SWB Martingale EA

Okay, I'm looking at this and see one problem: Once the lockin point is set and price swings back to close hedge at lockin point, what's to keep it from opening new hedge when it drops to point where first hedge opened? (1 point lower?) If that happens, and it goes on to retrace with new hedge open, it won't do any good.

I suppose you could specify a reset point after so many pips but, if you don't reach it and the price reverses, you're still going to find yourself vulnerable.

You might have heard of the Blessing EA, which is probably the best free grid trading EA available, hedging is included as an option in this but nobody has found settings that work, and there are quite a lot of people working on it. Hedging sounds good but difficult to make work. And locking in 1 point and stop loss at 5 points as suggested by the guy above, which broker is going to allow you to do that. Anyway take a look at Blessing there might be some code you can use for your EA.

http://www.jtatoday.info/forum/

Hi ecTrade,

I see your point re reset of the hedge if it closes at a 1 pip hidden take profit point.

Maybe consider case where we include the spread of 2 pips then the price will be 3 pips above the previous buy point. Might be worth looking at not having the next attempt at a hedge start a trade until the price has moved say +10 pips above the first trigger point.

So if we are stepping at 25 per trade and hedging after three levels, hedge kicks in at 100 pips of a loss and 2x the last lot size, hedge locks in and closes out for 1 pip profit, market reverses and we want another hedge, so next attempt at hedging will be at the 110 pips loss. Problem of course is if the market is just bouncing around and opens and closes a series of these trades then the hedge trade could just get too far away from the first three levels to be useful?

Regards,

Pipernet.

Or, maybe have it open a new hedge at 10 pips above, like you say, with a reset point 10 pips below, so it opens new hedge at original level? That might be do-able.

Yes.

Okay, this is a revised version of the original SWB. I restructured ordering process to make it easier to modify. It should maintain same functionality, however. Although you folks might want to check it out to see if it works to your satisfaction.

It also includes previous version of hedge, plus h_top_level, which allows you to select level to shut hedge off. It may actually come in handy with this other idea we're talking about.

Ectrade-
The idea of stepping the entry of the hedge I think is a good idea. Is it doing that now?

From what I can see during the backtest, although it seems to be working fine, it is not what I had in mind. I was thinking that when the hedge trade pulled enough pips to break even, it closed all of the trades at the same time.

The way it is currently running creates a ton of draw down and it seems to be working just like the other hedge version works.

You should only need 1 hedge trade, and that is the last one. Here you can see 5 hedge trades were fired while keeping the original BUY trades open.

(I have it set to only 7 levels deep, so that is why it didn't fire anymore)

I do think your way will make more actually money, but I am getting some huge draw down spikes. If I am running this on 10 different charts I think the less draw down will be very important.

Let me know what you think.

See screenshot.


http://i368.photobucket.com/albums/o...ge_example.jpg

Pudsey-
Yes - I have heard of Blessed, and I agree it has some serious math involved. It is extremely impressive. However, in my tests it still has the dreaded death trade on some currencies.

Here are some backtests using the Blessed EA:

The USDJPY looks good with default settings:

http://i1084.photobucket.com/albums/...djpy-2year.jpg

EURUSD, Not so much with default settings:

http://i1084.photobucket.com/albums/...rusd-2year.jpg

EURUSD, using the same settings found on the website:

http://i1084.photobucket.com/albums/...SETTINGS-1.jpg

This SWB grid I believe has the potential to run great on all currencies if we can implement this hedge feature. (at least on paper it should work)

Domino, I hope you are successful with your hedging work. I was merely pointing out that hedging is available in Blessing but not been succesful upto now, though the guys are still working on it. And yes it does have some serious code in it. I don't think any EA will overcome the death trade including Blessing, but its how you manage it, and Blessing has options to prevent total blow out and protect some of your account.

Anyway this thread is not about blessing and I wish you well with your work on your EA. And if your hedging works maybe the Blessing group can also learn something.

No, as I said in previous post, I just restructured ordering process so it was easier to modify. And, since I intend to use it to implement this new hedge idea (am still working on it), I thought folks might want to check it out to make sure it works correctly. And yeah, it has basically the same functionality as the original SWB, with the hedge.

Hey, I tried your idea with the initial 5 point SL, but I kept hitting it so many times (one right after the other) while backtesting that all it did was erode my balance, in no time. So I tried testing it without it and set SL to break even point (OrderOpenPrice()) after 5 point positive swing with much better results. Of course if break even point doesn't get set, and price swings the wrong way, you start losing money as a result of the hedge. In which case I suggest making the hedge lots equal to the total non-hedge lots. So, now all you have to do is try to find a way to get out of the hedge. Of course this isn't always easy either but, if the majority of trades closeout at level 3 (with hedge set at level 4), you are still making a profit and closing out the hedge (once set) may not need to be quite so exact?

