Three pairs hedging/Renko | Page 2 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Three Pairs Hedging 1,211 replies

RENKO Trading SYSTEM - RENKO Indicators 7 replies

Pairs Symbol Indi changer for Renko 9 replies

New Renko Bar and Reversal Renko Bar Indicator Alert Help Please 5 replies

Need help hedging Exotic Pairs 2 replies

Attachments: Three pairs hedging/Renko
Exit Attachments

Three pairs hedging/Renko

  • #21
  • Jan 6, 2012 7:25am Jan 6, 2012 7:25am
Quoting Indrek
As much as I understand from instructions, it is not usual hedging. If all trades - long EURUSD, long USDCHF and short EURCHF - are set to trigger after bullish candles then, if the first triggered entry is a good one, possibly other two will not trigger at all, and you just win the first one.

At least, that is how I see it logically. First entry is what you try to win and other two are just for backup.

But... I would not set SL if hedging three pairs like this. They equalize each other, and if money management is not too risky then sooner or...

You are correct Indrek, it is not a usual hedging. That’s why is different then The first approach of cash4u2. You are seeing the correct logic. Not using SL will also work, but with the risk of accumulating too many open orders that could become difficult to manage. Obviously, that could be avoid by using bigger pips range, thus with consequential longer time to come out in profits.

yes MM is everything in this type of system.
  • #22
  • Jan 6, 2012 8:59am Jan 6, 2012 8:59am
Quoting Indrek
No calculations needed at all. We may very well skip all math and use simply ATR values for each pair for setting pending orders.
And the system certainly needs no EA.
Here we may disagree a bit.
For the unforeseen future, and for systems like this (multi-pairs), I tend to trust the math more then any type of value given by any type of past history indicator. The bottom line is that price action does not respect any specific “past” level in a consistence basis. The probability of the price to move in any specific direction from “any” given point in the chart is always 50/50, regardless if it is setting in a MA, ATR, Pivots or whatever any other value you call it. But this would be another subject.

The math calculations is something that I have control over. At least one factor is known:
EUR/USD*USD/CHF= EUR/CHF, I can work with that, and make strategies around it. Then I decide the risk I want to take, at what point I cut my loss or take my profits. This is one reason why I will never understand how someone can blow a real money account.


I will never deny that your approach using values has validity. But we get to the same point.
To tests the validity with long periods of data, the system needs to coded (EA), then run it thru. Find the weakness and strengths, then decide if to use it with our real money.
  • #23
  • Jan 7, 2012 5:15am Jan 7, 2012 5:15am
Quoting leerees
Three pairs hedging doesn't work, if one of the pairs uncouples out of normal range your account will get blown to peices. This was one of the strategies I experimented with before I became profitable myself, I can assure you that you are barking up the wrong tree.

No doubht you'll follow the same path that I did and it will go like this:

1) add stop loss
2) add counter hedge pairs
3) remove counter hedge pairs and add martingale
4) end

Perhaps you never understood Three (or more) pairs Hedging correct.
It is impossible that 1 pair uncouples, I mean using the correct pairs and the correct lot size for each of them. (not possible with brokers that have 2 decimals lot size like most mt4 brokers). You could open 3 positions and keep them open years and never risk more then the initial spreads, swap and some tiny insignificant “inefficiency ” the market produce in a continuous basis, that sometimes for fraction of seconds absorb the spreads and place you P/L in profit. Obviously, there is no point in doing that. No risk, no profit.
Many have attempted to take advantage of those tiny inefficiency for many years but that is a complete different subject.
  • #24
  • Jan 8, 2012 3:25am Jan 8, 2012 3:25am
Hi jeuro,
Your strategy seems good. I have started demo and will post my results. Jeuro could please post your statement which will help us to understand more clearly. And what is your TP?
  • #25
  • Jan 8, 2012 9:05am Jan 8, 2012 9:05am
Quoting k4u
Hi jeuro,
Your strategy seems good. I have started demo and will post my results. Jeuro could please post your statement which will help us to understand more clearly. And what is your TP?
Hello K4u:

The details of TP and SL used for all 3 pairs are in post #12. But in top of that I used a GlobalEquityProfit of $20. (I think I mention that in other post). That means the following: When I started the account with $5000 and start opening trades, as soon the Equity reached $5020, "all" trades were closed, regardless of the amount of trades were open. Then, the next target was $5040.. etc.

But the account did not necessary kept that $20 increase in that order. Sometimes the balance would get ahead., say to $5100 but with a couple trades open in dd. say -$100, therefore Equity $5000. Later the system would reach the $5040 and CloseAll.

The system most of the time kept a very nice Equity curve. What is the most important part of any system.

I don’t know if I can post the statement. It is 40 pages long, over 750 trades. I doubt that re-viewing it would help anyone to understand better.

I you monitor your forward test, even a short time while executing a few trades you will understand the mechanics almost immediately.

Here it is the last page of the report, hardly fit the screen.
Attached Image (click to enlarge)
Click to Enlarge Name: last page.png Size: 62 KB
  • #26
  • Jan 9, 2012 12:24am Jan 9, 2012 12:24am
Hi jeuro,

As I'm sure you're aware backtesting using MT4 on Renko charts can be "sort of" done with a bit of mucking around, but trying to backtest multi-currency EA's is even more difficult (although not impossible). If you did away with the offline Renko charts and incorporated the appropriate maths into the EA to remember the Renko steps and act accordingly then the whole thing could be rewritten for MT5 to test the multi-currency aspect. Is that what you're asking for, someone to rewrite it for MT5?

I have written many EA's for Renko charts in the past and believe I could probably incorporate it all into one EA in MT4 (given enough time) but I have no experience with MT5.
  • #27
  • Jan 9, 2012 4:39am Jan 9, 2012 4:39am
Quoting skoda2008
Hi jeuro,

As I'm sure you're aware backtesting using MT4 on Renko charts can be "sort of" done with a bit of mucking around, but trying to backtest multi-currency EA's is even more difficult (although not impossible). If you did away with the offline Renko charts and incorporated the appropriate maths into the EA to remember the Renko steps and act accordingly then the whole thing could be rewritten for MT5 to test the multi-currency aspect. Is that what you're asking for, someone to rewrite it for MT5?

I have written many EA's for Renko...
Hello Skoda.
Thanks for the input. You hit all the right nails.

Yes, yes, yes and “sort of”.

Yes, we are aware BT renko in mt4 can be done and quite well. With help from the developer of a commercial renko and Birt help, we have been BT-ing on renko with tick data for a while. Thus getting 99% quality Renko BT.

Yes, BT Multi-currency is a pain on the neck.

Yes, we could re-write our renko tool EA (in mql) to do away with the non-existing renko in mt5, and then ask someone to rewrite the whole thing just to test the multi-currency aspect of this specific strategy. ...and here is the “Sort of”..

We also do not have experience in mt5 nor in any other similar platform.
Time is precious for all of us, and do not want to take it from anybody else by just asking rewrite this for mt5 or rewrite this for Jforex etc., because we are not sure what platform is the way to go.
In spite that mt5, Jforex and other commercial multi-currecy Strategy testers share similar or same coding formats . All of them have issues and small differences and takes time to master them.

Now if someone comes up and say, hey I am experienced in coding and BT in mt5. Or in this other etc. and I can do it with little efford or little time, then we would immediately alter our mql EA, send it and share all the fruits of our work and ideas.

Thanks for your input again. Do you have any experience with any other platform that allows multi-pairs BT?
  • #28
  • Jan 9, 2012 6:19am Jan 9, 2012 6:19am
Quoting jeuro
Hello Tarrini, Thanks for re-posting here.


Yes, I do believe this system has a lot of potential to be a constant profitable system but needs fine tuning. But the fine tuning is a painful and slow task if we are unable to back-test. This is one of those cases that back-test will help, because is matter of math. Placing Pending orders manually will do the trick, but difficult to keep. Also keep in mind that for the given results I did use renko and it gets tricky when price retrace, say 50 pips and the go back up again. You need to replace the...

Hi Jeuro,

Thanks for sharing your modified version. Really interesting. Just one doubt: why after an up bar, we sell EUR/CHF? A Counter Trend approach just for this pair? and why not SL for EUR/CHF?
Anyway your results are impressive, yes, just one month, but impressive.

Thanks again
  • #29
  • Jan 9, 2012 8:39am Jan 9, 2012 8:39am
Hello jeuro,

I am trying to implement Your system on my own API (mutlicurrency, Dukascopy tickdata), so that I might be able to do an optimization, but I need to have the correct logic implemented to be able to replicate your results. I do not have December data in my software yet, so the replication might not be exact as I am running it on November

However, I have problems in replicating the results, even close. My code places 4 times as many orders, it overtrades, and when global TP and SL are used the results are heavily in red. I suspect I have some logic error, or maybe not all was explained or understood by me. If you could post a link to the myfxbook site of Your system or attach the transactions file here that would be really helpful in examing the logic.

I attach two backtests results: BBS, SSB, modest MM and effectively no global closures, so the equity results are dependent on EUR/CHF behaviour. Maybe you spot what's wrong by looking on transactions list. I have not implemented swap, the slippage is 0.3 pips on every order and commission is 5$ per lot.

And of course, thank you for posting the system!

Regards,
Qooyend
Attached File(s)
File Type: rar JeuroTS_SSB.rar   355 KB | 1,086 downloads
File Type: rar JeuroTS_BBS.rar   358 KB | 872 downloads
  • #30
  • Edited 11:09am Jan 9, 2012 10:00am | Edited 11:09am
Quoting jeuro
Hello Tarrini, Thanks for re-posting here.


Yes, I do believe this system has a lot of potential to be a constant profitable system but needs fine tuning. But the fine tuning is a painful and slow task if we are unable to back-test. This is one of those cases that back-test will help, because is matter of math. Placing Pending orders manually will do the trick, but difficult to keep. Also keep in mind that for the given results I did use renko and it gets tricky when price retrace, say 50 pips and the go back up again. You need to replace the...

Hi, thanks for sharing your system


Lordorly
  • #31
  • Jan 9, 2012 10:55am Jan 9, 2012 10:55am
Quoting Caillou
Hi Jeuro,

Thanks for sharing your modified version. Really interesting. Just one doubt: why after an up bar, we sell EUR/CHF? A Counter Trend approach just for this pair? and why not SL for EUR/CHF?
Anyway your results are impressive, yes, just one month, but impressive.

Thanks again

Answer in post #17.
That is the calculated risk we are incurring.
  • #32
  • Jan 9, 2012 11:05am Jan 9, 2012 11:05am
Quoting jeuro
Answer in post #17.
That is the calculated risk we are incurring.
Thank you but this answer has no edge at all for me. Just Sell EUR/CHF after a 5 pips Up Renko Bar? And your risk is there? Sure, I see the risk but just that, the risk.

Anyway thank you and good luck.
  • #33
  • Jan 9, 2012 11:46am Jan 9, 2012 11:46am
hi

i would like to say my opinion on this method, because i used it a while ago ...and i am waiting for an ea to make it less boring ....

this is what i do :
open currency ring ( buy eurusd , buy usdchf e sell eurchf ) same lot size no stop loss or take profit .

place a grid of pending orders every 20 pips this way :


open buy eurusd 1.3020
pending buy 1.3040
pending buy 1.3060
pending buy 1.3080 and so on

open sell eurchf 1.2120
pending sell 1.2100
pending sell 1.2080
pending sell 1.2060 and so on

same for usd chf , pending buys, because it was a buy in the opening .

i increase the position of the best performer for example if eurusd after a while is + 20 pips , i add another eurusd and try to close the basket in profit .

if it retraces ther will be for sure another cross of the ring the will make 20 pips in positive and will hedge again...and so on

it may take a while to close a basket , but it usually does , maybe 1 or twice a day .

try to think if you run this method on 4 or more rings, and you can have a good pipping day....
  • #34
  • Jan 9, 2012 12:09pm Jan 9, 2012 12:09pm
Quoting qooyend
Hello jeuro,

I am trying to implement Your system on my own API (mutlicurrency, Dukascopy tickdata), so that I might be able to do an optimization, but I need to have the correct logic implemented to be able to replicate your results. I do not have December data in my software yet, so the replication might not be exact as I am running it on November

However, I have problems in replicating the results, even close. My code places 4 times as many orders, it overtrades, and when global TP and SL are used the results are heavily in red. I suspect I...
Hello Qooyend.


It is difficult to reverse engineering the way your are doing your BT by just looking the statement.


Yes, way too many trades compared to December.


These are some questions that come to mind:
Are you using renko bars or range bars? Range bars would give you a lot more trades then renko, and would not keep the math meant for a renko size 11/11/5 .


If you are using a different script or code to mimic the renko, make sure it takes the trades as a renko would print in the Chart. You probably know that Renko is not a plain grid , of 11 pips in this case, because When price reverse, it takes 22 pips to print a bar. ..and all that goes into factoring the numbers.


I think that using or not using Global stops would affect the results but not the number of trades. Lets focus in the number of trades first.


If you get December data, we could use skype or teamviewer to set and run your api. We should be able to replicate results in a very near ballpark. Once we do, then we would be ready to run a longer BT.
  • #35
  • Jan 9, 2012 2:53pm Jan 9, 2012 2:53pm
Quoting jeuro
Hello Qooyend.


It is difficult to reverse engineering the way your are doing your BT by just looking the statement.


Yes, way too many trades compared to December.


These are some questions that come to mind:
Are you using renko bars or range bars? Range bars would give you a lot more trades then renko, and would not keep the math meant for a renko size 11/11/5 .


