Joe. Swing-trade your way to a baby equity millipede

[SteveHopwood]

Worth considering, for sure. Maybe there should be a minimum pips in between new trades?

In the meantime, is anybody actively coding the buy_limit/Sell_limit thingy? Cos if not, I have just had a bath and will code it later.

[andydoc]

No I'm not but have done the signal verification as without it there is mostly no move left to stack!
I am also testing leaving out the initial trade as too many get faked out and what we are left with then is a straightforward MA cross confirmed by a preceding FB

Here it is with the BE only after MA Cross in our favour (lines 1190-1191,1231, 1314)
Also the shift on the looking for Trend weakening reduced at lines 3188, 3190

[DragosDanescu]

It may be a case of filtering the FBs but f so will clearly need to to run it on a number of different charts for any reasonable trade frequency

There's still some other work to do on the EA as well. We need to look at the move to break even and perhaps only do it when we are a certain number of pips away. Also, we don't strictly follow the rules. We don't do a limit order to get in on the retrace when price is below the 50% point of the previous bar (for sells), so we are exposing ourselves to larger SLs and less profit on those trades. Not sure if it is material, of course, but we'll need to code it to see.

Guys,
BE option with BE pips is already there, and it has no logic to use it elsewhere except for the stacking orders. Why not just simplify the code and send a stacking trade only when last opened position is at BE? No more OrderModify, we can get rid of a lot of problems...

[andydoc]

Hi Dragos
Problem at the moment is that we need to filter out False entries - there are a huge number IMHO many more than 60%. How we handle this is something I am looking at at the moment. The big thing about this strategy as far as I can see is not the entry, but the scaling algorithm, and we may need to be flexible about the entry signal

[andydoc]

Having said that, it is very important to alter this bit in the breakeven test after both ordermodify lines:

Code: Select all

if (!result) 

to

Code: Select all

if (!result && GetLastError()!=1) 

Otherwise you get to a grinding halt - I have not worked out where the unchanged parameters are coming from - this is easier

[garyfritz]

Problem at the moment is that we need to filter out False entries - there are a huge number IMHO many more than 60%.

Agreed, Andy. I'm downloading Dukascopy data to get reliable ST results -- my initial poor-quality ST run produced 100% losses -- but eyeballing the chart it's clear there are huge numbers of false FBs. I don't quite see how this thing can work given the number of false starts you'd take, only to get stopped out. Especially if that cut short your run on the other side.

E.g. in H4 EU there was a cluster of 3 down FBs on 14 Nov, and that kicked off a 500-pip move over the next few weeks. But there were 14 up FBs during that down move, at least 3-4 of which would have triggered a long entry if I understand the rules right. How do you stay in the move and stack up your millipede if the FBs keep telling you to reverse?

[SteveHopwood]

Latest update in post 1.

I have added Andy's fix from a couple of posts ago.

Add the ability to detect the need for a limit order in the stacking process. Limit order pending lines are LimeGreen/OrangeRed and HasBuyFilled() and its partner have some extra code to use the info in the colour field.

[gaheitman]

Problem at the moment is that we need to filter out False entries - there are a huge number IMHO many more than 60%.

Agreed, Andy. I'm downloading Dukascopy data to get reliable ST results -- my initial poor-quality ST run produced 100% losses -- but eyeballing the chart it's clear there are huge numbers of false FBs. I don't quite see how this thing can work given the number of false starts you'd take, only to get stopped out. Especially if that cut short your run on the other side.

E.g. in H4 EU there was a cluster of 3 down FBs on 14 Nov, and that kicked off a 500-pip move over the next few weeks. But there were 14 up FBs during that down move, at least 3-4 of which would have triggered a long entry if I understand the rules right. How do you stay in the move and stack up your millipede if the FBs keep telling you to reverse?

Didn't the original "Joe" just trade sells? I was thinking the same thing, the stronger the move in one direction, the more FBs there are in the opposite. I meant to test it with just sells (or just buys), but haven't yet.

As Gary mentioned previously, if the trick is really the stacking, perhaps we alter the bot to allow different direction signals and the fire off the stack based on them. Maybe there is a way to make an MA cross successful. :>

George

[andydoc]

Hi George

That may be it - but look at this - 2 way, not taking the original trade only the stack after the MA Cross, and Pyramiding the position
Please note this is too far from the original concept to really qualify but maybe there are some principles we can look at.
I would want to add Steve's limit order amendment which I haven't yet looked at

[garyfritz]

Maybe there is a way to make an MA cross successful. :>

I wouldn't go THAT far. But maybe we could use, say, Hanover Recent Strength as our directional signal, and stack as long as the RS goes our way? Add to the stack when the low TF pulls back, as long as the higher TF stays steady?

I think stacking is the magic behind Joe's approach. He just combines a very weak turn indicator (FBs) with his discretionary skills to find good opportunities. We can't use his discretionary skills so we need a better turn indicator. Hanover might do that.