The main issue when dealing with short lookback windows is stability. A lookback period of 2-3000 bars on a 1 minute chart is short (2 days) while a 2-3000 lookback period on a 60 minute is long (125 days). The stability of the coefficients in my ...

System to hedge every major dollar pair for profit / low DD

I'm curious as to the motivation of EA builders (those who spend the time writing an EA builder, creating a website, etc.). Why do they do it? Free ideas? The challenge? Because they can? (serious question by the way)

Expert Advisor Builders

Hugh, The comments so far are on target with my understanding. Generally speaking arbitrage (true arbitrage, not picking off slow brokers) is a game of the fastest and is typically out of reach for retail traders. But those points aside, many/most ...

Triangular Arbitrage

Try this format: (use forward slashes / not backward slashes for directory dividers in a valid R path. Default save workspace is '.RData', while the default script extension is '.R'. Rx("save.image('C:/yourdirectory/moredirectory/yourfile.RData')"); ...

Synthetic hedges, cointegration, mean reversion and similar stuff

This is actually a very good observation. Yes, it is repainting because the sizes (as Mediator pointed out) have changed. Everyone who tries this should study the following video and think about the implications. (make particular note of how much / ...

System to hedge every major dollar pair for profit / low DD

Bollinger bands are really standard deviation bands. Bollinger popularized them so the name stuck. I prefer calling mine standard deviation bands. The thing about standard deviations is that the assumed statistics at the 1st, 2nd and 3rd ...

System to hedge every major dollar pair for profit / low DD

I think I understand your comments. You're right that EURUSD by itself is not stationary. The video was posted to give ideas regarding how to implement a mean reversion strategy because it appears there are many new traders on this thread who are ...

System to hedge every major dollar pair for profit / low DD

Actually is isn't hard to say unless you're ignorant of the underlying concepts. I suggest those who are confused, spend some time educating themselves on what a mean reversion strategy looks like. This is not a case where the wheel needs to be ...

System to hedge every major dollar pair for profit / low DD

DB, To see R output you need to run DebugView.exe... (read THIS POST) The easiest way is to simply return the value to MT4 rather than trying to print it. so simply something like (off the top of my head and not error checked) : ...

System to hedge every major dollar pair for profit / low DD

That parameter controls whether the data is differenced before running the regression. Basically the way R's diff function works the current price is subtracted from the last price and so essentially the bar's net change is what the diff represents. ...

System to hedge every major dollar pair for profit / low DD

You can create error correction models in R as well using the vars package and requiring a ca.jo (Johansen) object. I never mastered this topic. Here is a link on the topic: url (I'll stop taking over the thread now)

System to hedge every major dollar pair for profit / low DD

And then just to continue with the code example, to do the ADF test you could call: library(urca) spread <- ns-ins[,'AU'] # store the spread ur <- ur.df(spread,type='trend',lags=3,selectlags='AIC') If you want to get help on the function try ?ur.df. ...

System to hedge every major dollar pair for profit / low DD

This is the output created from the previous analysis. This is about 34 days worth of data. This was lm(AU~NU) or testing the relationship between AUDUSD and NZDUSD. First plot is simply the AUDUSD price. Second plot is the predicted AUDUSD price ...

System to hedge every major dollar pair for profit / low DD

Here is some code to get the juices flowing for some of you who aren't that familiar with R. I hope it will inspire some of you who are unfamiliar with R to play with some of the packages and do your own analysis. Assume a price series AUDUSD, ...

System to hedge every major dollar pair for profit / low DD

It could be done - but this is where it turns into a more hands on approach by statisticians typically. It's been a while since I was into this material and I'm not a statistician. A good reference on the procedures used to perform a proper ADF ...

System to hedge every major dollar pair for profit / low DD

This basically comes down to what it means to run a regression with an intercept and without one. What are the assumptions of the model? Try reading the text on this page: (many others on the web, do a search on regression with / without intercept) ...

System to hedge every major dollar pair for profit / low DD

I took a look through Mediator's code a while back. I'm trying to be diplomatic... I wasn't impressed. Essentially the code turned a perfectly decent idea into a martingale grid strategy that sort of missed the boat on the underlying theory. A ...

System to hedge every major dollar pair for profit / low DD

sidhujag, The flaw in trend-o-mat is based on an unsound regression equation, not the reason you state. The formula in arb-o-mat uses sound mathematics to achieve cointegation. It appears to be the Engel-Granger formula. Does this guarantee success? ...

System to hedge every major dollar pair for profit / low DD

It seems to me that if the pairs are really cointegrated then mean reversion will be the best bet. You can do it in a number of ways, and the results from testing the different alternatives may make the decision more clear for this set of symbols. ...

Synthetic hedges, cointegration, mean reversion and similar stuff

Dirtybrown, Notice how different the eigen values are between the two latest pictures you posted. I've found that if you play around with arb-O-mat you can get it to produce very trend-like curves in some cases, simply due to the regression. I'm ...

Synthetic hedges, cointegration, mean reversion and similar stuff