|
不
经验
|
0
产品
|
0
演示版
|
|
0
工作
|
0
信号
|
0
订阅者
|
Retired
在
Owner
Engineer with master degree accomplishes since 1980. Mostly worked as software system engineer. Here mostly for fun as have been retired for years now. However this is my hobby. And I have been looking at several so called experts, but I don't regard them as 'experts' as they almost always do the same thing. Get a small(er) profit and take a huge risk with a large StoppLoss. As far as I can see thy all fail, sooner or later.
So I did this an own EA or I will call it a lazy man's help. Like a windrudder, it works 27/7 and I can do other things. So what is this doing that no one else is ?
This EA is not bragging about 80% or more profit rate. Win rate is more like 12 to 35% but that is huge when the tp/sl is more then 10. Like the one I bought, (of course that the way most of us start ), all have a large sl and way smaller tp. Example is 21points at win and 1500 as sl. Of course you also need an enormous high win rate to make that equation end up with profit over time. And that is how that nice curve going straight up at 30% makes us wonder how it's done. Of course it's done with a lot of fitting so tp and sl make it work and it sells.
A much better way is to make smaller losses and bigger wins. It is more sustainable and actually profitable. For example, I set up my TP and SL after 5 minutes per pair in the Optimizer covering 2020 to 2024. They all still make profit in 2025 and more important all the way since 2016 - 2025. It's easy to see when an EA is overfitted if there is variables to test. It is shown as horizontal lines in the heat map. So many now hides all the parameters, but still can't hide the tp/sl ratio.
When making this EA I learned a lot of things. (I have been trading on the stock market and futures for 20 years, but Forex is completely different. First I noticed there is not one classic Indicator that is actually working. I tested and registered the price for 50 - 100 bars after a 'signal' and did some data digging. Not one is working. On the contrary, doing the opposite will work even better maybe most of the times. You may fit the signal for some time but in the end the result isn't there.
I also pretty fast found that trailing isn't working as the size of the wriggling is so big that we loose most of the profit that could be. So trailing may look as a beneficiary feature but in reality it is not. It may help get some big wins but overall it's a loss in most cases even if we adjust the sl to break even even after 95% of the tp is achieved. There is a way to go by selling off halt the lot at some point and let the rest 'run'. However it also may be more profitable to gamble the full lot on reaching tp.
So what is there to do ? I set it up to make bigger wins and smaller losses. I may loose up to 90% of trades but still make good profit. Very small losses and big wins is a profitable concept that actually works. Why is it not common ? Because overfitting and a nice curve sells very good. However with so many trades my curve also look very nice and it works in the long run. Not on all pairs as my setup need volatility. The bigger the better. So the problem is mainly to find volatile pairs. GBP is commonly said to act erroneus, but that is good volatility and acts in my favor, as do gold and some indices and at least a few more currencies.
My EA now trades ten pairs, which is easily expanded. Everything is indexed and there is very few parameters. I will open up the tp and sl with time if you want better optimization, but it's my believe it's really no good. For the time now the only parameters is which pairs to trade and to set the lot size for each one. Indices and Energy like XBRUSD may only be traded by whole or sometimes half lots. (Which is ok with so small sl's).
All this comes with something to consider. Trading ten pairs at the same time demands a bigger balance. But it's fully possible to start with only one and grow into more pairs as balance grow. (Because it will, or at least it's a very good chance it will). It's impossible to guarantee, no one can know what will happen tomorrow. I can only say I think using this EA is my best chance to really earn money on Forex.
I give two examples. Nasdaq100 I trade with a tp of 22358.35 at the price 22106.35 with sl at 22082.35. TP is 252 points and SL is 24 points. The TP/SL ratio here is 10.5 and the win rate on this indices is 34%. that means I win on average 34 * 228 points and looses 66 * 24 points which results in a win of total 6168 points on one hundred trades. The best thing is there is never a big loss I can't handle.
Next example is on Gold ./. USD Price was 1240.03 sl at 1239.78 and tp at 1245.3 that means a risk of 25 points against a win at 500 points. Or a tp/sl ration of 20. With that setup I only need one win for every twenty losses, but the win rate has been closer to 15-16% making this very good.
So there is no use of any magic like grid or putting more gas on the fire if it goes against me, I know next win is not far away.
I will provide some nice 'profit curves' that is from the past and absolutely no guarantee for the future but a hint what it may be.
To begin I set the EA up to trade with small lots, smallest available and you may increase it with time as you get more confident this is working for you. I started with 2000$ trading just gold, one currency and one of the indices.