I also think you should open the hedge at same point as last level (without skipping a level), because the longer you wait to hedge, the bigger the drawdown gets. One thing that might help (which I am trying) is to trigger hedge using rsi filter, to ensure it has momentum to get past break even point. Or, maybe there are better filters here?

Another thing I think would only help is to add a trailing stop to hedge, to help insure you get some profit out of it. Then, after hedge closes out, via break even, trailing stop or, TP, it should be set and ready to open a new hedge, via rsi settings. That, I think, should pretty much cover all the bases. Except of course if break even point doesn't get set and price swings the wrong way. But, maybe we could try something similar (with rsi filter) here?

EC-
Have you tried opening the gap up a little bit?
Meaning....when the hedge fires (say level 4), it put a 15 pip stop loss (instead of 5) in case of emergencies. Then if it goes the right direction, maybe you could trail it up to break even. Or once it hits a 5/10 positive swing, it moves it to break even. But, that way you aren't left open.

I haven't seen any indicators that seem to call it right. The RVI seems ok, but it has flaws. I was checking to see if there was a cross after level 3, it could be the signal to enter, but it doesn't look that good. Maybe the RSI if it is above/below 80/20 after the 3rd level or something. That seems like it actually might work. It would probably work as an exit also. If the RSI crosses below/above 70/30 then exit hedge trade.

Do you think you can make the hedge trade a multiplier option?
That way we could backtest it and see if keeping the same or bumping it up would work better. Maybe something like:

Hedge_multiplier =1 (1 means it will fire the total of non-hedge lots as default. Then we could test 1.5x 2x, 2.5x etc)

As for the trailing stop, I guess we can try it. I haven't had very good luck with them though. I usually get closed out on a retrace, then the price continues and I am not in! Would the hedge trade fire again at that point?

My original thought was so it would never go more than 4-5 levels deep (instead of 7-10). The hedge trade wasn't really designed to take profit. It is designed to protect the account from the "death trade". I thought it would just go enough pips to make your "tp_in_money" over the other level losses and close all of the other previous level trades. The bigger the hedge trade lot size, the fewer pips you have to grab to make your "tp_in_money" and you move on. You would be out of the trade and you don't have to worry about the potential level 10 scenario. That way I would feel comfortable using much bigger lot sizes from the start if I knew I couldn't get closed out due to lack of margin. If the hedge trade takes a small loss because it got stopped out, maybe we take a little more pips to make up for it when the other 3 levels close out? If somehow you could tie that math together or something.....that would be nice! Obviously, the bigger the hedge trade lot size loss is, there is no way it can be made up for. But if you lost 15 pips on a .16 lot (level 5 on micro) = ~$24, then $24 + "tp_in_money" would be when the other 3 levels will close, instead of standard "tp_in_money". That might have to be an option so we can test it to see if it works. (Makeup_hedge_loss = True/False)
If it's too much to make up on the remaining 3 levels, maybe the next trade makes it up or something.

We can try the trailing stop as an option, that way we can test what works best.

I am open for ideas....

Thanks a bunch!

Okay, this is just preliminary, I just thought folks might want to see what I've been up to the last day or two, and give them a chance to tweak around a bit. Boy, this was a pain in the butt!

Instead of trailing stops, TP's, SL's and break-even points, I've decided to open and close hedge using RSI filter. So far so good! ... albeit it's far from perfect. It's a step in the right direction though.

Currently hedge is set to level 3, start lot set to .25, tp_in_money set to $15.00 and hedge factor (h_factor) set to 35% of total non-hedge lots. Am also testing with 15M TF and $2,500 initial deposit.

The opening RSI settings are 65 high and 35 low on 5M TF (internally), and closing settings are 40 low and 60 high. Currently these are not adjustable, but I will try and incorporate that within the next day or so. Am also using multi-TF RSI indicator which you will need to download below ... and save to indicator folder.

domino, I will try and address some of your points later. Right now I need to get some sleep!

Thanks.

ETA: Added TP mod per post below.

Okay, I just added TP settings to EA. It's set to 120 pips default, although it's variable (between 30 pips and 120 pips), depending on RSI strength. I think it's kind of neat!

If you downloaded EA earlier, you need to download it again if it was before edit stamp on post above.

Oh, I set hedge factor (h_factor) to 70% this time. You can change it back if you like. I also changed RSI entry settings to 60 high and 40 low, as it was opening too many trades right towards end of spike.

Thanks.

I'm new to this forum and I've been trying to get a good trending Ea but with limited success.

Has anyone tried a Trending martingale? I have been manually trading a indicator with re-entrys an actual case study, I'm still in it at 12:00 noon EST-GBPUSD 1m chart .
1. Enter on Trend 10:53 Sht- 1.6186 - 0.01
2. enter again 11:18 1.6165 - 0.01
3. enter again 11:13 1.6165 - 0.01
4. Out 11:42 1.6118 +162 pips

5. Enter Trend Rev 11:42 LG- 1.6118 - 0.01
6. enter again 11:46 1.6165 - 0.01
7. enter again 11:47 1.6165 - 0.01
8. OUT 12:02 1.6165 +7

I hope this explains the thought. I'm not a programmer, does anyone have the ability to help program this EA.

