Funny thing is, I'm taking an online course on neural networks and the professor showed how computers can guess images fairly accurately (see attached). And training a computer to see objects is in a different ballpark than asking a computer to identify chart patterns. In the latter, you, as a human do the learning and encode your observations in code - much easier than asking a computer to figure out chart patterns on its own.
And CAPTCHAs are used in processing old books to computer text. You know how you see two captchas. If you get both right, it uses one of them to encode words of old books into computer text. And I wouldn't be surprised if they're also training computers in vision.
Algorithmic/Automated trading is an interesting topic in the respect of this conversation.
In most cases, this is what prevents people from moving forward with their trading.
In my opinion, if you are looking to write a computer program to automate trading outright directional positions but you cannot actually do it yourself yet (i.e. you are not yet a profitable trader), then you have close to zero chance of success.
I think a lot of people try to replace learning to trade with trying to program a computer to do it for them. It has certain allure, I admit but the realities of it are quite harsh. Programming won't teach you how the markets work, trading will. If you don't trade, you will never understand how things work enough to know how to automate.
Once again, it comes down the individual being the reason they fail.
The following 3 users say Thank You to DionysusToast for this post:
For those of you who have automated in some fashion your trading method - whether by creating software to trade your mechanical system, or whatever - I'm curious to hear about whether it has provided any benefits over the way you traded previously? What would say those benefits are, if any?
The biggest benefit for me is that now I do not have to think about "should I exit now? what should I do? ..." The key is this need for decision making disappears in the heat of the moment. I still have to make trading decisions, but not in the middle of trades. It is much easier for most of us to make decisions when we are under little stress.
Biggest drawback to mech. trading is the loss of control. With auto systems, you turn everything over to a machine, and when the machine enters goofy trades, you sometimes still get anxious. Plus, there is always the lingering doubt "how long will this system last? when should I pull the plug?" Finally, when automation screws up, do you "sync" the system back up right away - that might change my risk:reward. Maybe it is better to wait until next trade.
What it comes down to is this: If one thinks that trading a mechanical system eliminates all stress, they are dead wrong. The stress is still there - just like discretionary systems - but the stress manifests itself in different ways.
The following 3 users say Thank You to kevinkdog for this post:
I don't think there is a more straightforward explanation than this, if profitability, and not academic satisfaction, is the goal. The desire to know relates little to profitability much of the time; when I have known the most, I have often traded the poorest, but when I accepted the unknown, I have had my best days.
The following 2 users say Thank You to josh for this post:
Did you read the thread? One obvious answer is bactesting and seeing whether a strategy (which could've been discretionary) is profitable or not. Breakeven or slighly losing trades can be profitable for months, and you wouldn't know it by pressing buttons day in and day out.