Oh, this EA is not cheap. It took me 2 years of hard work. (Ask muy understanding wife). So I can't let it be cheap. But if there is no win in the first year I will give that money back or send to charity whatever you choose. This is the best guarantee I can give. If I due to my age can't keep supporting this software I will provide the source. (Hopefully that will never have to take place, but who knows, I am now 73). Maybe my grandson can continue, I started this because of him and to provide for his college.
So I did this an own EA or I will call it a lazy man's help. Like a windrudder, it works 27/7 and I can do other things. So what is this doing that no one else is ?
This EA is not bragging about 80% or more profit rate. Win rate is more like 12 to 35% but that is huge when the tp/sl is more then 10. Like the one I bought, (of course that the way most of us start ), all have a large sl and way smaller tp. Example is 21points at win and 1500 as sl. Of course you also need an enormous high win rate to make that equation end up with profit over time. And that is how that nice curve going straight up at 30% makes us wonder how it's done. Of course it's done with a lot of fitting so tp and sl make it work and it sells.
A much better way is to make smaller losses and bigger wins. It is more sustainable and actually profitable. For example, I set up my TP and SL after 5 minutes per pair in the Optimizer covering 2020 to 2024. They all still make profit in 2025 and more important all the way since 2016 - 2025. It's easy to see when an EA is overfitted if there is variables to test. It is shown as horizontal lines in the heat map. So many now hides all the parameters, but still can't hide the tp/sl ratio.
When making this EA I learned a lot of things. (I have been trading on the stock market and futures for 20 years, but Forex is completely different. First I noticed there is not one classic Indicator that is actually working. I tested and registered the price for 50 - 100 bars after a 'signal' and did some data digging. Not one is working. On the contrary, doing the opposite will work even better maybe most of the times. You may fit the signal for some time but in the end the result isn't there.
I also pretty fast found that trailing isn't working as the size of the wriggling is so big that we loose most of the profit that could be. So trailing may look as a beneficiary feature but in reality it is not. It may help get some big wins but overall it's a loss in most cases even if we adjust the sl to break even even after 95% of the tp is achieved. There is a way to go by selling off halt the lot at some point and let the rest 'run'. However it also may be more profitable to gamble the full lot on reaching tp.
So what is there to do ? I set it up to make bigger wins and smaller losses. I may loose up to 90% of trades but still make good profit. Very small losses and big wins is a profitable concept that actually works. Why is it not common ? Because overfitting and a nice curve sells very good. However with so many trades my curve also look very nice and it works in the long run. Not on all pairs as my setup need volatility. The bigger the better. So the problem is mainly to find volatile pairs. GBP is commonly said to act erroneus, but that is good volatility and acts in my favor, as do gold and some indices and at least a few more currencies.
My EA now trades ten pairs, which is easily expanded. Everything is indexed and there is very few parameters. I will open up the tp and sl with time if you want better optimization, but it's my believe it's really no good. For the time now the only parameters is which pairs to trade and to set the lot size for each one. Indices and Energy like XBRUSD may only be traded by whole or sometimes half lots. (Which is ok with so small sl's).
All this comes with something to consider. Trading ten pairs at the same time demands a bigger balance. But it's fully possible to start with only one and grow into more pairs as balance grow. (Because it will, or at least it's a very good chance it will). It's impossible to guarantee, no one can know what will happen tomorrow. I can only say I think using this EA is my best chance to really earn money on Forex.
I give two examples. Nasdaq100 I trade with a tp of 22358.35 at the price 22106.35 with sl at 22082.35. TP is 252 points and SL is 24 points. The TP/SL ratio here is 10.5 and the win rate on this indices is 34%. that means I win on average 34 * 228 points and looses 66 * 24 points which results in a win of total 6168 points on one hundred trades. The best thing is there is never a big loss I can't handle.
Next example is on Gold ./. USD Price was 1240.03 sl at 1239.78 and tp at 1245.3 that means a risk of 25 points against a win at 500 points. Or a tp/sl ration of 20. With that setup I only need one win for every twenty losses, but the win rate has been closer to 15-16% making this very good.
So there is no use of any magic like grid or putting more gas on the fire if it goes against me, I know next win is not far away.
I will provide some nice 'profit curves' that is from the past and absolutely no guarantee for the future but a hint what it may be.
To begin I set the EA up to trade with small lots, smallest available and you may increase it with time as you get more confident this is working for you. I started with 2000$ trading just gold, one currency and one of the indices.
