紅買綠賣的交易策略 - 平均K線

EasyTrader ArtNo 017

讀者應記得在"變身的K棒"文章內提到了利用PlotPaintBar的功能可以讓K棒的顏色依據個人的獲利操作心法轉換成紅買綠賣的指標,近日在網路上看到一篇不同的平均K線計算方式,因此將幾個不同的計算方式作為交易策略的核心 ,並比較其差異(下圖為平均K線與傳統K線圖比較)



不同的平均K線的計算公式如下：(為了方便,就直接將筆者程式碼貼上)

1.近日網路上看到的,只有開盤及收盤價
if CurrentBar > 1 then
    NewOpen = ( ( Open[1] + High[1] + Low[1] + Close[1] ) / 4 + NewOpen[1] ) / 2
else NewOpen = Open ;

NewClose = (( Open + High + Low + Close ) / 4 + NewOpen + MaxList( High, NewOpen ) + Minlist( Low, NewOpen )) / 4 ;

2.標準的平均k線

MyClose = (Close + Open + High + Low )/4 ;
if CurrentBar > 1 then MyOpen = (MyOpen[1]+MyClose[1])/2 else MyOpen = Open;
MyHigh = High ;
MyLow = Low ;
NewOpen = MyOpen ;
NewClose = MyClose ;

3.書上看來的
RefOpen = Open ;
if CurrentBar > 1 then MyOpen =(RefOpen[1] + Close[1])/2 ;
MyClose = (Close + Open + High + Low )/4 ;
MyHigh = MaxList(High,MyOpen,MyClose) ;
MyLow = MinList(Low,MyOpen,MyClose) ;
NewOpen = MyOpen ;
NewClose = MyClose ;

4.自己編的(只有開盤及收盤價)NewOpen = (MaxList(Open,Close)+High)/2 ;
NewClose = (MinList(Open,Close)+Low)/2 ;

針對同一個進出場邏輯,將此四種不同計算方式作測試
基本設定: 5分K , 留倉策略 , 來回成本 1200 回測時間 2001~2013/8/30

結果如下

公式1




公式2(公式 3 結果雷同因為沒用到 High/Low)





公式4





從以上的結果只有公式1較佳,而從三個績效曲線看,不知道是否合成起來會比較好,因此修改了 NewClose的計算

公式 5 NewClose = 公式 1 , 2, 4的平均





未蒙其利先受其害,獲利沒增加,反而MDD增加, 是不是進場規則不合適 ? 動念之間改了個符號 , 哇! 怎會變化這麼大???





