Hedge and Correlation Strategy

Very well written post with thought clearly presented. As Caillou said though, I need to read it again for it to sink in but it clearly a result of a sharp analytical mind.

Possibly... i'm not familiar with the TMA indicator, and from what I can tell, it's as you said.... the math itself is just basic calculus, but converting it to code may come in handy from that TMA indicator...

still would need a lot of customization tho... a lot. like, factoring in a pair set for starters, as opposed to a single pair

Eh...i don't even know what i'm talking about tho...i'm no coder! i hope it can help...seems like a fair amount of calculation in my head

I agree with spinfoil, stochastic is the achilles's heel of this system. Basically you are trying to use a lagging indicator to predict future market behavior, it doesn't work. Try a leading indicator instead of a lagging indicator and you're certain to get better results.

Using currency correlation to trade on such a small time frame is way too risky, the correlation can change drastically in just a matter of seconds. And why correlate two pairs? Why not just correlate one pair? You cut your analysis in half. Check out this X Factor indicator, its splits the price of one pair, showing the weight of each currency, even at the smallest time frame. This indicator, a good leading indicator like ADX and multi-time frame analysis will bring you far more rewards for much less work. Save the stochastic for H4 or higher, once it pops 80 or 20 on an H4 chart then you know a trend is imminent...

Please direct me to where this indicator can be downloaded, Thanks.

Thank you Zr... I have been on this board for over 2 years now...but I almost never post. I rarely feel compelled...though this may be changing for me...

reading this type of thread reminded me a bit of where I was about 2 years ago, having just got seriously interested into forex...

it's the type where there is actually a profitable idea, but it seems like it might get muddy and ruined by tweeking this, or adding that, or modifying this, or altering that...

I know, because I would change an 18 MA to a 19EMA to a 17HGSUPER-MA, always hoping to find "just the right setting", or "just the right time frame" that would give me high probability trades. stuff that wins over 70%, and the winners are bigger than the losers. What I didn't realize at that time was that most of my trades were winners, AND losers. The problem was, I didn't have a method of ACCURATELY measuring risk and reward, so I never knew when a trade was just "working itself out"...or whether it was just gonna lose.

I wasted a lot of time that way...but I also ruined a "very" few good ideas.

I think this is a good idea... but lets face it:

IMO, most people will not manage a trade well if they lack conviction in the trade. I believe most people who have tried trading instinctively know that no big bank/hedge fund/pro-trader has any idea about what indicator settings they have on their screen, or whether it is "oversold".

When you know you don't move the market, and you know your information is not an accurate representation of the market forces at work...how can one possibly have the conviction necessary to really develop, have, and follow a method or system? I couldn't...maybe others can, but not me.

So, I set about to figure out what information exists out there that would show how the market REALLY works, not how we want it to work.

Pair trading is a time tested, profitable approach to the market, but you have to have a way to objectively measure the opportunity before you. AND, you have to manage risk.

if a large price event occurs, say, a 70 pip separation of two correlated currencies, and you recognize that when such an event occurred in the past, on average, price reversed over 75% of the time once it moved that much... and did so within 3-6 hours, by an average of 29 pips, you know you now have a very good chance to make a profit. However... VERY FEW trades are profitable forever....while most trades are profitable for at least a few moments.

SO...the trick is to have a way to know when a move is likely...and then have a way to make it a risk free trade. One that can't lose, but can win a lot. If you just do that...you'll break even a lot, and you will win some. Thats better than most people in the trading world ever will be. And in the example above, we know it won't take 4 days to do it. Opportunities exist in time alone.... and so do profits.

Measure an opportunity for what it ACTUALLY is, and if acted upon, know that your profits WILL eventually turn to a loss, but you DON"T KNOW WHEN (they don't ring a bell when the market hits a top) ...so what is your plan to protect your account when it does turn around?

Bottom line: this will make you money, if you can seperate your A+ opportunities from your B+ opportunities, and if you can find a way to just not lose... you will make money.

As a suggestion, to anyone who is a losing trader using this method...try this. Once you are in a trade.... once you have a single candle on each pair that does not touch your entry price, move a stop to break even.

Say, you enter EUR/USD long at 1.3810, and EUR/CAD short at 1.3623, and your trading off a 15 min candlestick chart... once you have a 15 minute candle finish printing that is COMPLETELY above 1.3810 on the eur/usd (this means neither wick nor body touched 1.3810) AND you have a 15 minute candle that finishes printing COMPLETELY below the 1.3623 level on the EUR/CAD. Move a stop to break even on each pair at this time. Leave your target(s) alone...just take profits, or break even.

Try it for a week. See how you do. It might surprise you.

Greg

P.S. If this is off topic, please let me know, and feel free to remove it. I don't want to hijack this thread, just provide some deeper insights into how to turn this system into something profitable for those who arn't right now.

P.P.S. By the way... many of those "black box" high frequency trading firms who pull billions out of the markets every year use statistical extremes such as if price moves up X distance in Y time then short it for Z pips/percent/tics/etc... and some of them make money every day

what I learned from my positive correlation trading is never fully trust the gap that we see in chart, if not we will fall in a trap that keep open another layer and layer and waiting it retrace, but the problem is, sometimes we cannot recover our negative floating, and maybe after 50 bars, u found there is nice correlated chart in your MT4, that time will panic how to recover our loss.

