ChaoZhang
概述
双均线交易策略通过计算不同周期的指数移动平均线,形成快线和慢线,并观察它们的金叉和死叉形态来产生交易信号。当快线从下方上穿慢线时,做多;当快线从上方下穿慢线时,做空。该策略捕捉了均线的趋势反转点,是一种较为常见的趋势跟踪策略。
策略原理
双均线交易策略的核心指标是计算快线和慢线。快线指短周期的指数移动平均线,默认参数为12日线;慢线指长周期的指数移动平均线,默认参数为26日线。指数移动平均线的计算公式为:
EMA(t) = (C(t) - EMA(t-1)) * SF + EMA(t-1)
其中,C(t)为当日收盘价,SF为smoothing factor 平滑因子。指数移动平均线与普通算术移动平均线的区别在于,指数移动平均线赋予了最近数据更大的权重,可以更快速地响应价格变化。
双均线策略的交易规则为:
- 当快线从下方上穿慢线时,即golden cross金叉形成,做多入场;
- 当快线从上方下穿慢线时,即dead cross死叉形成,做空入场;
- 快线和慢线发生背离时,平仓离场。
通过capture监控均线的交叉形态,及时反应市场供需关系和趋势的变化,实现盈利。
优势分析
双均线交易策略作为一种较为成熟的技术指标策略,具有以下优势:
- 思路清晰,易于理解和实现;
- 对市场供需关系的判断准确,胜率较高;
- 有效过滤市场噪音,捕捉主要趋势;
- 可在不同市场和时间框架下应用;
- 可与其他技术指标组合,丰富策略;
- 资金利用率高,符合大资金需求。
风险分析
双均线交易策略也存在一定的缺陷和风险:
- 无法应对剧烈行情,如快速熊市;
- 容易产生假信号和频繁小幅震荡带来密集交易;
- 需优化参数以适应不同品种和时间周期;
- 无法判断趋势反转的合理位置。
针对以上风险,可通过调整均线周期参数、引入附加过滤器等方法进行优化,确保策略更稳健。
优化方向
双均线交易策略可从以下几个方面进行优化:
- 引入MACD指标判断强弱趋势,避免弱势震荡行情的错误交易;
- 增加交易 volumes 作为确认指标,避免趋势反转的假突破;
- 结合其它如布林线、K线等技术指标,设定更精确的入场退场条件;
- 利用机器学习方法如LSTM自动优化均线参数,实现更好的市场适应性。
总结
双均线交易策略通过捕捉均线的金叉与死叉交易机会,判断价格趋势反转点,实现稳定盈利。该策略优势在于简洁清晰、资金效率高,是量化入门的首选策略。但也存在一定缺陷如产生假信号等,须引入更多指标进行优化,使其能更好地适应特定品种和交易环境。整体而言,双均线交易策略是一个非常实用的技术指标策略。
策略源码
/*backtest
start: 2022-11-24 00:00:00
end: 2023-11-30 00:00:00
period: 1d
basePeriod: 1h
exchanges: [{"eid":"Futures_Binance","currency":"BTC_USDT"}]
*/
// This source code is subject to the terms of the Mozilla Public License 2.0 at https://mozilla.org/MPL/2.0/
// © antondmt
//@version=5
strategy("Returns & Drawdowns Table", "R & DD", true, calc_on_every_tick = false, default_qty_type = strategy.percent_of_equity, default_qty_value = 100, process_orders_on_close = true)
i_eq_to_dd = input.string("Compound Equity", "Mode", ["Simple Equity", "Compound Equity", "Drawdown"], group = "R & DD Table")
i_precision = input.int(2, "Return Precision", group = "R & DD Table")
i_headers_col = input.color(#D4D4D4, "Headers Color", group = "R & DD Table")
i_headers_text_col = input.color(color.black, "Headers Text Color", group = "R & DD Table")
i_pos_col = input.color(color.green, "Positive Color", group = "R & DD Table")
i_neg_col = input.color(color.red, "Negative Color", group = "R & DD Table")
i_zero_col = input.color(#DDDDDD, "Zero Color", group = "R & DD Table")
i_cell_text_col = input.color(color.white, "Cell Text Color", group = "R & DD Table")
// TIME {
var month_times = array.new_int(0) // Array of all month times
new_month = month(time) != month(time[1])
if(new_month or barstate.isfirst)
array.push(month_times, time)
var year_times = array.new_int(0)
new_year = year(time) != year(time[1])
if (new_year or barstate.isfirst)
array.push(year_times, time)
//}
// SIMPLE EQUITY CALCULATIONS {
// Simple equity is strictly calculated from start to end of each month/year equity. There is no compound
var monthly_simp_pnls = array.new_float(0) // Array of all monthly profits and losses
var yearly_simp_pnls = array.new_float(0)
if(i_eq_to_dd == "Simple Equity")
var initial_monthly_equity = strategy.equity // Starting equity for each month
cur_month_pnl = nz((strategy.equity - initial_monthly_equity) / initial_monthly_equity) // Current month's equity change
if(new_month or barstate.isfirst)
