书接上两回
开始
终于我们可以可以来进行模拟了
复杂 的我不会，所以就写个最简单的策略吧：
- 买入条件
- 比两年前是上升的
- 比一年前是上升的
- 比一个月前上升更下降的区间在 ± 0.38% 内
- 比昨天上涨 0.1% 以上 0.5% 以下
- 卖出条件
- 如果当前下跌 5% 则全卖出
进行模拟
设计模拟时间
先进行模拟时间，即回测多少天
strStartDate = '20230101'
strEndDate = '20240407'
就是从 2023年的1月1日开始进行策略，2024年4月7日结束
选股范围
strIndustry = "'银行'"
只选择银行股进行回测
中间计算
因为需要计算 2年、1年、半年等时间窗口各股票的涨跌，用 Python 比较麻烦，于是使用 SQL 进行计算
完整代码
'''
Descripttion: 就简单回测
Version: 1.0
Author: BerryBC
Date: 2024-02-18 12:09:31
LastEditors: BerryBC
LastEditTime: 2024-04-07 11:07:43
'''
import datetime
from Lib.DBConn import claDBConn
from Lib.BackTestingDailyDeal import claBackTestingDailyDeal
btDaily = claBackTestingDailyDeal('研究银行股', '根据时间窗口看追涨', 100000)
dbConn = claDBConn()
strStartDate = '20240101'
strEndDate = '20240407'
strStock = '601988.SH'
strMaxDate = ''
# 时间序列
strSQLTime = "SELECT * FROM tb_calendar WHERE trade_date >=" + \
strStartDate+" AND trade_date <= "+strEndDate+" ORDER BY trade_date "
listCal = dbConn.ReturnSQLData(strSQLTime)
# 打印快照
btDaily.AssetSnapshot(listCal[0][0])
# 循环验证
for tubDay in listCal:
# 计算日期
strNowDate = tubDay[0]
strMaxDate=strNowDate
'''
.::::.
.::::::::.
:::::::::::
..:::::::::::'
'::::::::::::'
.::::::::::
'::::::::::::::..
..::::::::::::.
``::::::::::::::::
::::``:::::::::' .:::.
::::' ':::::' .::::::::.
.::::' :::: .:::::::'::::.
.:::' ::::: .:::::::::' ':::::.
.::' :::::.:::::::::' ':::::.
.::' ::::::::::::::' ``::::.
...::: ::::::::::::' ``::.
````':. ':::::::::' ::::..
'.:::::' ':'````..
'''
# 追涨杀跌
# # 每日数据
# strSQLQuote = "SELECT * FROM tb_daily_stock_quote WHERE trade_date=" + \
# strNowDate+" AND ts_code = '"+strStock+"'"
# listQuote = dbConn.ReturnSQLData(strSQLQuote)
# if listQuote[0][8] > 0.2:
# btDaily.BuyStock(strNowDate, strStock, 500,'追涨')
# elif listQuote[0][8] < -0.1:
# strSQLSell = "SELECT * FROM tb_bt_daily_tmp_my_stock_pool WHERE stock_code='" + strStock+"' AND bt_ind = '"+str(btDaily.intInd)+"'"
# listSell = dbConn.ReturnSQLData(strSQLSell)
# if len(listSell) > 0 and listSell[0][0] is not None:
# btDaily.SellStock(strNowDate, strStock, listSell[0][2], '全部卖了！！！')
'''
.::::.
.::::::::.
:::::::::::
..:::::::::::'
'::::::::::::'
.::::::::::
'::::::::::::::..
..::::::::::::.
``::::::::::::::::
::::``:::::::::' .:::.
::::' ':::::' .::::::::.
.::::' :::: .:::::::'::::.
.:::' ::::: .:::::::::' ':::::.
.::' :::::.:::::::::' ':::::.
.::' ::::::::::::::' ``::::.
...::: ::::::::::::' ``::.
````':. ':::::::::' ::::..
'.:::::' ':'````..
