ChaoZhang
本文将详细介绍一种利用多周期RSI指标判断反转点的量化交易策略。该策略同时分析多条RSI指标,识别市场拐点的形成。
一、策略原理
该策略使用3组不同参数设置的RSI指标,具体逻辑如下:
分别计算2周期、7周期、14周期的RSI值;
当RSI-2小于10、RSI-7小于20、RSI-14小于30时,判断为底部形成;
当RSI-2大于90、RSI-7大于80、RSI-14大于70时,判断为顶部形成。
根据RSI指标一致性,产生买入卖出信号。
可预设指标一致性要求的参数,从而控制信号频繁程度。
这样,多周期RSI指标的集合分析,可以提升对反转点位的判断准确性。
二、策略优势
该策略最大优势是利用多周期RSI指标进行集合分析,这可以提高对关键点位的判断准确性,滤除假信号。
另一优势是可通过调整一致性参数,控制交易频率,适应不同市场环境。
最后,不同周期RSI的组合也提供了更多参数空间进行优化。
三、潜在风险
但该策略也存在以下风险:
首先,RSI指标对价格反转的判断,本身存在滞后问题。
其次,多指标组合带来的信号判断困难,需要设置清晰的过滤规则。
最后,反转交易本身带有一定的失败率,这需要有心理准备。
四、内容总结
本文详细介绍了一种基于多周期RSI指标识别反转点的量化交易策略。它通过判断RSI指标的一致性,提高了对市场转折点的识别能力。但也需要防控滞后问题和信号判断错误。总体来说,它提供了一种参数灵活的RSI策略优化思路。
策略源码
/*backtest
start: 2023-09-06 00:00:00
end: 2023-09-13 00:00:00
period: 45m
basePeriod: 5m
exchanges: [{"eid":"Futures_Binance","currency":"BTC_USDT"}]
*/
//Noro
//2018
//@version=2
strategy(title = "Noro's Triple RSI Top/Bottom v1.1", shorttitle = "3RSI Top/Bottom 1.1", overlay = true, default_qty_type = strategy.percent_of_equity, default_qty_value = 100, pyramiding = 0)
//Settings
needlong = input(true, defval = true, title = "Long")
needshort = input(true, defval = true, title = "Short")
leverage = input(1, defval = 1, minval = 1, maxval = 100, title = "leverage")
indi = input(3, defval = 3, minval = 1, maxval = 3, title = "Indicators")
accuracy = input(3, defval = 3, minval = 1, maxval = 10, title = "accuracy")
fromyear = input(1900, defval = 1900, minval = 1900, maxval = 2100, title = "From Year")
toyear = input(2100, defval = 2100, minval = 1900, maxval = 2100, title = "To Year")
frommonth = input(01, defval = 01, minval = 01, maxval = 12, title = "From Month")
tomonth = input(12, defval = 12, minval = 01, maxval = 12, title = "To Month")
fromday = input(01, defval = 01, minval = 01, maxval = 31, title = "From day")
today = input(31, defval = 31, minval = 01, maxval = 31, title = "To day")
//RSI-2
fastup = rma(max(change(close), 0), 2)
fastdown = rma(-min(change(close), 0), 2)
fastrsi = fastdown == 0 ? 100 : fastup == 0 ? 0 : 100 - (100 / (1 + fastup / fastdown))
//RSI-7
middleup = rma(max(change(close), 0), 7)
middledown = rma(-min(change(close), 0), 7)
middlersi = middledown == 0 ? 100 : middleup == 0 ? 0 : 100 - (100 / (1 + middleup / middledown))
//RSI-14
slowup = rma(max(change(close), 0), 14)
slowdown = rma(-min(change(close), 0), 14)
slowrsi = slowdown == 0 ? 100 : slowup == 0 ? 0 : 100 - (100 / (1 + slowup / slowdown))
//Body
body = abs(close - open)
abody = sma(body, 10)
//Signals
acc = 10 - accuracy
signalup1 = fastrsi < (5 + acc) ? 1 : 0
signalup2 = middlersi < (10 + acc * 2) ? 1 : 0
signalup3 = slowrsi < (15 + acc * 3) ? 1 : 0
signaldn1 = fastrsi > (95 - acc) ? 1 : 0
signaldn2 = middlersi > (90 - acc * 2) ? 1 : 0
signaldn3 = slowrsi > (85 - acc * 3) ? 1 : 0
up = signalup1 + signalup2 + signalup3 >= indi
dn = signaldn1 + signaldn2 + signaldn3 >= indi
exit = ((strategy.position_size > 0 and close > open) or (strategy.position_size < 0 and close < open)) and body > abody / 3
//Trading
lot = strategy.position_size == 0 ? strategy.equity / close * leverage : lot[1]
if up
if strategy.position_size < 0
strategy.close_all()
strategy.entry("Bottom", strategy.long, needlong == false ? 0 : lot)
if dn
if strategy.position_size > 0
strategy.close_all()
strategy.entry("Top", strategy.short, needshort == false ? 0 : lot)
if time > timestamp(toyear, tomonth, today, 23, 59) or exit
strategy.close_all()