The best strategies recognize the very real element of randomness inherent in price action. In order to automate that, a programmer must know exactly when and where to expect randomness to avoid churn. Other strategies lend to automation, but only work under some conditions, and not others. Knowing when to stay home and when to get in the game, makes all the difference in the world. Again, the average programmer would be ignorant to the broader landscape and conclude winning strategies are losers (and visa versa), when in actuality, they're not. In order for a programmer to successfully automate a great strategy, they have to become a trader first, learn the ropes, decode the great riddle, then quantify, code, backtest and automate.
When people talk of long term profitable automating , it is not possible to automate on the lower time frames , and in ranging markets.The rules coded can not be identified with higher probability , in a fifty /fifty range.
Not true. Some of the best pay-dirt is in the sub-1 minute charts. Again, a trader has to know *when* it's profitable to dig, and when to stay home. Not all HFT shops quote stuff and front-run. Like Cornix said, range is relative. A 5 minute chart can range, but on the 30 or 10 second, there's big trends. And the opposite is also true. The second charts can be ugly, but on the 1 or 5 minute chart, a nice, clean trend. Knowing when, where and why to look is important.
I am not disputing the manual vision of such trades , I quite agree on a manual basis. Have you ever coded trends on 1 minute or 5 minute , or have you ever provided logic to a coder?Have you ever back tested such logic? Provide a sample code in plain English of a 5 minute trend or a 1 minute trend.How do you code it? When does the code trigger a trade ? I doubt if any of you will come up with a code to prove your opinions.
Not sure if it works in an automated form, but if one tries to code say 5-min inside bars which do not overlap with a few bars before and go for a fixed stop/target I think it could be profitable. But fully automated strategy will always lack one very important moment: context. I doubt it's possible to include context in a mechanical algorithm and it can make a huge difference in edge when trading the same signal on the same time-frame.
Yes, I have (backtested) and they work. Triggers are classic reversal signals. Everything is in the public domain. The simplest entry rules can be the most profitable, if you know when to stay in and stay out. As far as proving anything, why would I do your work for you?
I don't want you to work for me.I have ready made templates and codes and done thousands of tests. It is difficult to code pictures of art into machine language ,especially pictures on 5 min and 1 min.At least I can't do it without hitting the 50 /50 no edge code scenario.
Try lower time frames. I don't code art. I quantify simple relationships that compose the larger picture. It also helped to include a fail/reverse clause into programs. Another tip: add to winners, Livermore-style. Pyramiding works so the initial trade is only a fraction of the total position. That way, when you're wrong, you're wrong for a penny. Right, right for a pound etc..
np One of the keys is fail = reverse. Most traders can't tack when the market does. Either on a failed reversal (pullback), or a failed pullback (reverse). Largely, they follow patterns, but wed themselves to the trade. When the market throws it back, they can't adapt = stops get run. That's where algos are great.