help modification of an EA very profitable | Page 3 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Need help for this very very simple indicator 1 reply

Very Useful ATR Indi: Needs modification 8 replies

Very simple modification required 1 reply

I need help with very profitable EA 6 replies

profitable EA needs modification 1 reply

Attachments: help modification of an EA very profitable
Exit Attachments

help modification of an EA very profitable

  • #41
  • Jul 3, 2010 6:27am Jul 3, 2010 6:27am
Quoting akukaya
Consider martingale only if:
1)Your system tested and the occurance of consecutive straight loss almost always <7, martingale is a better strategy compare to fix lot.
Disagree. There is no difference. Martingale merely redistributes wins and losses**. I would need to see a solid mathematical proof to convince me otherwise. If you can't use martingale to gain an advantage at the casino, why would it be any different with forex?

3)You can bank in profit very high percentage return >100% in quick period of time, basically before reach death trade...
Disagree on the grounds that you can't possibly know when the death trade is going to occur. And withdrawing money reduces your ability to survive future losses (i.e. the number of martingale levels your account can cope with). Also, martingale (as I understand it) doesn't increase % return; it is merely a way of recovering from a recent sequence of losses by way of a single trade.

compare it with fix lot. you will get up and down results, end of the same period, you may slightly increase/decrease your account or breakeven. Some will also meet margin call.
Yes but with small fixed lot (e.g. 1% risk per trade) you have time to shut down the EA and escape with your remaining capital. With martingale, one death trade creates a margin call instantly and unexpectedly.

[** the XLS attached to this post demonstrates how and why this is the case].


Quoting Adal
Sorry for misunderstanding you.
No problem. Just expressing my opinions.

Yes, it doesn't provide an edge. But it makes the difference between getting 1% or getting 10% annually from your edge.
Disagree. You could simply trade larger fixed positions to achieve the same thing. Everything else being equal, variable sizing merely redistributes wins and losses**.

You make it sound like people should not worry about finding an optimum MM at all.
Optimum MM is a balance between maximizing return and minimizing risk of ruin, and can be achieved equally well by using fixed lot sizes.

A black swan will get you anyway, even with a very conservative MM. That's what Taleb argues.
Yes, but if you size at (for example) 1% then it will take a large number of losses to ruin you, and you have time to bale out, and escape with your remaining capital. With martingale, you can survive only a few losses because of the exponential sizing (1+2+4+8+16+.... etc), and if you're running an EA that trades intraday unattended, the death trade losing sequence can ruin you before you have a chance to react.

I could use a MM system where I would double down 2 or 3 times, but I would cut my loses there, at a maximum 10% account drawdown.
If you double down 3 times, then after 3 losses you've lost 1+2+4 = 7 units. Now, by returning to 1 unit sizing, you need 7 consecutive wins to compensate for the 3 losses. Needing a high ratio of wins makes your recovery system work unnecessarily hard, IMO.

[** the XLS attached to this post demonstrates how and why this is the case].


Anyway, I'm merely restating the same points, and reposting the same link, that I did earlier. If you good folk disagree, well that's fine, let's just leave it at that.

I wish you all the best with your trading.
  • #42
  • Jul 3, 2010 6:55am Jul 3, 2010 6:55am
Quoting hanover
Anyway, I'm merely restating the same points, and reposting the same link, that I did earlier. If you good folk disagree, well that's fine, let's just leave it at that.
I downloaded the Excel file and looked at it.

You are correct that MM doesn't change the expectancy. I agree with that. But this is not what I've said. I think that we are talking about two different issues.

What I'm saying is this: Assume that I have a system which outputs a number between -1 and 1 which represents a prediction of future price direction. The value of the number represents the quality of the prediction (or it's "confidence" level if you want). -1 and 1 represent maximum conviction.

Are you saying that all MM are equivalent here? What if I have a MM which takes the "confidence" level into account? So that it bets proportionally with this "confidence" level. I'm sorry, but all the books and research papers I've read suggest that in such a case MM does make a difference. I didn't try to follow all the math, but I trust their conclusions.

When I said that I would double down 2 or 3 times, it was because in those instances the "confidence" level of my system will increase.

http://tradermike.net/2005/07/position_sizing/
  • #43
  • Jul 3, 2010 7:10am Jul 3, 2010 7:10am
Quoting Adal
I downloaded the Excel file and looked at it.