績效變好,MDD也下降,真是"柳暗花明又一村" ~~

範例程式碼

input: Type(1) ; Vars:NewOpen(0),NewClose(0),MyClose(0),MyOpen(0),MyHigh(0),MyLow(0),RefOpen(0) ; Vars:OpenA(0),CloseA(0),OpenB(0),CloseB(0),OpenC(0),CloseC(0),MaxClose(0),MinClose(0); {公式1} if Type = 1 then Begin if CurrentBar > 1 then     NewOpen = ( ( Open[1] + High[1] + Low[1] + Close[1] ) / 4 + NewOpen[1] ) / 2 else  NewOpen = Open ; NewClose = (( Open + High + Low + Close ) / 4 + NewOpen + MaxList( High, NewOpen ) + Minlist( Low, NewOpen )) / 4 ; end; {公式2} if Type = 2 then Begin    MyClose = (Close + Open + High + Low )/4 ;    if CurrentBar > 1 then MyOpen = (MyOpen[1]+MyClose[1])/2 else MyOpen = Open;    MyHigh = High ;    MyLow = Low ;    NewOpen = MyOpen ;    NewClose = MyClose ; end; {公式3} if Type = 3 then Begin    RefOpen = Open ;    if CurrentBar > 1 then MyOpen =(RefOpen[1] + Close[1])/2 ;    MyClose = (Close + Open + High + Low )/4 ;    MyHigh = MaxList(High,MyOpen,MyClose) ;    MyLow = MinList(Low,MyOpen,MyClose) ;    NewOpen = MyOpen ;    NewClose = MyClose ; end; {公式4} if Type = 4 then Begin    NewOpen = (MaxList(Open,Close)+High)/2 ;    NewClose = (MinList(Open,Close)+Low)/2 ; end; {公式5} if Type = 5 then Begin     if CurrentBar > 1 then        OpenA = ( ( Open[1] + High[1] + Low[1] + Close[1] ) / 4 + OpenA[1] ) / 2     else OpenA = Open ;     CloseA = (( Open + High + Low + Close ) / 4 + OpenA + MaxList( High, OpenA )     + Minlist( Low, OpenA )) / 4 ;     CloseB = (Close + Open + High + Low )/4 ;     if CurrentBar > 1 then OpenB = (OpenB[1]+CloseB[1])/2 else OpenB = Open;     MyHigh = High ;     MyLow = Low ;     NewOpen = OpenB ;     NewClose = CloseB ;     OpenC = (MaxList(Open,Close)+High)/2 ;     CloseC = (MinList(Open,Close)+Low)/2 ;     NewClose = (CloseA+CloseB+CloseC)/3 ;     MinClose = MinList(CloseA,CloseB,CloseC);     MaxClose = MaxList(CloseA,CloseB,CloseC); end; {交易策略} inputs: TypPriceLen(5),NewCloseLen(8),StopLossRatio(25),BreakEvenRatio(15); vars: MP(0),AvgCost(0),EntryCount(0),EntryHigh(0),EntryLow(0),StopLossLong(0), StopLossShort(0),LongTrigger(0),ShortTrigger(0),AvgOpen(0),AVGTyp(0),AVGClose(0) ,LineCorss(-1),AvgMaxClose(0),AvgMinClose(0) ; { 變數設定 } MP = MarketPosition ; AvgCost = AvgEntryPrice ; EntryCount = BarsSinceEntry ; { 進場後高低點 } if MP <> 0 then begin{ in a postion }     if EntryCount = 0 then begin{ entry bar }        EntryHigh = High ;        EntryLow = Low ;     end else begin        EntryHigh = MaxList( EntryHigh,High ) ;        EntryLow = MinList( EntryLow,Low ) ;     end ; end ; {停損設定} StopLossLong = EntryPrice {AvgCost} * ( 1 - StopLossRatio * 0.01 ) ; StopLossShort = EntryPrice {AvgCost} * ( 1 + StopLossRatio * 0.01 ) ; LongTrigger = EntryPrice {AvgCost} * ( 1 + BreakEvenRatio * 0.01 ) ; ShortTrigger = EntryPrice {AvgCost} * ( 1 - BreakEvenRatio * 0.01 ) ; { 計算幾個測試用變數 } AVGTyp = XAverage( TypicalPrice,TypPriceLen) ; AVGOpen = XAverage( NewOpen,TypPriceLen) ; AVGClose = XAverage( NewClose, NewCloseLen ) ; AvgMaxClose = XAverage( MaxClose, NewCloseLen ) ; AvgMinClose = XAverage( MinClose, NewCloseLen ) ; {進場規則設計} if AVGTyp > AVGClose and Close > Open then LineCorss = 1     else if AVGTyp < AVGClose and Close < Open then LineCorss = 0 ; if CurrentBar > 1 then begin { 多單進場 }    if OpenD(0) > CloseD(1) and Close > OpenD(0) and LineCorss = 1 and LineCorss[1] = 0        then Buy ( "MQS01_L" ) next bar market ; { 空單進場 }    if OpenD(0) < CloseD(1) and Close < OpenD(0) and LineCorss = 0 and LineCorss[1] = 1        then Sell( "MQS01_S" ) next bar at market ; {停損機制 }    if MP = 1 then ExitLong ( "MQS01_LX" ) next bar at StopLossLong stop ;    if MP = -1 then ExitShort( "MQS01_SX" ) next bar at StopLossShort stop ; { Breakeven exits }    if MP = 1 and EntryHigh >= LongTrigger then        ExitLong ( "MQS01_BEL" ) next bar at AvgCost stop ;    if MP = -1 and EntryLow <= ShortTrigger then       ExitShort ( "MQS01_BES" ) next bar at AvgCost stop ; end ;

後記:

交易策略開發過程有時碰牆撞壁時,將概念轉個彎,就像連續劇內常聽到的台詞"代誌不是你所想的那樣",也會有意料不到的結果

(程式碼變更位置說明)

if AVGTyp < AVGClose and Close > Open then LineCorss = 1

   else if  AVGTyp > AVGClose and Close < Open then LineCorss = 0 ;

就是這個符號變動,好玩吧!!

MagicQS001