日频回测demos
由qxiao创建，最终由qxiao 被浏览 4 用户
股票日频策略一：基本面选股策略
非ST、非科创板、非退市、非停牌，流通市值大于5亿元，净利润同比>30%，营业收入同比>30%，0<市盈率<50，市净率<5，总市值从小到大排列，每5个交易日调仓，持有前20只。
from bigquant import bigtrader, dai
def initialize(context: bigtrader.IContext):
context.set_commission(bigtrader.PerOrder(buy_cost=0.0003, sell_cost=0.0013, min_cost=5))
rebalance_days = 5
stock_num = 20
sql = """
SELECT
date,
instrument,
total_market_cap,
float_market_cap,
pe_ttm,
pb,
net_profit_growth_rate,
revenue_growth_rate,
1.0/$stock_num AS weight
FROM (
SELECT
date,
instrument,
total_market_cap,
float_market_cap,
pe_ttm,
pb,
net_profit_ttm,
operating_revenue_ttm,
st_status,
suspended,
list_sector,
list_days,
(net_profit_ttm / m_lag(net_profit_ttm, 250) - 1) * 100 AS net_profit_growth_rate,
(operating_revenue_ttm / m_lag(operating_revenue_ttm, 250) - 1) * 100 AS revenue_growth_rate
FROM cn_stock_prefactors
) AS t
WHERE
st_status = 0
AND suspended = 0
AND list_sector IN (1, 2)
AND list_days > 365
AND float_market_cap > 500000000
AND pe_ttm > 0 AND pe_ttm < 50
AND pb > 0 AND pb < 5
AND net_profit_growth_rate > 30
AND revenue_growth_rate > 30
ORDER BY date, total_market_cap ASC
"""
df = dai.query(
sql,
filters={"date": [
context.add_trading_days(context.start_date, -250),
context.end_date
]},
params={"stock_num": stock_num}
).df()
df = df.groupby('date').head(stock_num)
df = bigtrader.TradingDaysRebalance(rebalance_days, context=context).select_rebalance_data(df)
context.data = df
performance = bigtrader.run(
market=bigtrader.Market.CN_STOCK,
frequency=bigtrader.Frequency.DAILY,
start_date="2024-01-01",
end_date="2026-01-21",
capital_base=1000000,
initialize=initialize,
handle_data=bigtrader.HandleDataLib.handle_data_weight_based,
)
股票日频策略二：KDJ 指标策略（农业银行）
仅 601288.SH，计算 KDJ(9,3,3) 指标，KDJ_K < 20 为买入信号（加仓至100%），KDJ_K > 80 为卖出信号（减仓至80%），首次建仓 80% 仓位，每1个交易日调仓。
from bigquant import bigtrader, dai
from datetime import datetime
def initialize(context: bigtrader.IContext):
context.set_commission(bigtrader.PerOrder(buy_cost=0.0003, sell_cost=0.0013, min_cost=5))
context.stock = '601288.SH'
context.kdj_n = 9
context.kdj_m1 = 3
context.kdj_m2 = 3
context.buy_threshold = 20
context.sell_threshold = 80
context.base_position = 0.8
context.full_position = 1.0
context.current_position_target = context.base_position
context.is_first_trade = True
sql = """
SELECT
date,
instrument,
close,
m_ta_kdj_k(high, low, close, fastk_period := $kdj_n, slowk_period := $kdj_m1, slowd_period := $kdj_m2) AS kdj_k,
m_ta_kdj_d(high, low, close, fastk_period := $kdj_n, slowk_period := $kdj_m1, slowd_period := $kdj_m2) AS kdj_d,
m_ta_kdj_j(high, low, close, fastk_period := $kdj_n, slowk_period := $kdj_m1, slowd_period := $kdj_m2) AS kdj_j
FROM cn_stock_prefactors
WHERE instrument = $stock AND suspended = 0
ORDER BY date
"""
df = dai.query(
sql,
filters={"date": [
context.add_trading_days(context.start_date, -60),
context.end_date
]},
params={"stock": context.stock, "kdj_n": context.kdj_n, "kdj_m1": context.kdj_m1, "kdj_m2": context.kdj_m2}
).df()
df = df[df['date'] >= context.start_date]
context.data = df
def handle_data(context: bigtrader.IContext, data: bigtrader.IBarData):
current_date = data.current_dt.strftime('%Y-%m-%d')
today_data = context.data[context.data['date'] == current_date]
if today_data.empty:
return
kdj_k = today_data.iloc[0]['kdj_k']
kdj_d = today_data.iloc[0]['kdj_d']
kdj_j = today_data.iloc[0]['kdj_j']
if kdj_k is None or kdj_k != kdj_k:
return
current_position = context.portfolio.positions.get(context.stock)
current_value = current_position.market_value if current_position else 0
total_value = context.portfolio.portfolio_value
current_weight = current_value / total_value if total_value > 0 else 0
target_weight = context.current_position_target
signal_desc = "持有"