You are correct that MM doesn't change the expectancy. I agree with that. But this is not what I've said. I think that we are talking about two different issues.

What I'm saying is this: Assume that I have a system which outputs a number between -1 and 1 which represents a prediction of future price direction. The value of the number represents the quality of the prediction (or it's "confidence" level if you want). -1 and 1 represent maximum conviction.

Are you saying that all MM are equivalent...
I like your posted link..

I agree with most of your state opinion.
  • #44
  • Jul 3, 2010 7:35am Jul 3, 2010 7:35am
Quoting hanover
1)Your system tested and the occurance of consecutive straight loss almost always <7, martingale is a better strategy compare to fix lot.
Disagree. There is no difference. Martingale merely redistributes wins and losses**. I would need to see a solid mathematical proof to convince me otherwise. If you can't use martingale to gain an advantage at the casino, why would it be any different with forex?

I am not mathematician, cannot proof you with mathematical model.
BTW martingale at casino is totally different with martingale in forex.
1)There are more variants of martingale ideas in forex
2)Forex with technical analysis/fundamental analysis that can built the "confident" level or understanding in market movement via past data.


3)You can bank in profit very high percentage return >100% in quick period of time, basically before reach death trade...
Disagree on the grounds that you can't possibly know when the death trade is going to occur. And withdrawing money reduces your ability to survive future losses (i.e. the number of martingale levels your account can cope with). Also, martingale (as I understand it) doesn't increase % return; it is merely a way of recovering from a recent sequence of losses by way of a single trade.

No, disagree. your initial deposit should already cover your overall level. Any profit can be bank in and should not interfere with your sequence.
if it frequently over your sequence(death trades), meaning you not find a suitable system to make use of martingale.


compare it with fix lot. you will get up and down results, end of the same period, you may slightly increase/decrease your account or breakeven. Some will also meet margin call.
Yes but with small fixed lot (e.g. 1% risk per trade) you have time to shut down the EA and escape with your remaining capital. With martingale, one death trade creates a margin call instantly and unexpectedly.

I remember the movie " BRAVE HEART".
There is one scene before the fighting, coward guy afraid and want to return back. Wallace speech is very good: " Yeah, you can choose to fight or going back home, if you back home, one day you will still die on your bed! If you fight, you may die but you will die with dignity!

Haha..may not suitable in this scenario, but if you continue small fixed lot, you still die as dying in the bed!!


[** the XLS attached to this post demonstrates how and why this is the case].
  • #45
  • Jul 3, 2010 8:00am Jul 3, 2010 8:00am
Quoting Adal
Martingale like techniques can be useful, but you have to know what you're doing. First, they should be employed only in relation with another statistical edge, like hanover says.

I have an experiment proposal for you. You said that you couldn't find more than 7 levels. There is a subtle trap here, the fact that that number, 7, occured at one particular point in the past. Before that point, the maximum number of levels was 6. And sometimes even before that, it was 5. And somewhere in the next years that number will grow to 8. If you really want...
thank you so much this is one very constructive post .
I SEE YOU !!
  • #46
  • Jul 3, 2010 8:18am Jul 3, 2010 8:18am
Quoting akukaya
[color=Blue][i]1)Your system tested and the occurance of...
I started to like this thread , some debates and intelligent people are posting , and I thank every one in this conversation.

1. for the casino part I disagree , my view is that its random , it has no rules behind it,.
But in forex there are rules, and the very basic rule is that its a market with buyers and sellers and prices vary based on the supply and demand theory . it doesn't vary randomly , other wise we would see EU @ 1.2400 and next minute 1.5000 and minute after 0.8000.

that's why if you entered a casino making only one bet on any game there you have near zero chance of wining
but if you enter forex market making only one bet you have 50% chance of success .



2 . I agree with you with that martingale is a way to trying to recover from a losing trade but what do you say if the percentage for example is 97% of success in recovering and 3% losing the sum of 20 wining trades . then mathematically its giving you an edge .
assume you win 50% and lose 50%
in 200 trades 100 lose
out of the 100 lost trades you managed to recover 97
and 3 made you lose sum of 60 trades
then in total you managed to increase the efficiency of your trade by 20%
and won 140 trades out of 200.

and converted a 50-50 system to a 70-30 .
I SEE YOU !!
  • #47
  • Jul 3, 2010 1:26pm Jul 3, 2010 1:26pm
Quoting dbfusion
I started to like this thread , some debates and intelligent people are posting , and I thank every one in this conversation.

