Discussion re Pure Renko strategy

GraHal · Tue, 25 Feb 2020 10:10:02 +0000

edit : new topic from https://www.prorealcode.com/topic/why-is-backtesting-so-unreliable/ According to IG Status Page … all trading platforms are now sorted!! See attached. This Renko code is the only code I am still getting Rejects / Stoppages on (up to now anyway!?). Only major difference I can see between this System and my others is the use of While, Wend? Maybe While. Wend has recently become a non-approved term anymore?? @Nonetheless would it be easy for you to re-code without While, Wend? I know the code is a DocTrading code so maybe it not easy for you either? If one of our coding wizards might be able to help it would be much appreciated? Below is the code of the System that is showing +£1.8K profit since October (see attached). //------------------------------------------------------------------------- // Main code : Nneless Renko DJI 1M v1 //------------------------------------------------------------------------- //https://www.prorealcode.com/topic/why-is-backtesting-so-unreliable/#post-110889 // Definition of code parameters DEFPARAM CumulateOrders = False // Cumulating positions deactivated // The system will cancel all pending orders and close all positions at 0:00. No new ones will be allowed until after the "FLATBEFORE" time. //DEFPARAM FLATBEFORE = 143000 // Cancel all pending orders and close all positions at the "FLATAFTER" time //DEFPARAM FLATAFTER = 210000 boxSize = 110 //50 once renkoMax = ROUND(close / boxSize) * boxSize once renkoMin = renkoMax - boxSize IF high > renkoMax + boxSize THEN WHILE high > renkoMax + boxSize renkoMax = renkoMax + boxSize renkoMin = renkoMin + boxSize WEND ELSIF low < renkoMin - boxSize THEN WHILE low < renkoMin - boxSize renkoMax = renkoMax - boxSize renkoMin = renkoMin - boxSize WEND ENDIF buy 1 CONTRACT at renkoMax + boxSize stop sellshort 1 CONTRACT at renkoMin - boxSize stop // Stops and targets SET STOP PLOSS 75 SET TARGET PPROFIT 150

This topic has 345 replies, 24 voices, and was last updated 3 years ago by bertrandpinoy.

Tagged: renko

Viewing 15 posts - 61 through 75 (of 346 total)

← 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 … 20 21 22 23 24 →

03/12/2020 at 7:42 PM #121991

Francesco

No one is working trying to convert the strategy on higher timeframes? Maybe we can share ideas i would like to support it, it seems very interesting but i think will never put on live a strategy with 1 sec

Francesco

Participant

Topics: 32

Replies: 439

Been thanked: 200 times

03/12/2020 at 8:36 PM #122005

Paul

maybe what can be tested is trading the equity curve

if performance falls, stop live trading, but keep counting fake trade results.

If the fake results pick up, start trading live again.

@Francesco second timeframe appears to be excellent to iron out the shortcomings and fix those. Later on I go back to a slower timeframe.

Today anyway was a great day, but I’am not so sure if the results hold, when trading a much smaller range.

1 user thanked author for this post.

Francesco

Paul

Participant

Topics: 40

Replies: 812

Been thanked: 660 times

03/12/2020 at 8:58 PM #122006

GraHal

Paul wrote:

f performance falls, stop live trading, but keep counting fake trade results. If the fake results pick up, start trading live again.

Good idea as I have a few times today thought … I’ll put this live on Lot size 0.2 as it is doing so well then it goes through a losing phase! 🙂 But still in profit overall!

If it wasn’t for the “broker does not know the status of your last order” System Stop then I would have 5 or 6 versions running at the same time and pick the best!

GraHal

Participant

Topics: 204

Replies: 8397

Been thanked: 2038 times

03/14/2020 at 11:52 PM #122154

Paul

Found a code from Nicolas and adopted it slightly to fit this strategy. It slows it down on 200k bars though.

i.e. Can be used to trade only in the trend

//moving average
once usema=1
if usema then
once priceMAperiod=14
count=0
sum=0
lastbox=0
for i = 0 to barindex do
if renkomax[i]<>lastbox then
lastbox=renkomax[i]
count=count+1
median=((renkomax[i]+renkomin[i])/2)
sum=sum+median
if count=priceMAperiod+1 then
sum=sum-last
last=median
break
//
endif
endif
next
once avg=undefined
if last<>last[1] then
avg = sum/priceMAperiod
endif
graphonprice avg as "average renko"
endif

//moving average

once usema=1

if usema then

once priceMAperiod=14

count=0

sum=0

lastbox=0

for i = 0 to barindex do

if renkomax[i]<>lastbox then

lastbox=renkomax[i]

count=count+1

median=((renkomax[i]+renkomin[i])/2)

sum=sum+median

if count=priceMAperiod+1 then

sum=sum-last

last=median

break

endif

once avg=undefined

if last<>last[1] then

avg = sum/priceMAperiod

endif

graphonprice avg as "average renko"

endif

there was always something off with the “dynamic box”. Removed the +/- 1.

