 转换为 DataFrame import pandas results_pd = pandas.DataFrame(results_corrPlusRevs).T.sort_index() results_pd = results_pd[[u'alpha', u'beta', u'information_ratio', u'sharpe', u'annualized_return', u'max_drawdown', u'volatility', u'hedged_annualized_return', u'hedged_max_drawdown', u'hedged_volatility']] for col in results_pd.columns: results_pd[col] = [np.round(x, 3) for x in results_pd[col]] cols = [(u'风险指标', u'Alpha'), (u'风险指标', u'Beta'), (u'风险指标', u'信息比率'), (u'风险指标', u'夏普比率'), (u'纯股票多头时', u'年化收益'), (u'纯股票多头时', u'最大回撤'), (u'纯股票多头时', u'收益波动率'), (u'对冲后', u'年化收益'), (u'对冲后', u'最大回撤'), (u'对冲后', u'收益波动率')] results_pd.columns = pd.MultiIndex.from_tuples(cols) results_pd.index.name = u'五分位组别' results_pd5.7 更长回测时间 —— 06年开始回测
量价因子和反转因子分别标准化，之后直接等权相加生成叠加因子
此处选择叠加因子最小的20%股票作为持仓组合
start = '2006-01-01' # 回测起始时间 end = '2016-08-01' # 回测结束时间 benchmark = 'ZZ500' # 策略参考标准 universe = set_universe('A') # 证券池，支持股票和基金 capital_base = 2000000 # 起始资金 freq = 'd' # 策略类型，'d'表示日间策略使用日线回测 refresh_rate = 15 # 调仓频率，表示执行handle_data的时间间隔 corr_data = pd.read_csv('VolPriceCorr_W15_FullA.csv') # 读取量价因子数据 corr_data = corr_data[corr_data.columns[1:]].set_index('tradeDate') corr_dates = corr_data.index.values revs_data = pd.read_csv('BackwardReturns_W20_FullA.csv') # 读取反转因子数据 revs_data = revs_data[revs_data.columns[1:]].set_index('tradeDate') quantile_five = 1 # 选取股票的因子五分位数，1表示选取股票池中因子最小的20%的股票 commission = Commission(0.0002,0.0002) # 交易费率设为双边万分之二 def initialize(account): # 初始化虚拟账户状态 pass def handle_data(account): # 每个交易日的买入卖出指令 pre_date = account.previous_date.strftime("%Y-%m-%d") if pre_date not in corr_dates: # 只在计算过量价因子的交易日调仓 return # 拿取调仓日前一个交易日的量价因子和反转因子，并按照相应分位选择股票 pre_corr = corr_data.ix[pre_date] pre_corr = pre_corr[(pre_corr<=1.0) & (pre_corr>=-1.0)] pre_revs = revs_data.ix[pre_date] # 量价因子和反转因子只做简单的等权叠加 pre_data = pd.Series(standardize(pre_corr.to_dict())) + pd.Series(standardize(pre_revs.to_dict())) pre_data = pre_data.dropna() pre_data_min = pre_data.quantile((quantile_five-1)*0.2) pre_data_max = pre_data.quantile(quantile_five*0.2) my_univ = pre_data[pre_data>=pre_data_min][pre_data<pre_data_max].index.values # 调仓逻辑 univ = [x for x in my_univ if x in account.universe] # 不在股票池中的，清仓 for stk in account.valid_secpos: if stk not in univ: order_to(stk, 0) # 在目标股票池中的，等权买入 for stk in univ: order_pct_to(stk, 1.1/len(univ))
fig = plt.figure(figsize=(12,5)) fig.set_tight_layout(True) ax1 = fig.add_subplot(111) ax2 = ax1.twinx() ax1.grid() bt_quantile = bt data = bt_quantile[[u'tradeDate',u'portfolio_value',u'benchmark_return']] data['portfolio_return'] = data.portfolio_value/data.portfolio_value.shift(1) - 1.0 data['portfolio_return'].ix[0] = data['portfolio_value'].ix[0]/ 2000000.0 - 1.0 data['excess_return'] = data.portfolio_return - data.benchmark_return data['excess'] = data.excess_return + 1.0 data['excess'] = data.excess.cumprod() data['portfolio'] = data.portfolio_return + 1.0 data['portfolio'] = data.portfolio.cumprod() data['benchmark'] = data.benchmark_return + 1.0 data['benchmark'] = data.benchmark.cumprod() # ax.plot(data[['portfolio','benchmark','excess']], label=str(qt)) ax1.plot(data['tradeDate'], data[['portfolio']], label='portfolio(left)') ax1.plot(data['tradeDate'], data[['benchmark']], label='benchmark(left)') ax2.plot(data['tradeDate'], data[['excess']], label='hedged(right)', color='r') ax1.legend(loc=2) ax2.legend(loc=0) ax2.set_ylim(bottom=0.5, top=5) ax1.set_ylabel(u"净值", fontproperties=font, fontsize=16) ax2.set_ylabel(u"对冲指数净值", fontproperties=font, fontsize=16) ax2.set_ylabel(u"对冲指数净值", fontproperties=font, fontsize=16) ax1.set_title(u"量价因子反转因子叠加选股的前20%股票回测走势", fontproperties=font, fontsize=16)
上图可以看到从06年起的回测结果，展示出量价因子反转因子叠加后的稳定的alpha输出
我们根据量价因子叠加反转因子选取股票组合，表现最好的组合其06年以来年化收益达到41.4%，alpha达到22.6%，beta仅为0.88，展示出稳定盈利的能力。
引用地址:https://www.gupiaohao.com/202307/36500.html
tags: