IndoRun exit solution

[name000]

As you all probably aware that Indo Run is great EA, you are also aware that it lacks good exit strategy when news filter miss to filter because position is already open or when some Asian session spike happen. IR is vulnerable with floating positions and high impact news announcements are incoming. No matter how good/safe setup you have, it is always possible to extend the number of open positions and put your account into more dangerous position. I must say that this could happen extremely rarely, but I would rather avoid that situation on my accounts and try to heal before it happens.

 

I found one simple exit strategy which could be integrated with IR perfectly.

It's modified 'anti-martingale hedge'. With example bellow I believe it will be much easier to follow.

 

Assuming that IndoRun opened buy position on [email protected], lot 0.1. TP is on 1.5745. Now, instead to place buy order (as its currently does) if price go in other direction (1.5715), it should place sell order with 15 pips target (SELL 1.5715, TP 1.5700, lot 0.3). So, if price reaches 1.5700 both should be closed, with $15 profit (BUY -30 and SELL +45).

http://img816.imageshack.us/img816/8154/antimartingalehedge.gif

Further more, if price initiate sellstop on 1.5715, and trace back up after that, we should have ready BUYSTOP order on 1.5730 (BUY 1.5730, TP 1.5745, lot 0.6). So, if price reaches 1.5745 all open orders should be closed and score would be 0.1 and 0.6 in profit with 15 pips ($15+$90), and 0.3 in loss with 30 pips (-$90). Total will be again +$15 profit. Further more, if market reaches and open 3rd position ([email protected], lot 0.6) and goes down again, there should be SELL order ready on 1.5715 (TP 1.5700, lot 1.2)...

 

Drawback of this strategy is if price goes too long sideways, reaches buy and sell stops and doesn't reach TP either side. That way, we would come up to margin call at some point. Good side of all that is that GBPUSD average daily range is approximately 150+, and when IndoRun 'misses' direct hit, it is usually because position is opened before some high impact news or there is some unexpected movement on market. Good this also is that IndoRun already have dynamic profit (which could be activated after e.g. 3 open positions) and break even (which could be activated after 4 open positions).

 

I believe this addon will make IR much more safer, and it could trade in more volatile market time. Also, it should significantly lower down avarage holding time, which will give the EA more time for new trades.

 

Unfortunately I'm not programmer, other way I would try to do that by my self... I really hope that somebody will jump into this and make it work, because I smell something really good. I contacted expat with this about 2-3 weeks ago, but he is unfortunately too busy with some other projects. And as I would rather wait for him to do that, I can't wait anymore to see what will become of this.

 

Numbers that I presented are only for explanation purposes, and we should tweak/test after this is added further more (I didn't count spread, but that was not the point here).

 

Look forward to to hear other suggestions and opinions about this idea.

Edited January 28, 2011 by name000

[askalas]

Hi name000, here some feedback. If I understand you correctly your suggestion means that you no longer use "baskets" with one-direction (long or short) orders. Instead you add opposite orders until goal is reached. A completely different strategy than IR (although I must confess that I haven't tried the martingale functionality in IR). IR only adds a same-size order if price goes against the basket. This is why it can take fairly large movements in the wrong direction.

 

In your example above, the next order would be 1.2 lots and bringing your total lots to 2.2 lots (even though exposure is no more than 0.8 lots)? Thats 22 times the first order only after 4 orders :) In comparison, IR only gets you total lots and exposure of 0.4 lots after 4 orders in a basket. And after 5 orders... Your example above is probably a context where IR is best suited, ie a ranging market.

 

I don't mean to be all negative, it obviously don't suffer from the same problem as IR. I just don't see it as a fix to IR but as a different strategy. One that can probably be quite good if adding some market condition filters. Maybe I'll do an EA for this as "proof of concept" next week.

 

Thanks!

[name000]

Hi thanks for your feedback.

IR only adds a same-size order if price goes against the basket. This is why it can take fairly large movements in the wrong direction.

Fairly large movement is right word. With large movements, account will be blown away. Of course, it depends on your MM, but if you have 16 (default) price offset, and your MM allows approximately 25 open positions (movement of 400 pips), where you could be quite safe... but, when you can expect (or rather guess) that market will retrace to have your BE after 250-300 pips movement (which is quite usual according to statistics)? That is completely unacceptable position because you have no control over it at all.

In your example above, the next order would be 1.2 lots and bringing your total lots to 2.2 lots (even though exposure is no more than 0.8 lots)? Thats 22 times the first order only after 4 orders :) In comparison, IR only gets you total lots and exposure of 0.4 lots after 4 orders in a basket. And after 5 orders... Your example above is probably a context where IR is best suited, ie a ranging market.

Even after 4th position open (and 2.2 lots in total), you can control profit/loss and you can use BE or fixed stop loss if goes in 'wrong' direction. With 22 open positions is same direction, you can only start to pray (usually it doesn't help on FX). There shouldn't be 5th position in this strategy. Everything must be solved with 4th position open. BE, or fixed stoploss. Something like that. I would rather take 10% loss than margin call.

I don't mean to be all negative, it obviously don't suffer from the same problem as IR. I just don't see it as a fix to IR but as a different strategy. One that can probably be quite good if adding some market condition filters. Maybe I'll do an EA for this as "proof of concept" next week.

Thanks!

I agree with you it will be quite different strategy. IR is set for ranging market and with this addon it will be more prepared for volatile movements and it should avoid ranging time (like assian session).

 

Session and news filters are must for this kind of strategy. That is why I just wanted to add this to IR. If you know how to add this, and with proper optimization, we can make this happen.

[askalas]

OK, I'll see if I can make a simple proof-of-concept EA from this strategy tomorrow.

[askalas]

Here's a quick'n dirty testmod using IR1.5. Its certainly not perfect but give it a testrun and see if its is near what you were aiming for...

 

http://www.4shared.com/file/Fj7ogPAk/Indo_Run_testmod_askalas.html

[name000]

Thanks for your quick turnaround. Order open works good. Profits are doubled with every new position open. Maybe we should keep it to ProfitAllOrder value (I don't know how now its calculated)?

One idea more... is it possible to just reverse what triggers in EA to place initial buy-sell orders?

 

Great start!

 

edit: I figured what is happening with profits. ProfitAllOrder is multiplied also. E.g. when there is only one open order, and its closed, distance from open to close is ProfitAllOrder value. When two orders are open, distance between second open and closed position is tripled. Profits should stay to original value set with ProfitAllOrder, no matter how many orders are open.

Edited January 30, 2011 by name000

[askalas]

I've added your suggestion and reuploaded the file to the same place. Check param 'ReverseInitalOrders' to control inital buy-sell orders. As for double profit every position, set FixedProfitAO to true and all ATR-related profits to false. This is a quick'n dirty one so I've not considered the advanced functionality such as ATR-related TP. If the core concept of this strategy seems solid we can add more functionality and make more solid EA out of this.

[name000]

Basic concept is now looking great! This is it!

About reverse option I mentioned, I meant to revert condition which puts initial buy-sell orders (take a look attached image). Orders are not placed if indicators doesn't send trigger and 'overall filter status' is filtering. You will notice that initial buy-sell orders are placed mostly on ranging time, and its filtered out on volatile time. So I thought that we could just reverse how its filtered out and to get good result?

http://img11.imageshack.us/img11/5158/ofsfiltering.gif

 

 

What you added (ReverseInitalOrders) will be also useful while we testing it, don't remove it.

[askalas]

Alright, same place, new file, new option (ReverseFilters) and some other bug fixes.

 

Makes sense to reverse the filter :) Didn't reverse newsfilter.

[jonnnny23]

Great work on increasing the capability of the EA.

 

Is the Price offset is working properly with the new settings?

 

If you are BUY, the price offset should be -16 (for a new SELL)

 

If your are SELL, should be +16 (for a new BUY)?

 

Or am I misunderstanding?

 

http://i53.tinypic.com/2wgc5rc.jpg

 

PriceOffset: -16

ReverseInitialOrders: False

ReverseFilters: True

 

2010.01.04-2010.10.05

[askalas]

Hi jonnny23!

 

Tired today but I'll try to answer your question. You are correct in that distance(step) to new opposite order is PriceOffset. But only if you haven't enabled ATRStepOn/ATRPipsToStep and StepFactorOn. Looking at your screenshot there shouldn't be a "Sell 3" with this mod as "Sell 2" will have 3 times the lotsize of "Buy1". So basically, you are just waiting for TP when going south from "Sell2".

 

And guys, please post your results (I can see that several people have downloaded the mod). It is especially interesting to see difference from original version. It is by no means a finished EA but by identifying where it fails in comparison to original version we might be able to make a merged version that uses both strategies depending on market conditions. But I just don't have the time to test this myself.

 

One way could be to find a shorter testperiod where original IR fails. Don't use any filter as we want to dig into the basic characteristics of this. Then use this modversion on the exact same testperiod. How does it compare? I've made a version where this mod "takes over" when original fails but it needs work.

 

How is it going for you name000?

[askalas]

I just noticed that I can't send any PM because I'm such a newbie. name000: sorry, just noticed msg in my inbox. So far I can read but not answer PM. If you wan't you can send me PM with your email for future private correspondance.

[askalas]

How many posts do one have to make to enable PM... ;)

[name000]

How many posts do one have to make to enable PM... ;)

Sorry about that, I forgot you can't respond to it yet. I'll send you email.

I just didn't wanted to post here ping-pong correspondence, which is not that much important.

[fx4a]

Hi,

Just running this week after seeing this interesting mod.

All my back testing fails using default though.

Below are results so far on demo.. Hope it's some use

 

Gross Profit: 457.24 Gross Loss: 261.63 Total Net Profit: 195.61

Profit Factor: 1.75 Expected Payoff: 10.87

Absolute Drawdown: 0.00 Maximal Drawdown: 111.27 (9.78%) Relative Drawdown: 9.78% (111.27)

 

Total Trades: 18 Short Positions (won %): 8 (50.00%) Long Positions (won %): 10 (90.00%)

Profit Trades (% of total): 13 (72.22%) Loss trades (% of total): 5 (27.78%)

Largest profit trade: 107.37 loss trade: -111.27

Average profit trade: 35.17 loss trade: -52.33

Maximum consecutive wins ($): 5 (146.69) consecutive losses ($): 1 (-111.27)

Maximal consecutive profit (count): 146.69 (5) consecutive loss (count): -111.27 (1)

Average consecutive wins: 3 consecutive losses: 1

[name000]

It is especially interesting to see difference from original version. It is by no means a finished EA but by identifying where it fails in comparison to original version we might be able to make a merged version that uses both strategies depending on market conditions.

Without indi's it can't survive for a long time. I used CCI and momentum, I also had to limit trading time from 6am - 18pm (GMT). I was trying to find combination which will lower number of open positions to no more than 4. Result is not that bad. First tests were done in one month period.

Best result was: 1 position open/close - 48%, 2 positions - 16%, 3 positions - 19%, 4 positions - 12%, 5 positions - 4%, 6 positions - 0%, 7 positions - 1%. Profit with this setup was also encouraging - 31%.

I will continue to test this setup through whole 2010. I'm concerned how we will control EA so it can not have more than 4-5 positions open. Is it possible to have dynamic BE? Or to have at least one more option for BE so we can lower it down if we get in bad position. In this setup, BE is activated after 4th open position. But when 5th position is open, we need to make some 'soft' wall there, so we won't break our teeth :).

 

To make these percentages better, we will need better trend/ranging indi. Maybe Schaff (I didn't try it so far, but I read its good), maybe to just add simple H-high and H-low for previous 10 candles and volume condition?

 

I would really like to hear some suggestion since I'm not sure which indi is best to use in order to detect trend or sideways.

 

Test and setup is here:

http://www.mediafire.com/?gwrxgmtanhbudp6

[askalas]

Not too bad results there. To try and stimulate ideas here is a combo-version of the two strategies. It moves from original IR to this modversion after a basket size is larger than x orders. I haven't made a proper lotsize calculation when moving to alternative method but it servers as a fairly good example of how this can be done.

http://www.4shared.com/file/pTdMHjor/Indo_Run_testmod_askalas_combo.html

Edited February 1, 2011 by askalas
Fixed link

[fx4a]

askalas...check this pdf diagrams and explanation..not long..better way to hedge

 

http://www.multiupload.com/8M0D2CGW11

[name000]

fx4a, same thing is in 1st post... that is the base of this mod.

[fx4a]

Yeah so it is. The whole think looks logical and you would think it would work. This hedge method was used with the pendulum EA originally. That EA was specifically set for EU, because of the lower spread. Nobody I know got it to work properly.

[name000]

I was testing combo and first idea, but none of them worked safe enough with any setting I tried. I also tested dozen indi's which I thought we could use to better filter out sideways... I couldn't decide which one will be good because each one should be tested with rest of the code, otherwise we can only guess. I can't ask anyone to do this further only by a guess, so I call this (from my self at least) a draw... Thank you askalas for your help.

[⭐ E2U]

Hi

I have tested “Indo Run testmod askalas” with 0.02 lot and TP 1.5 Euro.

I have run it for 4-5 weeks and only on EURUSD.

Profit between 50-100 Euro/week

Max DD last week 40 Euro

[Richo]

I tried to download the files “Indo Run testmod askalas” that askalas put up but its no longer valid, can someone please re-upload

[⭐ E2U]

I wonder how many good EAs has gone to the bin thanks to back tests. I don't believe in back test at all.

Need to be forward tested before gives up.

I hope you guys continue to work on this EA.

 

http://www.4shared.com/file/cunTDqM_/Indo_Run_testmod_askalas__1_.html

[danny_pip]

Guys, I'm running Indo life.

I advise you to switch on the martingale function; it's giving faster profit and does not stay endlessly long in the trades