Thank You
Ray

Traderduke-
When you trade with the trend, it is called Anti-Martingale. This is when you add lots to your winning position instead of doubling down on a losing position, known as the martingale.

However, you are not actually doing a martingale unless you double each lot size as you enter. Trading an anti-martingale can become risky because you can get in with your biggest lot size at the worst price and take the biggest loss, wiping out all of your other gains.

I would check other threads regarding anti-martingale's, they might have something posted already.

This thread is specifically about SWB Grid.

EC-
How is the hedge trade exiting the trade?
Is it just variable TP or the closing of all of the trades? One or the other?

As you see here, I think maybe if it had a RSI exit in addition to the TP exit, it could be much more effective.

The 3 lines on the RSI are 60, 50 & 40. I think if the entry was above 80 (it was about 83 in this picture at the point of entry) and the exit was a 60 cross, it would be pretty effective. It seems to exit with all of the other trades along with this hedge. The problem with this trade is all of the big winnings from the last sell trade are completely wiped out by the hedge trade closing at the same time. If the hedge trade would have closed sooner, it would have been a winner. I think if the hedge trades hits the TP line it should close all of the other trades with it. If it hits the RSI exit criteria, then only the hedge closes and the others remain open to close like normal.

http://i1084.photobucket.com/albums/...e-example1.jpg

The other huge issue is this one. I am sure you are working on it, but I am just bringing it up.

The account ended up at zero because it kept firing hedge trades for 2 months and never closing the original buy trades. The hedge trade wins will have to be tied to the closing of the original trades after the "tp_in_money" it reached.

As you can see here, the hedges made WAY more than the other trades open. If fact it more than doubled the account, but still ended up at zero because it was about a 1200 pip move over 2 two months.

http://i1084.photobucket.com/albums/...ro-balance.jpg

P.S. - I ran it with 7 levels and started the hedging at level 7 also. The default 4 was not enough. It would open trades and just continually fire hedge trades, similar to what ended the account in the above picture.

Looking good.....almost there.


Thanks,
Domino

Primarily I'm focusing on closing out the hedge, in an attempt to "pad" the profit, although it will fluctuate relative to the position of the hedge and where it closes out ... sometimes at a loss, because I've set a -10 point threshold (invisible SL), which is set after price rises above hedge open price and hits RSI cross on the way back down. Or else you can be in for a long wait if there's a 200-300 gap between the hedge open price and last non-hedge open price. This is basically the problem we have now. So, we have to try and find a way to minimize the gap or, at least close out the hedge when it is retracing.

Also, so long as tp_in_money target is not reached, it will only close out the hedge trades. But, once tp_in_money is reached, it closes out all the trades and starts over again.

One thing you can do you here, is allocate a certain percentage of hedge profit to be set aside, and not applied to tp_in_money, say, 25%? so, once everything does close out, you will realize that above and beyond the tp_in_money ... which could be a sizable chunk of change! So, you were not just sitting around waiting for nothing all this time.

Oh, hedge is currently existing with both TP and RSI settings. If you set TP to 0, it will exit with RSI settings only. What I need to do is make user-definable inputs for RSI entry and exists, which I'm in the process of doing. That way, maybe you can try some of the things you're talking about above.

Another thing to consider, which might be the real cincher here, is set it up to hedge 1 to 1, and treat both sides with same dynamics as the one side we are treating currently. That way you can close out whichever side peaks, and you don't have to wait forever for the price to swing back up and down. Heck, if we could do that, we might not even need the Martingale!

We'll see.

I think there may be bug here.

Here you can see where it closes 7 levels deep on the original version.

http://i1084.photobucket.com/albums/...EVEL-CLOSE.jpg

Here is the hedge version starting 7 levels deep and it doesn't close the original trades.

http://i1084.photobucket.com/albums/...EL-NOCLOSE.jpg


Any idea why this would happen?

Try seeing what happens if you turn the hedge off. If hedge is set to open at level 7, and it does, it probably won't close anything out until after hedge closes and it retraces. If that's not the problem, although I think it is, it could be because spread changed between backtests. That could potentially effect it as well.

I will turn hedging off and let you know. It can't be the spread because there are at least 30 pips in there it could have closed on.

I am running another backtest now so give me a couple hours.

If that IS the case, then I think that would need to be changed. If the standard set of 7 trades are ready to close, there is no need for the hedge trade anymore.
The hedge trade should close with them.

UPDATE: Yes - if I turn off the hedging then it functions like the original. So, that means it needs to close with the rest of the trades as mentioned above.