1....
I think the points made about the cons of martingale were pretty valid. Eventually there will be a death trade. Adjusting the bet sizing, doesn't provide an edge either as stated. It makes you a little money, at the ever increasing cost of blowing up your account.. it is pure gambling. I think martingale would only be useful if you could get the probability of the death trade to be so low, that such dangers of blowing up your account, would be less then say, dying in a car crash or something. With 7 steps, its really only similar to say, getting bitten by a spider, or seeing lightning, or something. Perhaps if you could figure a way to get 20 or 30 steps, it may be a viable long term way to trade. Simple math in my head, I think it would be very hard to get that many steps. Maybe there's another solution though, I don't know. Sure is a fun idea, but would be pretty scary to risk real cash on, unless you really got the probabilities of busting to be almost nothing.
  • #48
  • Jul 3, 2010 2:35pm Jul 3, 2010 2:35pm
Quoting Adal
Are you saying that all MM are equivalent here? What if I have a MM which takes the "confidence" level into account? So that it bets proportionally with this "confidence" level. I'm sorry, but all the books and research papers I've read suggest that in such a case MM does make a difference. I didn't try to follow all the math, but I trust their conclusions.
At the risk of sounding pedantic, I've already agreed with you on this point. See my point #3 in post #38:

"3. Position sizing in itself can't provide any kind of edge (unless you can somehow calculate the probability of each setup in advance, and place larger bets on higher probability setups). Otherwise, all sizing variations do is redistribute wins and losses."

That's how blackjack pros during the 1970s/80s beat the casinos - by placing higher bets when there were an excess of high cards remaining in the deck, and hence the probabilities were in their favor (i.e. greater probability of obtaining a natural 21, and dealer busts more often). However, their bet size was always commensurate with the odds being offered by the cards showing and remaining, i.e. not a system based on their previous history of wins and losses. When you already have an edge, doubling up to recover from prior losses does not alter the probabilities on offer; it merely increases the risk of blowout. That is a major problem I have with martingale: the bet size rapidly grows to the point where it's no longer commensurate with probabilities.

The analogy with trading is hopefully obvious: when you have a high probability setup, you trade (bet size = 1); when the market direction is indeterminate, you don't trade (bet size = 0). You could of course, vary bet sizes (1,2,3,...) according to the quality of the setup, if you're somehow able to determine this quality. Or add to an existing trade if the probabilities subsequently improve (exactly like doubling down in blackjack), and subtract from it (scale out) if they worsen. But you'd need to have a way of of determining these probabilities in an activity whose underlying drivers are intangibles like fear and greed, rather than the turn of a card.

The difference between blackjack and roulette can help us to understand sizing in forex. You can (well, at least you used to be able to, when there were fewer countermeasures in place) beat the casino at blackjack, because it's possible to gain a statistical edge by counting cards and applying an appropriate betting strategy; but (unless the wheel is somehow physically biased) it's impossible to gain an edge with roulette. The same principle applies to forex. If your entries and exits provide an edge, you will win, for the same reasons as the blackjack pro. If you don't have an edge, no betting system can save you, just like the roulette player.

The Van Tharp comments in your link bear out all of the above. The % risk model beats flat size because it exploits the effect of compounding. The % volatility model further improves bottom line because it is exploiting underlying probabilities, just like the blackjack pro does.

My summary is this: Varying bet size commensurately with the probabilities will improve your bottom line, and is therefore a smart thing to do. In complete contrast, increasing bet size merely to recover recent losses lacks mathematical validity, and doing so without limit increases the risk of ruin exponentially.
  • #49
  • Jul 3, 2010 2:59pm Jul 3, 2010 2:59pm
Quoting akukaya
.....
I am not mathematician, cannot proof you with mathematical model.
BTW martingale at casino is totally different with martingale in forex.
1)There are more variants of martingale ideas in forex
2)Forex with technical analysis/fundamental analysis that can built the "confident" level or understanding in market movement via past data.