boxsize=25 // 5 to 35
boxdiff=0  // -5 to 5

if dynamicbox then
if close>close[1] and close[1]>close[2] then
boxsize=boxsize+boxdiff
elsif close<close[1] and close[1]<close[2] then
boxsize=boxsize-boxdiff
endif
endif

boxsize=25 // 5 to 35

boxdiff=0 // -5 to 5

if dynamicbox then

if close>close[1] and close[1]>close[2] then

boxsize=boxsize+boxdiff

elsif close<close[1] and close[1]<close[2] then

boxsize=boxsize-boxdiff

endif

2 users thanked author for this post.

Bard,

GraHal

Paul

Participant

Topics: 40

Replies: 812

Been thanked: 660 times

03/15/2020 at 2:00 PM #122169

Paul

it’s based on code what’s already used to determine performance.

Takes the last completed trade result from a long (green) or short (red) and plot’s it as 1 for a win and -1 for a loss.

Not sure yet if it could be usefull.

if (longperf>longperf[2]) then
previouslong=1
elsif (longperf<longperf[2]) then
previouslong=-1
else
previouslong=0
endif
if (shortperf>shortperf[2]) then
previousshort=1
elsif (shortperf<shortperf[2]) then
previousshort=-1
else
previousshort=0
endif

graph previouslong coloured(0,200,0) as  "long result  (1=win;-1=loss)"
graph previousshort coloured(200,0,0) as "short result (1=win;-1=loss)"

if (longperf>longperf[2]) then

previouslong=1

elsif (longperf<longperf[2]) then

previouslong=-1

else

previouslong=0

endif

if (shortperf>shortperf[2]) then

previousshort=1

elsif (shortperf<shortperf[2]) then

previousshort=-1

else

previousshort=0

endif

graph previouslong coloured(0,200,0) as "long result (1=win;-1=loss)"

graph previousshort coloured(200,0,0) as "short result (1=win;-1=loss)"

Paul

Participant

Topics: 40

Replies: 812

Been thanked: 660 times

03/15/2020 at 3:23 PM #122181

GraHal

Paul wrote:

Not sure yet if it could be usefull.

If anybody can … you can Paul!

I’m sold on these 1 second Systems …. loved it last week, so interesting due to loads trades!

And with the DJi up and down by 1oo points and more in 1 min the 1 Sec TF could be the way to go!?

Btw I found Hour > 3 and Hour < 23 giving consistent good results re times to trade.

2 users thanked author for this post.

nonetheless,

Paul

GraHal

Participant

Topics: 204

Replies: 8397

Been thanked: 2038 times

03/16/2020 at 4:06 PM #122258

Paul

GraHal wrote:

If it wasn’t for the “broker does not know the status of your last order” System Stop then I would have 5 or 6 versions running at the same time and pick the best!

When trying different positionsizes. like 1 contract, 1,01 etc I’am testing currently. Doesn’t seem to work though?

Interesting is the south african 40 to test too, in the timezone with the lowest spread.

Still backtest difference from the forward demo test, even with the fixes I tried.

1 user thanked author for this post.

GraHal

Paul

Participant

Topics: 40

Replies: 812

Been thanked: 660 times

03/16/2020 at 4:36 PM #122267

Paul

same strategy in forward demo test

one 0.27 positionsize, other 0.28. Rest is the same.

Paul

Participant

Topics: 40

Replies: 812

Been thanked: 660 times

03/16/2020 at 5:39 PM #122274

fifi743

Hi .
I don’t understand this
It is not a comparison.

myindex + nbbarlimit

if rule1 and myindex + nbbarlimit and mylimitshort > 0  and not shortonmarket then
sellshort positionsize contract at market
endif

if rule1 and myindex + nbbarlimit and mylimitshort > 0 and not shortonmarket then

sellshort positionsize contract at market

endif

fifi743

Participant

Topics: 9

Replies: 847

Been thanked: 401 times

03/16/2020 at 6:35 PM #122279

Paul

Good evening Fifi,

The code comes from the prorealtime manual on page 11.

You can set the validity length of limit and stop orders.

The following example shows how it is possible to create a limit order with a validity set to a specific number of bars by using variables. The code places a buy limit order at the close price of the bar on which a moving- average crossover occurred. This limit is valid for 10 bars after the bar of the crossing. If it is not executed during these 10 bars, it is cancelled.

