ianzeng123
这不是普通的DCA策略，而是一个会思考的交易机器人
看了上千个Pine Script代码，这个”Master Trading Bot”确实有两把刷子。作者把DCA（定投）玩出了新高度：不是无脑定投，而是基于技术指标的智能加仓系统。初始仓位5%，每次DCA增加2.5%，最高不超过100%。这种渐进式资金管理比传统等额定投更科学。
关键在于DCA触发条件：价格必须跌破平均成本价，且跌幅要达到2%+步数×4%的动态阈值。第一次DCA需要跌2%，第二次需要跌6%，第三次需要跌10%。这种设计避免了在小幅波动中频繁加仓，只在真正的回调中增持。
多重技术指标组合，但逻辑清晰不冗余
策略使用3/7/18周期EMA构建趋势框架，配合20周期布林带判断价格位置，52/200/3的MACD参数设置偏向中长期信号，14周期RSI用于超买超卖判断。这个组合覆盖了趋势、动量、波动性三个维度，比单一指标策略更可靠。
买入条件严格：快EMA>慢EMA+MACD金叉+价格在布林中轨上方+RSI<65。这四个条件同时满足才开仓，过滤掉了大部分假信号。卖出条件同样严谨：必须有最低2%利润+趋势转弱+MACD死叉。这种”有利润才卖”的设计避免了无意义的止损。
100%止损设置看似激进，实则合理
代码中100%的止损看起来很夸张，但注释说得很清楚：”价格必须跌到0才触发”。这实际上是关闭了传统止损，完全依靠技术指标和利润目标来管理风险。对于DCA策略来说，这种设计是合理的——既然要在下跌中加仓，传统止损就失去了意义。
真正的风险控制在于：2%价格下跌信号+动态DCA阈值+强制盈利退出。策略会追踪500周期内的最高价，一旦当前价格较最高点下跌超过2%就触发卖出信号。这比固定止损更灵活，能够适应不同的市场环境。
资金管理是这个策略的核心竞争力
每次买入金额=当前权益×DCA百分比÷当前价格，这种按权益比例而非固定金额的设计，让策略能够随着账户增长而扩大仓位。初始5%的仓位控制了单次风险，而渐进式加仓确保了在真正的机会面前有足够的火力。
最精妙的是”just_sold”状态管理：刚卖出后不会立即重新买入，除非出现强烈的看涨信号。这避免了在震荡市场中的频繁交易，降低了手续费成本和情绪化操作的风险。
适用场景明确，不是万能策略
这个策略最适合中长期上涨趋势中的回调买入，在熊市或长期横盘中表现会很一般。MACD的52/200参数设置决定了它更适合较大级别的趋势判断，不适合短线交易。
RSI超卖阈值设在25而非30，说明策略偏向在更深度的回调中买入。这种设计在牛市中能获得更好的买点，但在熊市中可能会”接飞刀”。建议在明确的上升趋势中使用，避免在市场顶部或下降趋势中启动。
回测表现需要关注最大回撤和连续亏损
策略的理论逻辑很完善，但实际表现还要看具体的回测数据。需要重点关注：最大回撤是否在可接受范围内，连续亏损次数是否过多，不同市场环境下的表现差异。
DCA策略的天然特点是会在下跌过程中持续加仓，这意味着账户净值会先下降再上升。投资者需要有足够的心理承受能力和资金储备。建议先在小资金上测试，确认策略特性后再逐步加大投资规模。
风险提示：任何量化策略都存在亏损风险，历史回测不代表未来收益，需要严格的风险管理和适当的资金配置。
策略源码
/*backtest
start: 2024-10-20 00:00:00
end: 2025-10-18 08:00:00
period: 1d
basePeriod: 1d
exchanges: [{"eid":"Futures_Binance","currency":"BTC_USDT","balance":500000}]
*/
// This Pine Script™ code is subject to the terms of the MPL 2.0 at https://mozilla.org/MPL/2.0/
// © MTB by Neurodoc
// By Nicolás Astorga
//@version=5
strategy("Master Trading Bot by Neurodoc",
shorttitle="MTB Adaptation",
overlay=true,
initial_capital=10000,
pyramiding=100,
commission_value=0.1,
commission_type=strategy.commission.percent,
default_qty_type = strategy.cash)
// —————— CONFIGURATION (Based on ve.env) ——————
// Purchase and DCA Percentages
var GRP_DCA = "DCA Configuration"
start_percentage = input.float(5.0, "Initial Buy Percentage (%)", group=GRP_DCA)
increment_percentage = input.float(2.5, "Increment per DCA Buy (%)", group=GRP_DCA)
max_percentage = input.float(100.0, "Maximum Buy Percentage (%)", group=GRP_DCA)
min_profit_percent = input.float(2.0, "Minimum Profit for Sell (%)", group=GRP_DCA)
// Stop Loss and Drop Signal
var GRP_RISK = "Risk Management"
stop_loss_percent = input.float(100.0, "Stop Loss (%)", group=GRP_RISK, tooltip="A value of 100 means there’s no real stop loss, as price would have to go to 0.")
