Please take a look at this chart:
It shows the number of pips of change from one minute to the next minute on EU, GU and EG.
I think I found why EU/GU works better than EG (at least on my tests, and it is the original method proposed by Kelton).
First, suppose you have EUR and GBP, and that both have 100 % correlation. We know that correlation measures the change from one period to the next period in percent. So with 100 % correlation if EUR is up 1 % then GBP will be up that 1 % also.
If you agree with this, continue
Now we know that the price of EUR/USD is about 1.3100 and GBP/USD is 1.6100 AND EUR/GBP 0.8136
If we have a 100 % correlation then 1 % change on EURUSD will be 131 PIPS and on GBPUSD will be 161 PIPS. Same percent move, 100 % correlation, but different pip move.
Now let suppose that we are looking at the charts with vitrite technique of overlaying, we have EU below GU for about 20 "visible" pips, so we buy EU and sell GU. Because we are comparing with EG, and we are buying EUR and selling GBP, then we buy EG.
Both move a -1 % , and our divergence increases a factor of 1.61/1.31. In fact we now have EU at 1.2969 and GU at 1.5939, because we bought EU it will have a loss of about 131 pips, while a GU will have a profit of 161 pips, with a net profit of 30 pips.
At the same time you will see that EG new value is 0.8136.... well, we are having 100 % correlation, so we expect that, isn't it? so while with EU/GU we are profiting with 30 pips, on EG we are not profiting at all, we still have our original position, without any profit, well, not only it is not in profit, it is in a loss because the spread.
So based on this observation, I can declare that for this method it is not the same to trade EU/GU and EG.
Also, because the correlation isn't 100 %, but about 80 %, we are going to have situations in which EG is going to have some profit or some more loss.
So what I think this method uses to win is precisely the difference in pip change of both pairs, because if we are on a trend (let suppose a downtrend) sometime in the future it will have a correction, we all know that market never moves straight down or up, but have some type of corrections or retracements. That retracement makes that the faster currency in PIPS crosses the slower currency in PIPS generating a profit.
If the prices were equal, then the moves will be parallel lines, but because we have some difference in the price of the currencies, we have lot of crosses. On the chart I provided the crosses are small, about 3 pips, not really useful, but on bigger time frame like 15M or 1H the crosses are 10-50 pips.
What do you think about my observation?