if context.is_first_trade and current_weight == 0:
target_weight = context.base_position
signal_desc = "首次建仓"
context.is_first_trade = False
elif kdj_k < context.buy_threshold:
target_weight = context.full_position
signal_desc = "买入信号"
elif kdj_k > context.sell_threshold:
target_weight = context.base_position
signal_desc = "卖出信号"
context.current_position_target = target_weight
if abs(target_weight - current_weight) > 0.01:
context.order_target_percent(context.stock, target_weight)
context.logger.info(
f"{current_date}: {signal_desc} | KDJ(K={kdj_k:.2f}, D={kdj_d:.2f}, J={kdj_j:.2f}) | "
f"当前仓位: {current_weight*100:.1f}% -> 目标仓位: {target_weight*100:.1f}%"
)
performance = bigtrader.run(
market=bigtrader.Market.CN_STOCK,
frequency=bigtrader.Frequency.DAILY,
start_date="2020-01-01",
end_date=datetime.now().strftime("%Y-%m-%d"),
capital_base=1000000,
initialize=initialize,
handle_data=handle_data,
)
ETF 日频策略一：宽基 ETF 低波动策略
标的池：沪深300ETF / 中证500ETF / 科创50ETF，因子：近30个交易日日收益率标准差（波动率），每次调仓选波动率最低的2只ETF，等权配置（各50%），每10个交易日调仓一次。
from datetime import datetime
from bigquant import bigtrader, dai
def initialize(context: bigtrader.IContext):
context.set_commission(bigtrader.PerOrder(buy_cost=0.0003, sell_cost=0.0013, min_cost=5))
fund_pool = ["510300.SH", "510500.SH", "588000.SH"]
vol_window = 30
sql = """
SELECT
date,
instrument,
-m_stddev(close / m_lag(close, 1) - 1, $vol_window) AS score,
0.5 AS weight
FROM cn_fund_bar1d
WHERE instrument IN $fund_pool
QUALIFY score IS NOT NULL
ORDER BY date, instrument
"""
df = dai.query(
sql,
filters={"date": [
context.add_trading_days(context.start_date, -60),
context.end_date,
]},
params={"fund_pool": fund_pool, "vol_window": vol_window},
).df()
df = df.sort_values(["date", "score"], ascending=[True, False])
df = df.groupby("date").head(2)
df["weight"] = 0.5
df = bigtrader.TradingDaysRebalance(10, context=context).select_rebalance_data(df)
context.data = df
performance = bigtrader.run(
market=bigtrader.Market.CN_FUND,
frequency=bigtrader.Frequency.DAILY,
start_date="2020-01-01",
end_date=datetime.now().strftime("%Y-%m-%d"),
capital_base=1_000_000,
initialize=initialize,
handle_data=bigtrader.HandleDataLib.handle_data_weight_based,
order_price_field_buy="open",
order_price_field_sell="open",
)
ETF 日频策略二：宽基 ETF 动量轮动
标的池：沪深300ETF / 中证500ETF / 科创50ETF，因子：近20个交易日收益率（动量因子），每次调仓选动量最强的1只ETF全仓持有，每5个交易日调仓一次。
from datetime import datetime
from bigquant import bigtrader, dai
FUND_POOL = ["510300.SH", "510500.SH", "588000.SH"]
MOMENTUM_DAYS = 20
REBALANCE_DAYS = 5
def initialize(context: bigtrader.IContext):
context.set_commission(bigtrader.PerOrder(buy_cost=0.0003, sell_cost=0.0013, min_cost=5))
sql = """
SELECT date, instrument, close / m_lag(close, $momentum_days) - 1 AS momentum
FROM cn_fund_bar1d
WHERE instrument IN $fund_pool
QUALIFY momentum IS NOT NULL
ORDER BY date, instrument
"""
df = dai.query(
sql,
filters={"date": [
context.add_trading_days(context.start_date, -30),
context.end_date,
]},
params={"fund_pool": FUND_POOL, "momentum_days": MOMENTUM_DAYS},
).df()
df = df.sort_values(["date", "momentum"], ascending=[True, False])
df = df.groupby("date").head(1).set_index("date")
context.signal_df = df
context.rebalance_counter = 0
def handle_data(context: bigtrader.IContext, data):
context.rebalance_counter += 1
if context.rebalance_counter % REBALANCE_DAYS != 1:
return
today = data.current_dt.strftime("%Y-%m-%d")
if today not in context.signal_df.index:
return
target_instrument = context.signal_df.loc[today, "instrument"]
for instrument, position in context.portfolio.positions.items():
if position.amount > 0 and instrument != target_instrument:
context.order_target_percent(instrument, 0)
context.order_target_percent(target_instrument, 1.0)
performance = bigtrader.run(
market=bigtrader.Market.CN_FUND,