If you are using a different script or code to mimic the renko, make sure it takes the trades as a renko would print in the Chart. You probably know...
I am coding the Renko behaviour, and I am aware of the difference between Renko and Range bars. It is possible that there is a bug in the code, I'll check on that. Apart from this I will have to get the current tick data and run the December test, but it'll take a couple of days.

Could you export Your statement from myfxbook into cvs, zip it and attach here? It would be helpful.

Regards,
Qooyend
  • #36
  • Jan 9, 2012 3:12pm Jan 9, 2012 3:12pm
Quoting Caillou
Thank you but this answer has no edge at all for me. Just Sell EUR/CHF after a 5 pips Up Renko Bar? And your risk is there? Sure, I see the risk but just that, the risk.

Anyway thank you and good luck.
Sorry Caillou, I don’t know how to explain better but this may help you and others.

If we hedge with 1 pair, say buy EUR/USD and sell EUR/USD simultaneously 0.30 lot. We have no risk, we lose the spreads. Will never make any money. No risk, no profit. No point in doing that.

If we hedge with 2 pairs EUR/USD and USD/CHF, say buy EUR/USD and BUY USD/CHF (buy, buy because they are negative correlated pairs) it is a very risky hedging. Therefore You can make a lot or lose a lot. Historical correlation (in my opinion) is not a good base to develop a system. The risk in unknown. High risk, High profit, no point in doing that. (unless you want )


If we hedge with 3 pairs EUR/USD, USD/CHF, EUR/CHF (some call them rings because in essence EUR/USD multiplied by USD/CHF will always equal EUR/CHF ) mathematically we can create a perfect combo of lot sizes that would make a “perfect” hedge if we place the trades simultaneously. Again we lose the spreads (more in this case) but again No risk , no profit. No point in doing that.


If we want to make money, we need to incurred risk. So, we need to take the 3 pairs out of equilibrium. As as part of designing the system, after making some basic math calculations and considering the 3 pairs equation, renko size, individual TP, individual SL, Lot size for the Global TP, normal spreads etc. I placed the risk in the EUR/UCH. ( I could have place it in any of the other 2). It is a calculated risk based in that the equation of the 3 pairs will never change.

And that was the answer I gave you.

I can not help if you “just” see risk in the counter trend sell every 5 pips in the EUR/CHF disregarding the fact that at least 1 of the other pairs, would close a 15 pips TP before the EUR/CHF would trigger a second entry.
  • #37
  • Jan 9, 2012 4:29pm Jan 9, 2012 4:29pm
Quoting tarrini65
hi

i would like to say my opinion on this method, because i used it a while ago ...and i am waiting for an ea to make it less boring ....

this is what i do :
open currency ring ( buy eurusd , buy usdchf e sell eurchf ) same lot size no stop loss or take profit .

place a grid of pending orders every 20 pips this way :


open buy eurusd 1.3020
pending buy 1.3040
pending buy 1.3060
pending buy 1.3080 and so on

open sell eurchf 1.2120
pending sell 1.2100
pending sell 1.2080
pending sell 1.2060 and so on

same for usd chf ,...
Let us know the results of your forward test after you get your EA going.
  • #38
  • Jan 9, 2012 7:42pm Jan 9, 2012 7:42pm
Quoting jeuro
Do you have any experience with any other platform that allows multi-pairs BT?
No, sorry. Although, this platform looks like it might be worth learning - with an API to VB.net

http://www.fastbrokers.com/pathfinder/
  • #39
  • Jan 9, 2012 9:03pm Jan 9, 2012 9:03pm
Does the fact that eur /chf ' floor is pegged to 1.20 make any difference in its relation to the other two pairs being considered here???

thank you
  • #40
  • Jan 10, 2012 12:10am Jan 10, 2012 12:10am
Quoting Gothamcity
Does the fact that eur /chf ' floor is pegged to 1.20 make any difference in its relation to the other two pairs being considered here???

thank you
It makes no difference to the equations... meaning EUR/USD*USD/CHF=EUR/CHF will continue to be the same.


But, we should keep in mind facts like this. The more facts/factors we know, makes it easy to minimize the risk or maximize profits. On September 5 last year we didn’t know the Swiss were going to intervene and place that floor limit on September 6. Therefore E/C spiked 1000+ pips. Now we do. Therefore I doubt will spike again.


Perhaps at the present E/C level of 1.21 we will have The E/C ranging between 1.22 and 1.20 and that would be nice for this set.


The risk of interventions/spikes is not because the equation change, rather it is because spikes causes, fill problem, non execution of trades, slippage etc. And that mess up the math.
1 trader viewing now