Please post indicator - thank you

sticky statement.
Ph.

sticky statement.
Ph.

ahhhh sorry for double post !

Hey Spinfoil, sorry to ask this but are you somewhat related to a person named "Edgetrader" or Mr. Keith Brown ? I can't help myself but your way to write things reminds me of him..

Fortheex,

awaiting your examples. If I'm understanding correctly, you were open sell/buy trades on same pair, so this kind of trading is only possible with broker accounts, which are accepting hedging (non us).

josch

The blue line shows Eur strength and the red line shows dollar strength/ weakness ? Do you know how it is calculated ?

I suppose you use it as a leading indicator, when blue above red then buy eur/usd and when red line above blue line sell eur/usd ...... ?

Look closer, the column on the right shows the various currencies, on this chart the indicator shows the weight of the USD as 4.96 and the AUD as 4.92-virtually equal, this means that price is ranging. You're right the red line represents USD and the blue line EUR (on this chart AUD), but you don't necessarily wait for them to cross in order to enter. You have to wait for the weight of one currency to be greater than the other. Check out the middle of the chart where price starts to drop. When this price drop started the weight of the AUD was less than 4 and the the weight of the USD was about 4.5-this kind of difference is what signals an entry or the beginning of a trend. One currency driving the price of the other down. Sometimes its not necessarily that one currency is stronger, its that the other is suddenly weaker. Fundamental announcements like USD employment report and others can have a rapid effect.

As far as hedging, I use an offshore ECN that allows me to take a long and short on the same pair, but I never hedge anyway. As far as brokers go, they are all buckets, I learned that lesson the hard way about seven years ago...

Interesting indicator.
Do you know how it calculates the 2 values ?

CCFP ? http://codebase.mql4.com/5862

Can you share the indicator ?

( What would also be worthwile getting programmed is a pip difference indicator between 2 currency pairs which are positively correlated. Ex: eur/usd and gbp/usd. Both are in an uptrend, however one of the 2 outperforms the other. Similar to stochastics but only in regards to pip values. Of course one would have to get a median price for the two as gbp/usd usually has larger moves. Then one could sell the currency pair "overperforming" and buy the one which is lagging behind.)

Nah....no relation Tex. I've only known two keiths that I can recall, and neither were traders.

I'm sorry to hear that tho... I figured I was the only one who would write a book with each post, and now I know there are two of us who plague the world

Have you tried this correlated markets strategy Tex?

Yes i tried it and i am still trying it. Your suggestions to solve the issues with the stoch sound good and i hope our ingenious coders here will implement some of them in their ea's and indicators.

cheers

dear squalou

what do you think ?
is better don't trade during the news ?

Squalou :

Looking for your updates on EA. Thanks.

I'll chime in on this one, if you don't mind wallstreet...

IMHO: trading this strategy during any major news on any pair you are in is about the worst possible thing you could do.

here's why: Pairs that are highly correlated for a reason. Although the specifics are always going to differ from one country to another... they all add up to one thing -

correlated pairs = very similar economic circumstance.

Most often this is due to geography (think CAD and USD, or EUR and GBP)
but there are other reasons as well, like similar primary industries, similar consumer habits, similar political structures...etc.

So, the ENTIRE REASON correlated market trading works, is when otherwise closely related currencies become overvalued or undervalued relative to their "current perceived value".

So... when two countries have currencies with a high level of correlation, and one of those currencies rises a great deal in price compared to the other...but there has been no economic reason for this to happen... it doesn't "make sense", and the market will correct it...by buying the lagging currency, selling the leading one...or both.

BUT, major economic news is the one thing that can substantiate a large move in otherwise correlated currencies...and that may permanently alter the degree of correlation...

News of changing economic conditions is the one thing that actually destroys this entire systems reason for existing... because changing economic conditions can affect one currency, while leaving the other untouched...therefore permanently altering the degree of correlation between the two.

Give you an extreme example. USD and CAD are very correlated when compared to EUR.

But...say canada raises interest rates, and the U.S. doesn't. Say...canada raises interest rates a full 2% in 6 months, and the U.S. does not raise interest rates at all.

Now, if you were in a trade from 7 months ago that went short EUR/CAD, but long EUR/USD.... what are the odds that your trade is showing a good profit?

Furthermore, what are the odds that the two currency pairs no longer have such a high correlation (it usually is at or above 80%)

I'd figure a good 10% - 20% drop in correlation is a minimum...AND, it can last a VERY long time.... until the U.S. raises interest rates at least...and then, there may be such a large difference, it can take months or years to correct... and all that while, correlation will be greatly reduced.

An even more basic example. Honda and Toyota are public companies. They both make import cars for U.S. consumers. They are the two top asian car companies in the U.S. As stocks go, these would both have a fairly high degree of correlation.

Then, toyota has a "news release" where it says it's cars have defective braking systems, and it will be recalling the last 3 years to replace the brakes.

Do you think these pairs will still have a high degree of correlation when the market opens tomorrow?

So, again... this system can (and will) have terrible breakdowns due to news announcements. Just ask anyone who trades it through an interest rate announcement

Greg

ok, but we can implement the rule of not trading
1 hour before and 1 hour after the news !