…age isn’t always a disadvantage. With more and more auto trading stuff appearing, I started tempering with the thought that I may end up throwing my hat in the ring.

Some back testing, and I’m already headed into unexpected directions.

The problem with automated trading is that you always end up with something overoptimized, as you want to present good results. In a bull market buys would tend to make money and sells in the bears. The core question remains the same: what is the current market? By the time you can identify it, you may be too late for the party. 

There is only one game, getting both disances right: the distance to the stop loss and the target (statistical edge), and you need the stop loss to be a whole lot closer due to trading costs, yet your probability needs to exceed a coinflip just to tread water.

Beam Around CRE>

Max draw down: 13.89%

14 years of indicator / EA programming is a master’s degree in automation.
I have written my first few million lines of MT4 code, and I picked a few of my latest and best filters to make a fully automated routine.
It was made for EUR/USD with raw/ flexible spread. You could try it on other pairs at your own risk, but the two decimals shift for Yens is not included.
You get to set the size, a blank-out server time (to exclude broker spread ballooning) and if you want the holding to be closed out on Friday if in gains before the close.
The program opens 1 or 2 positions per signal. The second would only be opened if the trigger candle’s high / low gets re-visited within 5 hours.
All opens are at market.
For figuring the size, factor in the possibility of a maximum 45-pips move agaist you.
Yet another option is to lock in better than break even (5 pips) when in 25+pips of gains. Trail stop at 1/5 th of the gains.
The target is 90-pips for the first position and 230 for the second.
There can be 2 directional holdings simultaneously at max.

More about the signals? Propietary combinations of oscillators, moving average displacements and sequences. It works with 30-minute data.