initial_monthly_equity := strategy.equity
array.push(monthly_simp_pnls, cur_month_pnl)
else
array.set(monthly_simp_pnls, array.size(monthly_simp_pnls) - 1, cur_month_pnl)
var initial_yearly_equity = strategy.equity
cur_year_pnl = nz((strategy.equity - initial_yearly_equity) / initial_yearly_equity)
if (new_year or barstate.isfirst)
initial_yearly_equity := strategy.equity
array.push(yearly_simp_pnls, cur_year_pnl)
else
array.set(yearly_simp_pnls, array.size(yearly_simp_pnls) - 1, cur_year_pnl)
// }
// COMPOUND EQUITY CALCULATIONS {
// Compound equity is strictly calculated based on equity state from the beginning of time until the end of each month/year equity. It shows the exact equity movement through time
var monthly_comp_pnls = array.new_float(0) // Array of all monthly profits and losses
var yearly_comp_pnls = array.new_float(0)
if(i_eq_to_dd == "Compound Equity")
var initial_equity = strategy.equity
cur_month_pnl = nz((strategy.equity - initial_equity) / initial_equity) // Current month's equity change
if(new_month or barstate.isfirst)
array.push(monthly_comp_pnls, cur_month_pnl)
else
array.set(monthly_comp_pnls, array.size(monthly_comp_pnls) - 1, cur_month_pnl)
cur_year_pnl = nz((strategy.equity - initial_equity) / initial_equity)
if (new_year or barstate.isfirst)
array.push(yearly_comp_pnls, cur_year_pnl)
else
array.set(yearly_comp_pnls, array.size(yearly_comp_pnls) - 1, cur_year_pnl)
// }
// DRAWDOWN CALCULATIONS {
// Drawdowns are calculated from highest equity to lowest trough for the month/year
var monthly_dds = array.new_float(0) // Array of all monthly drawdowns
var yearly_dds = array.new_float(0)
if (i_eq_to_dd == "Drawdown")
total_equity = strategy.equity - strategy.openprofit
var cur_month_dd = 0.0
var m_ATH = total_equity // Monthly All-Time-High (ATH). It is reset each month
m_ATH := math.max(total_equity, nz(m_ATH[1]))
m_drawdown = -math.abs(total_equity / m_ATH * 100 - 100) / 100 // Drawdown at current bar
if(m_drawdown < cur_month_dd)
cur_month_dd := m_drawdown
if(new_month or barstate.isfirst)
cur_month_dd := 0.0
m_ATH := strategy.equity - strategy.openprofit
array.push(monthly_dds, 0)
else
array.set(monthly_dds, array.size(monthly_dds) - 1, cur_month_dd)
var cur_year_dd = 0.0
var y_ATH = total_equity
y_ATH := math.max(total_equity, nz(y_ATH[1]))
y_drawdown = -math.abs(total_equity / y_ATH * 100 - 100) / 100
if(y_drawdown < cur_year_dd)
cur_year_dd := y_drawdown
if (new_year or barstate.isfirst)
cur_year_dd := 0.0
y_ATH := strategy.equity - strategy.openprofit
array.push(yearly_dds, 0)
else
array.set(yearly_dds, array.size(yearly_dds) - 1, cur_year_dd)
// }
// TABLE LOGIC {
var main_table = table(na)
table.clear(main_table, 0, 0, 13, new_year ? array.size(year_times) - 1 : array.size(year_times))
main_table := table.new(position.bottom_right, columns = 14, rows = array.size(year_times) + 1, border_width = 1)
t_set_headers() => // Sets time headers of the table
// Set month headers
table.cell(main_table, 0, 0, "", text_color = i_headers_text_col, bgcolor = i_headers_col)
table.cell(main_table, 1, 0, "Jan", text_color = i_headers_text_col, bgcolor = i_headers_col)
