Basket 14 EA System

This thing is a money machine so far. Up 5(+)% and not thru the 2nd day yet!!

MC, I highlighted the biggest winners.

Do it taz!

My current stream of consciousness about the forex market...

When we view currency pairs as numerator/denominator entities where a single currency is either a numerator or a denominator (that's all they are, correct?), the two 7-pair sides of our 14-pair hedge are mathematically equivalent. So if we multiply the exchange rates of both sides at any point in time, we should get the same numerical result.

I've run EAs that do these multiplications and comparisons in real time, monitoring for distortions. Remarkably, the two sides of the hedge stay equivalent within a range that doesn't allow for profitable exploitation (we have to overcome spread, remember), save for very occasional distortions that are quickly corrected.

However, if you map the movement of the two sides of the hedge, they DO drift apart over time, then drift back together. There is no telling which side of the hedge will move more than the other, we can only observe it when it does happen. We observe it when we open trades where one side is bought and the other is sold, and our account sees a floating loss or a quick profit, neither of which should happen. But it happens because of the sides moving to different extents.

WAIT A MINUTE, you say to yourself. What is going on?

If at single-timepoint snapshots the two sides of the hedge stay roughly equivalent almost always, yet they move apart and back together over time, wouldn't that mean at least one currency pair among the 14 pairs might appear to be used as a "pressure relief valve" or something like that?

Let's step back for a moment and try looking at it from the banks' and the brökers' points of view. They know that smart traders out there are looking at these synthetic hedges and are trying to profit from momentary distorations that should not exist. They can't give away free money just for being smart enough to do a little multiplying and comparing!

Some mechanism has to be in place to keep the distortions under control. One thing that can be done is to play with spreads. If a distortion gets large enough that a sophisticated trader could profit from it, the spreads can be increased to defeat the trader's advantage.

Another thing that can be done to keep the distortions in check is to continually adjust exchange rates of selected pairs, a little at a time, to keep things in equilibrium despite the movement that's happening.

My belief at this time is that EURCHF, EURGBP and GBPCHF are among the pairs that are adjusted to keep the distortions in check as the individual pairs move due to trading action. This is why they pop up as biggest phantom winners as often as they do.

Thoughts?

Open every half hour if trades are the biggest phantom winners or the biggest phantom losers (traditional direction reversed if losers). Require that all preexisting trades on the pair be in profit.
|
v

Hi there,

I came across your postings and the code that you posted on the 14 currency pairs. I found your code to be very professional and your ideas are very close to the work that I am doing although from a different angle..We have difficulty in implementing and testing our project. what is the best way to get in touch with you...

Thanks

Here is a fresh picture and it shows the both sides of the basket in pips. The time is from 0 min to 5 weeks. It shows the buy side has been stronger than the sell side but its not many pips different. They are nearly mathematically equivalent

Sir, my energies are focused on this thread. Thanks, Bob.

My 02-taz demo has already closed again since I posted the spreadsheet.

I will have this one going tonight.

As I understand it, this is the same as 02-taz, but adds the requirement that all previous trades on the pair be winners. Is that correct? Or does it also limit which pairs are traded to 3 or 6?

LOL, after I spend a good 30 minutes figuring out how to do this myself, save it and put it into place, I come back and see you've done it for me!! Darn you!!

Thanks MC!

Taz's code already had the requirement all previous trades on the pair be winners. The biggest phantom winner/loser criterion replaces the +/-20 pip criterion.

It trades all 14 pairs.

Don't know if it was exactly what you were thinking of doing.

yeah it was, and it looks like i actually did do it right, you just added additional code to show the biggest phantom loser/winner to the onscreen display, which is a nice touch

On another note:
What do you (and anyone else that understands the mod) think of this possibility?

If a series of trades is open based on my mod, and a >20pip phantom manages to move all the way to the other side and go -20pips, doing a trim and reverse scenario? Granted, we'd probably want to go back to every hour instead of every half hour to prevent the whiplash of a bouncing pair to get out of control on us, but it might help in some of the long progressions I (or we) have been seeing sometimes?

Thoughts?

I would like to try a lot of things before I go back to trim'n'reverse.

First post, I know... I have been following the thread for a while and am a fan of the ideas here and MetaCoder's coding. Thought I would chime in with my update to the code in case it helps future iterations:

1. Added an extern variable for add interval. This is in minutes and a MathMod is done to determine if we should check for adding positions:

...
if (MathMod(TimeMinute(TimeCurrent()), AddMinuteInterval) == 0) ClearSymbolMarks();

if ((BidsHaveChanged &&
(MathMod(TimeMinute(TimeCurrent()), AddMinuteInterval) == 0) &&
!CloseAll))
...

- Example Values: 60 = every hour; 30 = every 30 minutes; 15 = every 15 minutes, etc. (don't set it to 0...)

2. Added an extern variable called OnlyAddAllInProfit. Possible values TRUE or FALSE. If true, then all trades have to be in profit for additions. If false, don't add on:

...
if (SymbolArray[i] == BiggestPhantomWinnerSymbol &&
(!OnlyAddAllInProfit || (OnlyAddAllInProfit && AllSymbolTradesInProfit(SymbolArray[i]))) &&
!SymbolMarked(SymbolArray[i]))
...

Hope I'm not stepping on anyone's feet and I hate to add another version to the mix, but it seemed that some people like the every hour addition, some people like the all-in-profit option, and some don't.

Sounds good, thanks for the input

That's remarkable, but I'm wondering what can thwart this strategy.

What about a widespread reversal where we have too many pairs committed to one direction to effectively go with the new direction, assuming we leave same-pair hedging off the table (I personally won't explore anything that employs same-pair hedging)?

I suppose this is why taz is thinking of using trim'n'reverse.

Version 01: Every hour, add a trade to any phantom pair that is a biggest phantom winner during 59 minutes of the hour, not just minute 0. Require that all preexisting trades on the pair be in profit.

Version 01a: Same as 01, but also reset the phantom trades during minute 0 of every hour. This should let us detect direction changes much sooner.

Letting addsix 03 taz run a little longer, but doesnt seem like it's working here. Unless I am mistaken, there isn't a set amount a phantom needs to be in profit right? just needs to be the biggest winner or biggest loser?
If that is correct, it's gone through 2 cycles already without opening any trades. Will let it go through one more (in 17 mins) before trying to figure out whats wrong.

Since the actions are only taken on minute 0, it's very possible that none of the six pairs are popping up as biggest phantom winner.

k, will keep watching it, but am pretty sure there's a biggest loser/winner at every hour half hour so far, just no action taken. I've increased it to every 10 minutes now to try to troubleshoot the issue