'''
# 长期来看是否可以买
# 针对某个行业的股票
listPeriod = [720,360,180,90,30,5]
strIndustry = "'银行'"
strEndDate ="DATE'"+(datetime.datetime.strptime(strNowDate, "%Y%m%d")).strftime('%Y-%m-%d')+"'"
# 计算趋势中间表
strSQLTruncate = "TRUNCATE TABLE TB_BT_MID_DAILY_PERIOD_TREND"
dbConn.RunSQLOnce(strSQLTruncate)
# 根据不同时间长度，如2年、1年、半年等计算股价升降
for intCalPeriod in listPeriod:
strSQLInsert ='''
-- CREATE TABLE TB_BT_MID_DAILY_PERIOD_TREND as
INSERT INTO TB_BT_MID_DAILY_PERIOD_TREND
WITH CHECK_STOCK_LIST AS (
-- 获取行业对应的所有股票
SELECT
TS_CODE,
NAME,
INDUSTRY,
EXCHANGE
FROM
TB_ALL_STOCK_LIST
WHERE
1=1
AND INDUSTRY = '''+strIndustry+'''
)
SELECT
'''+str(intCalPeriod)+''' AS DAY_GAP,
'''+strEndDate+''' AS END_DATE,
S_L_O.TS_CODE,
S_L_O.NAME,
S_L_O.INDUSTRY,
S_L_O.EXCHANGE,
-- 获取整体的上升下降趋势，以及简单统计
MIN(TTL_U.TRADE_DAY) AS TRADE_DAY,
MIN(TTL_U.MIN_COLSE) AS MIN_COLSE,
MIN(TTL_U.MAX_COLSE) AS MAX_COLSE,
MIN(TTL_U.MIN_PCT_CHG) AS MIN_PCT_CHG,
MIN(TTL_U.MAX_PCT_CHG) AS MAX_PCT_CHG,
MIN(TTL_U.DOWN_DAY) AS DOWN_DAY,
MIN(TTL_U.UP_DAY) AS UP_DAY,
-- 期间内的最早收盘价以及最终收盘价
MIN(TTL_U.BEGIN_CLOSE) AS BEGIN_CLOSE,
MIN(TTL_U.END_CLOSE) AS END_CLOSE,
(MIN(TTL_U.END_CLOSE) / MIN(TTL_U.BEGIN_CLOSE) - 1)*100 AS ALL_PCT_CHG,
-- 期间内涨跌的天数占比
MIN(TTL_U.UP_DAY) / MIN(TTL_U.TRADE_DAY) AS UP_DAY_PCT,
MIN(TTL_U.DOWN_DAY) / MIN(TTL_U.TRADE_DAY) AS DOWN_DAY_PCT,
MIN(TTL_U.END_CLOSE_PCT_C) AS END_CLOSE_PCT_C
FROM
(
(
-- 整体统计趋势
SELECT
S_DQ.TS_CODE,
COUNT(1) AS TRADE_DAY,
MIN(S_DQ.CLOSE) AS MIN_COLSE,
MAX(S_DQ.CLOSE) AS MAX_COLSE,
MIN(S_DQ.PCT_CHG) AS MIN_PCT_CHG,
MAX(S_DQ.PCT_CHG) AS MAX_PCT_CHG,
COUNT(
CASE
WHEN S_DQ.PCT_CHG < 0 THEN 1
ELSE NULL
END
) AS DOWN_DAY,
COUNT(
CASE
WHEN S_DQ.PCT_CHG > 0 THEN 1
ELSE NULL
END
) AS UP_DAY,
NULL AS BEGIN_CLOSE,
NULL AS END_CLOSE,
NULL AS END_CLOSE_PCT_C
FROM
TB_CALENDAR S_T
INNER JOIN TB_DAILY_STOCK_QUOTE S_DQ
ON S_T.TRADE_DATE = S_DQ.TRADE_DATE
INNER JOIN CHECK_STOCK_LIST S_LIST
ON S_DQ.TS_CODE = S_LIST.TS_CODE
WHERE
1=1
AND S_T.TRADE_DATE_DT < '''+strEndDate+'''
AND S_T.TRADE_DATE_DT >= DATE_SUB('''+strEndDate+''', INTERVAL '''+str(intCalPeriod)+''' DAY)
GROUP BY
S_DQ.TS_CODE
)