You can apply different Martingale variants, or betting systems, just as easily to casino games. However, it's true that the 'gameplay rules' in forex are much more open-ended (no house limits; massive liquidity and scalability; add to, or subtract from, open trades; vary the trade odds by applying different TP and SL levels; etc). I also agree that with forex you can use analysis to gain an edge (just like with blackjack - see my comments in post #48). But I would respectfully argue that MM is pure math, and the same math concepts apply to any activity that involves uncertainty, i.e. to both casino games and forex.

No, disagree. your initial deposit should already cover your overall level. Any profit can be bank in and should not interfere with your sequence.
OK, fair enough, it depends on your system.

if it frequently over your sequence(death trades), meaning you not find a suitable system to make use of martingale.
I believe that all martingales are 'unsuitable' because the possibility of the death trade, no matter how small, is always there. But if you're happy living with that kind of risk, it's a personal decision.

I remember the movie " BRAVE HEART".
There is one scene before the fighting, coward guy afraid and want to return back. Wallace speech is very good: " Yeah, you can choose to fight or going back home, if you back home, one day you will still die on your bed! If you fight, you may die but you will die with dignity!

Haha..may not suitable in this scenario, but if you continue small fixed lot, you still die as dying in the bed!!
Agree, you will die one way or another, but only if you don't have an edge.
And also agree, Braveheart was a great movie!

____________________

@dbfusion: sorry, looks like we've hijacked your thread. However, the constant bumping is keeping it at the top of the forum for you. Hopefully a programmer will see it, and offer to code your EA.
  • #50
  • Jul 3, 2010 3:29pm Jul 3, 2010 3:29pm
There may be a EA that could help you backtest your theories, its very similar to what ur trying to achieve. Its a commercial EA so can't mention it's name here, but u can get a free demo. I don't use it myself but i use another in their stable.

PM me and i will send u the link.

mike
Believing is seeing
  • #51
  • Jul 4, 2010 8:31am Jul 4, 2010 8:31am
Quoting Jlr
I think the points made about the cons of martingale were pretty valid. Eventually there will be a death trade. Adjusting the bet sizing, doesn't provide an edge either as stated. It makes you a little money, at the ever increasing cost of blowing up your account.. it is pure gambling. I think martingale would only be useful if you could get the probability of the death trade to be so low, that such dangers of blowing up your account, would be less then say, dying in a car crash or something. With 7 steps, its really only similar to say, getting...
I have a plan for your last statement "unless you really got the probabilities of busting to be almost nothing"
it it has cut loss scenario but also in martingale way .

I can't say the probabilities or percentages unless I test it through 10 years data .
but theoretically it looks nice in my head.
some one is already trying to build the EA lets see if it will see the light .
I SEE YOU !!
  • #52
  • Jul 8, 2010 11:27pm Jul 8, 2010 11:27pm
Dbfusion,
heres the link to a similar 'stop and reverse heding martingale'.

http://www.forexfactory.com/showthre...t=12068&page=5

i have found range bars a better alterantive to tf bars with this system.
And i have tested an instance where this method went down to level 10.
and if u wana see worst case scenario, do test it on eur/chf.

All the best..
regards,
KS.

Quoting dbfusion
if you have seen such system before please send me link , I can't find it in the forum.

I thought it was unique Idea
  • #53
  • Edited Jul 9, 2010 12:04am Jul 8, 2010 11:47pm | Edited Jul 9, 2010 12:04am
correction- the level 10 drawdown was a result of stop and reverse hedging martingale and not ur method. The entry was made at the breach of previous swing moreover it was not hourly chart.

tc.
  • #54
  • Last Post: Jul 9, 2010 12:20am Jul 9, 2010 12:20am
Quoting King Singh
Dbfusion,
heres the link to a similar 'stop and reverse heding martingale'.

http://www.forexfactory.com/showthre...t=12068&page=5

i have found range bars a better alterantive to tf bars with this system.
And i have tested an instance where this method went down to level 10.
and if u wana see worst case scenario, do test it on eur/chf.

All the best..
regards,
KS.
I have a link on the first post of this system and it makes huge draw down on side ways .
could reach level 20 or more .
the only thing we need is a programmer who is willing to help us in this system or modify that old one to make it work the way I described it.
I SEE YOU !!
0 traders viewing now