table.cell(main_table, 2, 0, "Feb", text_color = i_headers_text_col, bgcolor = i_headers_col)
table.cell(main_table, 3, 0, "Mar", text_color = i_headers_text_col, bgcolor = i_headers_col)
table.cell(main_table, 4, 0, "Apr", text_color = i_headers_text_col, bgcolor = i_headers_col)
table.cell(main_table, 5, 0, "May", text_color = i_headers_text_col, bgcolor = i_headers_col)
table.cell(main_table, 6, 0, "Jun", text_color = i_headers_text_col, bgcolor = i_headers_col)
table.cell(main_table, 7, 0, "Jul", text_color = i_headers_text_col, bgcolor = i_headers_col)
table.cell(main_table, 8, 0, "Aug", text_color = i_headers_text_col, bgcolor = i_headers_col)
table.cell(main_table, 9, 0, "Sep", text_color = i_headers_text_col, bgcolor = i_headers_col)
table.cell(main_table, 10, 0, "Oct", text_color = i_headers_text_col, bgcolor = i_headers_col)
table.cell(main_table, 11, 0, "Nov", text_color = i_headers_text_col, bgcolor = i_headers_col)
table.cell(main_table, 12, 0, "Dec", text_color = i_headers_text_col, bgcolor = i_headers_col)
table.cell(main_table, 13, 0, str.tostring(i_eq_to_dd), text_color = i_headers_text_col, bgcolor = i_headers_col)
// Set year headers
for i = 0 to array.size(year_times) - 1
table.cell(main_table, 0, i + 1, str.tostring(year(array.get(year_times, i))), text_color = i_headers_text_col, bgcolor = i_headers_col)
t_set_months() => // Sets inner monthly data of the table
display_array = switch i_eq_to_dd
"Simple Equity" => monthly_simp_pnls
"Compound Equity" => monthly_comp_pnls
=> monthly_dds
for i = 0 to array.size(month_times) - 1
m_row = year(array.get(month_times, i)) - year(array.get(year_times, 0)) + 1
m_col = month(array.get(month_times, i))
m_color = array.get(display_array, i) == 0 ? color.new(i_zero_col, transp = 30) : array.get(display_array, i) > 0 ? color.new(i_pos_col, transp = 30) : color.new(i_neg_col, transp = 30)
table.cell(main_table, m_col, m_row, str.tostring(math.round(array.get(display_array, i) * 100, i_precision)), bgcolor = m_color, text_color = i_cell_text_col)
t_set_years() => // Sets inner yearly data of the table
display_array = switch i_eq_to_dd
"Simple Equity" => yearly_simp_pnls
"Compound Equity" => yearly_comp_pnls
=> yearly_dds
for i = 0 to array.size(year_times) - 1
y_color = array.get(display_array, i) == 0 ? color.new(i_zero_col, transp = 30) : array.get(display_array, i) > 0 ? color.new(i_pos_col, transp = 20) : color.new(i_neg_col, transp = 20)
table.cell(main_table, 13, i + 1, str.tostring(math.round(array.get(display_array, i) * 100, i_precision)), bgcolor = y_color, text_color = i_cell_text_col)
t_set_headers()
t_set_months()
t_set_years()
// }
// PLACE YOUR STRATEGY CODE HERE {
// This is a sample code of a working strategy to show the table in action
fastLength = input(12)
slowlength = input(26)
MACDLength = input(9)
MACD = ta.ema(close, fastLength) - ta.ema(close, slowlength)
aMACD = ta.ema(MACD, MACDLength)
delta = MACD - aMACD
if (ta.crossover(delta, 0))
strategy.entry("MacdLE", strategy.long, comment = "MacdLE")
if (ta.crossunder(delta, 0))
strategy.entry("MacdSE", strategy.short, comment = "MacdSE")
// }