UNION ALL
(
-- 计算时间窗口内最早那天的收盘价
SELECT
MIN_MID.TS_CODE,
NULL AS TRADE_DAY,
NULL AS MIN_COLSE,
NULL AS MAX_COLSE,
NULL AS MIN_PCT_CHG,
NULL AS MAX_PCT_CHG,
NULL AS DOWN_DAY,
NULL AS UP_DAY,
MIN_MID.CLOSE AS BEGIN_CLOSE,
NULL AS END_CLOSE,
NULL AS END_CLOSE_PCT_C
FROM
(
SELECT
S_DQ_MIN.TS_CODE,
S_DQ_MIN.CLOSE,
RANK() OVER(PARTITION BY S_DQ_MIN.TS_CODE ORDER BY S_T_MIN.TRADE_DATE_DT ASC) AS RNK
FROM
TB_CALENDAR S_T_MIN
INNER JOIN TB_DAILY_STOCK_QUOTE S_DQ_MIN
ON S_T_MIN.TRADE_DATE = S_DQ_MIN.TRADE_DATE
INNER JOIN CHECK_STOCK_LIST S_LIST
ON S_DQ_MIN.TS_CODE = S_LIST.TS_CODE
WHERE
1=1
AND S_T_MIN.TRADE_DATE_DT < '''+strEndDate+'''
AND S_T_MIN.TRADE_DATE_DT >= DATE_SUB('''+strEndDate+''', INTERVAL '''+str(intCalPeriod)+''' DAY)
) MIN_MID
WHERE
MIN_MID.RNK = 1
)
UNION ALL
(
-- 计算当前收盘价
SELECT
MAX_MID.TS_CODE,
NULL AS TRADE_DAY,
NULL AS MIN_COLSE,
NULL AS MAX_COLSE,
NULL AS MIN_PCT_CHG,
NULL AS MAX_PCT_CHG,
NULL AS DOWN_DAY,
NULL AS UP_DAY,
NULL AS BEGIN_CLOSE,
MAX_MID.CLOSE AS END_CLOSE,
MAX_MID.PCT_CHG AS END_CLOSE_PCT_C
FROM
(
SELECT
S_DQ_MAX.TS_CODE,
S_DQ_MAX.CLOSE,
S_DQ_MAX.PCT_CHG,
RANK() OVER(PARTITION BY S_DQ_MAX.TS_CODE ORDER BY S_T_MAX.TRADE_DATE_DT DESC) AS RNK
FROM
TB_CALENDAR S_T_MAX
INNER JOIN TB_DAILY_STOCK_QUOTE S_DQ_MAX
ON S_T_MAX.TRADE_DATE = S_DQ_MAX.TRADE_DATE
INNER JOIN CHECK_STOCK_LIST S_LIST
ON S_DQ_MAX.TS_CODE = S_LIST.TS_CODE
WHERE
1=1
AND S_T_MAX.TRADE_DATE_DT < '''+strEndDate+'''
AND S_T_MAX.TRADE_DATE_DT >= DATE_SUB('''+strEndDate+''', INTERVAL '''+str(intCalPeriod)+''' DAY)
) MAX_MID
WHERE
MAX_MID.RNK = 1
)
) TTL_U
INNER JOIN CHECK_STOCK_LIST S_L_O
ON TTL_U.TS_CODE = S_L_O.TS_CODE
GROUP BY
S_L_O.TS_CODE,
S_L_O.NAME,
S_L_O.INDUSTRY,
S_L_O.EXCHANGE
'''
dbConn.RunSQLOnce(strSQLInsert)
# 计算趋势
strSQLQuote = '''
SELECT
END_DATE,
TS_CODE,
NAME,
INDUSTRY,
EXCHANGE,
MIN(END_CLOSE) AS END_CLOSE,
MIN(
CASE
WHEN DAY_GAP = 720 THEN ALL_PCT_CHG
ELSE NULL
END
) AS ALL_PCT_CHG_720,
MIN(
CASE
WHEN DAY_GAP = 360 THEN ALL_PCT_CHG
ELSE NULL
END
) AS ALL_PCT_CHG_360,
MIN(
CASE
WHEN DAY_GAP = 180 THEN ALL_PCT_CHG
ELSE NULL
END
) AS ALL_PCT_CHG_180,
MIN(