Oh, this EA is not cheap. It took me 2 years of hard work. (Ask muy understanding wife). So I can't let it be cheap. But if there is no win in the first year I will give that money back or send to charity whatever you choose. This is the best guarantee I can give. If I due to my age can't keep supporting this software I will provide the source. (Hopefully that will never have to take place, but who knows, I am now 73). Maybe my grandson can continue, I started this because of him and to provide for his college.
Rolf Nylander
From now I have only one source with one #ifdef __MT4__ tag. All the engine and logic of DragonPower is now the same code. Only one thing differ, the executable. That's a major feat and the way to have way less work keeping two versions. I don't even have to test the result with MT4 even just working under MT5. A huge relief as one of my brokers only allow MT4.
And I made one more decision, this EA will not be sold. Not even for hire. I will keep it just for myself and the small interest club I will fund. Future is going to be interesting !
And I made one more decision, this EA will not be sold. Not even for hire. I will keep it just for myself and the small interest club I will fund. Future is going to be interesting !
Rolf Nylander
Now also running in single mode on MT4. It has been an interesting move. A new wrapper trade class executing all the MT5 code also under MT4, works wonder and now I have only one code to maintain, and there is more or less only one #define tag. That's the MT4 for the wrapper class. This makes the code way more easy to handle and the risk for severe bugs much smaller. (Good for a lazy coding guy). As before I will show the report on the backchecking since 2020. It shows a net profit of 12.800 without funny things like multiply volume when balance rise. No all trades with the same volume, but there is a switch to satisfy your greed ;-) Then this profit goes up to 100.000 from a 1.000 account. Still there is only 16% win rate but that's not bad with tp/sl of more than 20. That means on average this ea only need one win for 20 losses to still show profit. There is advantages with this approach. You never take a huge risk for a small profit and your balance will never suddenly go blasted to zero. You will always have plenty of time to stop the proceedings with minor loss. I started this on a live MT4 account yesterday . So far two losses of total 2.02$. That is just the fee and the cost to maybe get about 100$. With an average of 6 losses for one win, it's a very good equation. As the report show, even with 72 straight losses the total amount was only 196$, However on the other side there was 9 win trades in a row getting 433$, netting a total of 239$ on the balance. Gross profit was 19.794$ and Gross Loss 6.961$ Netting 12.833$. With more money on the account we may set the stake higher as the losses is limited and we may always change back at any moment. This is why I call this an AntiScalper. Scalper always have a very small TP/SL ratio and need more than 95% win rate consistently to make profit. However in the long run the market usually upset them all. This anti scalper turn the table around and you will never risk more than you can afford. (Not many can boost this on Forex ) !
Rolf Nylander
This EA will not be for sale. It will be free to download upon agreeing to a license that will charge 22.5% of net profits. No profit no pay. It will work as long as license payment is done at end of each month. EA itself will keep track of your trades and the individual result. Automatically sending a paypal invoice. High Net traders or Funds will have a lower rate depending on the size of trades. The fee will be in the bracket 10-15%. I will close this offer eventually and only admit a limit user base. The backend is now in work and when officially released there will be an introductory percentage for the 10 first agreeing a license valid for one year. The EA is not for very small accounts. It may trade 0.01 lots if you set it up that way, but Indices, Energy and Gold trading at the same time may take your margin down too much and trades will not open. My advise is to start with at least 3000 USD if trading more than 3 pairs. (I am extending the pairs available for trade and may add new strategys as well. At the moment there are three working together. None depending on anything but price movement. (No indicators). It also have dynamic stops and spacing depending on market conditions and price levels. So the settings is very simple, trade or not and the size of your lots. There is also a choice if you want to use the bayesian probability help that may add 30% profit. This will change the volume settings to internal decision affordable according to your balance. I hope to finish this now over the summer with the backend. Until then !
Rolf Nylander
After added a bayesian probability on the trading profit went up about 30%. It's really interesting and a new way of looking at trading. AI is very in the loop, but how should AI predict a trade if there is really nothing to make a forcast on where it's going. It's more like a horserace with some low odds that wins sometimes, but any bet may win. So making a probability influence the trading and see profit up 30% was really an eyeopener. This is exacly the same strategy, just influenced by probability of the hour on the day if the trade will be a win or loss. Not Bad and next step is to make it adaptive and self learning. That work is on the way for next version. I am also working on more pairs and maybe stocks and crypto.
Rolf Nylander
Note curve is from trading with 0.01 lot on Gold. Of course the is a parameter to adjust the lot or just change the setting. Then the curve will go exponential. But I don't think that's a fair way to show the performance.