// Definition of the validity length of the order
ONCE NbBarLimit = 10
MM20 = Average[20](close) MM50 = Average[50](close)
// If MM20 crosses over MM50, we define 2 variables "MyLimitBuy" and "MyIndex" containing the close price at that time and the index of the bar of the cross.
IF MM20 CROSSES OVER MM50 THEN MyLimitBuy = close
MyIndex = Barindex
ENDIF
IF BarIndex >= MyIndex + NbBarLimit THEN MyLimitBuy = 0
ENDIF
// Place an order at the price MyLimitBuy valid as long as this variable is greater than 0 and we are not in a long position.
// Remember: MyLimitBuy is greater than 0 for the 10 bars after the bar of the crossing.
IF MyLimitBuy > 0 AND NOT LongOnMarket THEN
   BUY 1 SHARES AT MyLimitBuy LIMIT
ENDIF

// Definition of the validity length of the order

ONCE NbBarLimit = 10

MM20 = Average[20](close) MM50 = Average[50](close)

// If MM20 crosses over MM50, we define 2 variables "MyLimitBuy" and "MyIndex" containing the close price at that time and the index of the bar of the cross.

IF MM20 CROSSES OVER MM50 THEN MyLimitBuy = close

MyIndex = Barindex

ENDIF

IF BarIndex >= MyIndex + NbBarLimit THEN MyLimitBuy = 0

ENDIF

// Place an order at the price MyLimitBuy valid as long as this variable is greater than 0 and we are not in a long position.

// Remember: MyLimitBuy is greater than 0 for the 10 bars after the bar of the crossing.

IF MyLimitBuy > 0 AND NOT LongOnMarket THEN

BUY 1 SHARES AT MyLimitBuy LIMIT

ENDIF

In case the order was not executed, it is possible to replace the expired buy limit order with a buy at market price order. This could be done by adding the following code to the previous one:

IF MyIndex + NbBarLimit AND MyLimitBuy > 0 AND NOT LongOnMarket THEN
   BUY 1 SHARES AT MARKET
ENDIF

IF MyIndex + NbBarLimit AND MyLimitBuy > 0 AND NOT LongOnMarket THEN

BUY 1 SHARES AT MARKET

ENDIF

1 user thanked author for this post.

GraHal

Paul

Participant

Topics: 40

Replies: 812

Been thanked: 660 times

03/16/2020 at 6:45 PM #122281

fifi743

I do not know where you are in the updates.
here’s mine

3 users thanked author for this post.

Francesco,

Paul,

GraHal

fifi743

Participant

Topics: 9

Replies: 847

Been thanked: 401 times

03/16/2020 at 6:47 PM #122284

GraHal

Link to Paul’s post above added as Log 200 here …

Snippet Link Library

1 user thanked author for this post.

fifi743

GraHal

Participant

Topics: 204

Replies: 8397

Been thanked: 2038 times

03/16/2020 at 8:10 PM #122290

fifi743

Right above the trailling stop, I put this code.

if longonmarket and barindex<=myindex + nbbarlimit and strategyprofit>strategyprofit+(close-positionprice)*round(COUNTOFLONGSHARES/positionsize) then
trailingstoptype=0
else
trailingstoptype=1
endif
if shortonmarket and barindex>=myindex + nbbarlimit and strategyprofit>strategyprofit+(positionprice-close)*round(COUNTOFSHORTSHARES/positionsize) then
trailingstoptype=0
else
trailingstoptype=1
endif

if longonmarket and barindex<=myindex + nbbarlimit and strategyprofit>strategyprofit+(close-positionprice)*round(COUNTOFLONGSHARES/positionsize) then

trailingstoptype=0

else

trailingstoptype=1

endif

if shortonmarket and barindex>=myindex + nbbarlimit and strategyprofit>strategyprofit+(positionprice-close)*round(COUNTOFSHORTSHARES/positionsize) then

trailingstoptype=0

else

trailingstoptype=1

endif

fifi743

Participant

Topics: 9

Replies: 847

Been thanked: 401 times

03/16/2020 at 8:36 PM #122293

Paul

Thanks I will included it! Working on last part.

I’ve 2 different versions, one with the code in the manual above, other normal approach.

If using nnbarlimit=1 then both systems have the same result, as it should be.

Also 3 ways to calculate the boxes. pfff and an attempted fix to run correctly.

3 users thanked author for this post.

GraHal,

Francesco,

fifi743

Paul

Participant

Topics: 40

Replies: 812

Been thanked: 660 times

03/16/2020 at 9:51 PM #122295

Paul

Renkoboxes default uses close, I experimented with totalprice & Heiken Ashi.

Removed dynamic box. It added nothing good enough.

results are with spread 2.4 between 8am and 22pm.

Original renko based on “close” with boxsize_25, 6k

renko based on “Heiken Ashi” with boxsize__ 20, 10k

more important I still see differences between live & active orders from the backtest. Maybe because I’am running different versions. Tomorrow I will stick to one.