CASE
WHEN DAY_GAP = 90 THEN ALL_PCT_CHG
ELSE NULL
END
) AS ALL_PCT_CHG_90,
MIN(
CASE
WHEN DAY_GAP = 30 THEN ALL_PCT_CHG
ELSE NULL
END
) AS ALL_PCT_CHG_30,
MIN(
CASE
WHEN DAY_GAP = 5 THEN ALL_PCT_CHG
ELSE NULL
END
) AS ALL_PCT_CHG_5,
MIN(
CASE
WHEN DAY_GAP = 720 THEN UP_DAY_PCT
ELSE NULL
END
) AS UP_DAY_PCT_720,
MIN(
CASE
WHEN DAY_GAP = 360 THEN UP_DAY_PCT
ELSE NULL
END
) AS UP_DAY_PCT_360,
MIN(
CASE
WHEN DAY_GAP = 180 THEN UP_DAY_PCT
ELSE NULL
END
) AS UP_DAY_PCT_180,
MIN(
CASE
WHEN DAY_GAP = 90 THEN UP_DAY_PCT
ELSE NULL
END
) AS UP_DAY_PCT_90,
MIN(
CASE
WHEN DAY_GAP = 30 THEN UP_DAY_PCT
ELSE NULL
END
) AS UP_DAY_PCT_30,
MIN(
CASE
WHEN DAY_GAP = 5 THEN UP_DAY_PCT
ELSE NULL
END
) AS UP_DAY_PCT_5,
MIN(END_CLOSE_PCT_C) AS END_CLOSE_PCT_C
FROM
TB_BT_MID_DAILY_PERIOD_TREND
WHERE
1=1
AND END_DATE = '''+strEndDate+'''
GROUP BY
END_DATE,
TS_CODE,
NAME,
INDUSTRY,
EXCHANGE
'''
listQuote = dbConn.ReturnSQLData(strSQLQuote)
# 针对不同的股票，看整体趋势，再看买还是卖
for listStockQuote in listQuote:
# 买入条件
# 1. 比两年前是上升的
# 2. 比一年前是上升的
# 3. 比一个月前上升更下降的区间在 ± 0.38% 内
# 4. 比昨天上涨 0.1% 以上 0.5% 以下
if listStockQuote[18] is not None and listStockQuote[6]>0 and listStockQuote[7]>0 and listStockQuote[10]<0.38 and listStockQuote[10]>-0.38 and listStockQuote[18]>0.1 and listStockQuote[18]<0.5:
# 判断有没有持仓，有则不买
strSQLSell = "SELECT * FROM tb_bt_daily_tmp_my_stock_pool WHERE stock_code='" + listStockQuote[1]+"' AND bt_ind = '"+str(btDaily.intInd)+"'"
listSell = dbConn.ReturnSQLData(strSQLSell)
if len(listSell) == 0 :
btDaily.BuyStock(strNowDate, listStockQuote[1], 1000,'追涨')
# 如果当前下跌 5% 则全卖出
if listStockQuote[18] is not None and listStockQuote[18]<-0.5 :
strSQLSell = "SELECT * FROM tb_bt_daily_tmp_my_stock_pool WHERE stock_code='" + listStockQuote[1]+"' AND bt_ind = '"+str(btDaily.intInd)+"'"
listSell = dbConn.ReturnSQLData(strSQLSell)
if len(listSell) > 0 and listSell[0][0] is not None:
btDaily.SellStock(strNowDate, listStockQuote[1], listSell[0][2], '全部卖了！！！')
# 打印快照
btDaily.AssetSnapshot(strMaxDate)
结果
不得不说，结果实在让人忧伤，我决定就不展示了
反正亏得裤衩都不剩了
1862
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