drop_percent_signal = input.float(2.0, "Price Drop for Sell Signal (%)", group=GRP_RISK)
// Indicator Parameters
var GRP_INDICATORS = "Indicator Parameters"
ema_fast_period = input.int(3, "Fast EMA", group=GRP_INDICATORS)
ema_mid_period = input.int(7, "Medium EMA", group=GRP_INDICATORS)
ema_slow_period = input.int(18, "Slow EMA", group=GRP_INDICATORS)
bb_length = input.int(20, "Bollinger Bands Length", group=GRP_INDICATORS)
bb_stddev = input.float(2.0, "BB Standard Deviation", group=GRP_INDICATORS)
macd_fast = input.int(52, "MACD Fast", group=GRP_INDICATORS)
macd_slow = input.int(200, "MACD Slow", group=GRP_INDICATORS)
macd_signal = input.int(3, "MACD Signal", group=GRP_INDICATORS)
rsi_length = input.int(14, "RSI Length", group=GRP_INDICATORS)
rsi_oversold_threshold = input.int(25, "RSI Oversold (for divergence)", group=GRP_INDICATORS)
// —————— INDICATOR CALCULATIONS ——————
// EMAs
ema_fast = ta.ema(open, ema_fast_period)
ema_mid = ta.ema(open, ema_mid_period)
ema_slow = ta.ema(open, ema_slow_period)
// Bollinger Bands
[bb_middle, bb_upper, bb_lower] = ta.bb(close, bb_length, bb_stddev)
bb_width = (bb_upper - bb_lower) / bb_middle * 100
is_bb_expanding = bb_width > bb_width[1]
// MACD
[macd_line, signal_line, _] = ta.macd(close, macd_fast, macd_slow, macd_signal)
// RSI
rsi = ta.rsi(close, rsi_length)
// Price drop signal from highest price (similar to `cummax` in Python)
highest_price = ta.highest(high, 500) // Using 500-bar lookback to approximate the high
price_drop_percent = ((highest_price - close) / highest_price) * 100
is_price_drop_signal = price_drop_percent >= drop_percent_signal
// —————— TRADING LOGIC ——————
// Trend Conditions
is_bullish = ema_fast > ema_slow and macd_line > signal_line and close > bb_middle
is_bearish = ema_fast < ema_slow and macd_line < signal_line and close < bb_middle
is_weakening = rsi < rsi[1]
// Variables to manage strategy state
var bool just_sold = false
var int dca_step = 0
// Determine next buy percentage of capital
dca_buy_percentage = start_percentage + (dca_step * increment_percentage)
if dca_buy_percentage > max_percentage
dca_buy_percentage := max_percentage
avg_buy_price = strategy.position_avg_price
// Initial Long Condition
long_signal_initial = strategy.position_size == 0 and is_bullish and macd_line > signal_line and rsi < 65
// DCA Condition
price_drop_from_avg = ((avg_buy_price - close) / avg_buy_price) * 100
dca_required_drop = 2.0 + (dca_step * 4.0) // DCA price drop start and increment logic
long_signal_dca = strategy.position_size > 0 and is_bearish and close < avg_buy_price and price_drop_from_avg >= dca_required_drop
// Manage `just_sold` state
if strategy.position_size > 0
just_sold := false
if strategy.position_size == 0 and strategy.position_size[1] > 0
just_sold := true
// Avoid immediate repurchase after sell unless bullish condition is strong
long_signal = (just_sold and is_bullish) ? long_signal_initial : (not just_sold ? (long_signal_initial or long_signal_dca) : false)
// Sell (Close) Condition
current_profit_percent = ((close - avg_buy_price) / avg_buy_price) * 100
has_min_profit = current_profit_percent >= min_profit_percent
stop_loss_price = avg_buy_price * (1 - stop_loss_percent / 100)
is_stoploss_triggered = close <= stop_loss_price
short_signal = strategy.position_size > 0 and has_min_profit and ((is_bearish and is_weakening) or is_price_drop_signal or is_stoploss_triggered or (macd_line < signal_line))
// —————— ORDER EXECUTION ——————
if (long_signal)
// Calculate how much MONEY (USDT) to invest in this trade
cash_to_invest = (strategy.equity * dca_buy_percentage / 100) / close
strategy.entry("Buy", strategy.long, qty=cash_to_invest)
dca_step := dca_step + 1
if (short_signal)
strategy.close_all(comment="Sell")
dca_step := 0 // Reset DCA counter after selling
// —————— VISUALIZATION ——————
// Background color by trend
bgcolor(is_bullish ? color.new(color.green, 90) : is_bearish ? color.new(color.red, 90) : na)
// Plot EMAs and Bollinger Bands
plot(ema_fast, "Fast EMA", color.blue)
plot(ema_slow, "Slow EMA", color.orange)
p1 = plot(bb_upper, "Upper BB", color=color.gray)
p2 = plot(bb_lower, "Lower BB", color=color.gray)
fill(p1, p2, color=color.new(color.gray, 90))
// Plot average buy price when in position
plot(strategy.position_size > 0 ? avg_buy_price : na, "Average Buy Price", color.yellow, style=plot.style_linebr, linewidth=2)
// Plot Take Profit target
plot(strategy.position_size > 0 ? avg_buy_price * (1 + min_profit_percent / 100) : na, "Sell Price (TP)", color.aqua, style=plot.style_linebr, linewidth=2)
// Plot Stop Loss level
plot(strategy.position_size > 0 ? stop_loss_price : na, "Stop Loss", color.fuchsia, style=plot.style_linebr, linewidth=2)
相关推荐