frequency=bigtrader.Frequency.DAILY,
start_date="2020-01-01",
end_date=datetime.now().strftime("%Y-%m-%d"),
capital_base=1_000_000,
initialize=initialize,
handle_data=handle_data,
order_price_field_buy="open",
order_price_field_sell="open",
)
可转债日频策略一：低溢价率 + 低价格周度轮动
筛选条件：转股溢价率 < 10%，转债价格 < 110 元，剩余余额 > 1 亿元，剩余天数 > 120 天。按转债价格升序排列，持有价格最低的 8 只，等权持仓（各 12.5%），每 5 个交易日调仓一次。
from bigquant import bigtrader, dai
from datetime import datetime
def initialize(context: bigtrader.IContext):
context.set_commission(bigtrader.PerOrder(buy_cost=0.0003, sell_cost=0.0013, min_cost=5))
context.bond_num = 8
sql = """
SELECT
m.date,
m.instrument,
b.close,
m.conversion_premium_rate,
m.bond_balance,
m.remaining_days
FROM cn_cbond_analyze_metric AS m
INNER JOIN cn_cbond_bar1d AS b USING (date, instrument)
WHERE
m.conversion_premium_rate < 10
AND b.close < 110
AND m.bond_balance > 1
AND m.remaining_days > 120
QUALIFY c_rank(b.close, ascending := true) <= $bond_num
ORDER BY m.date, b.close ASC
"""
df = dai.query(
sql,
filters={"date": [context.start_date, context.end_date]},
params={"bond_num": context.bond_num}
).df()
df["date"] = df["date"].astype(str).str[:10]
context.daily_candidates = {
date: group["instrument"].tolist()
for date, group in df.groupby("date")
}
context.rebalance_period = bigtrader.TradingDaysRebalance(5, context=context)
def handle_data(context: bigtrader.IContext, data: bigtrader.IBarData):
if not context.rebalance_period.is_signal_date(data.current_dt.date()):
return
today = data.current_dt.strftime("%Y-%m-%d")
target_instruments = context.daily_candidates.get(today, [])
weight = 1.0 / context.bond_num
target_set = set(target_instruments)
current_positions = {
pos.instrument: pos
for pos in context.portfolio.positions.values()
if pos.amount > 0
}
current_set = set(current_positions.keys())
# 先卖出不在目标池中的持仓
for instrument in current_set - target_set:
context.order_target_percent(instrument, 0)
if not target_instruments:
return
# 买入/调整目标券至固定等权仓位
for instrument in target_instruments:
context.order_target_percent(instrument, weight)
performance = bigtrader.run(
market=bigtrader.Market.CN_CBOND,
frequency=bigtrader.Frequency.DAILY,
start_date="2021-01-01",
end_date=datetime.now().strftime("%Y-%m-%d"),
capital_base=500000,
initialize=initialize,
handle_data=handle_data,
order_price_field_buy='open',
order_price_field_sell='open'
)
可转债日频策略二：双低策略
双低值 = 转股溢价率（%）+ 转债价格（元），筛选双低值 < 120 的转债，过滤剩余余额 < 1亿和剩余天数 < 120天，按双低值升序排列取前10只等权持有，每5个交易日调仓一次。
from bigquant import bigtrader, dai
from datetime import datetime
def initialize(context: bigtrader.IContext):
context.set_commission(bigtrader.PerOrder(buy_cost=0.0003, sell_cost=0.0013, min_cost=5))
rebalance_days = 5
bond_num = 10
sql = """
SELECT
date,
instrument,
close,
conversion_premium_rate,
(conversion_premium_rate + close) AS dual_low,
1.0 / $bond_num AS weight
FROM cn_cbond_analyze_metric
INNER JOIN cn_cbond_bar1d USING (date, instrument)
WHERE
(conversion_premium_rate + close) < 120
AND bond_balance > 1
AND remaining_days > 120
AND conversion_premium_rate > -10
QUALIFY c_rank(conversion_premium_rate + close, ascending := true) <= $bond_num
ORDER BY date, dual_low ASC
"""
df = dai.query(
sql,
filters={"date": [
context.add_trading_days(context.start_date, -30),
context.end_date
]},
params={"bond_num": bond_num}
).df()
df = bigtrader.TradingDaysRebalance(rebalance_days, context=context).select_rebalance_data(df)
context.data = df
performance = bigtrader.run(
market=bigtrader.Market.CN_CBOND,
frequency=bigtrader.Frequency.DAILY,
start_date="2021-01-01",
end_date=datetime.now().strftime("%Y-%m-%d"),
capital_base=500000,
initialize=initialize,
handle_data=bigtrader.HandleDataLib.handle_data_weight_based,
order_price_field_buy='open',
order_price_field_sell='open'
)
\