Is 100% Mechanical Trading Possible?

[ratmpower 10]

To answer original question - Yes, 100% mechanical trading is possible. The more important question however is - how effective or profitable is 100% mechanical trading.

 

Most the the automated trading system today are based on historical prices (you do some giant complex math-magic on the past data to predict the future price). IMO simply basing your automated strategy on historical prices is often not enough. Generally, the future price is driven by response to the events (earnings announcement, FDA drug approval, interest rate, government decisions etc). You must be able to take these into consideration while developing your strategy. That is why we encourage people to develop strategy on their own in any programming language of their choice. We only provide REST API to keep track of strategy and its performance, and leave strategy development to the users.

Edited January 14, 2012 by MadMarketScientist
urls removed

[BlueHorseshoe 10]

I run two systems that I've developed that are 100% automated in TradeStation.

I also do discretionary trading on the side. Actually, I 'm learning discretionary trading on a simulator during the market hours. I hope to actively trade non correlated markets against my automated systems for diversification.

 

Not that I'm wanting to suggest that this is what Sansjr is doing, but this can be a great approach for those who struggle against a need to watch the markets and constantly be involved . . .

 

Assuming that you have enough capital, it's perfectly plausible to run several profitable autotrading systems trading multiple contracts, and then to trade around these with a single contract on a discretionary basis. The latter can almost become pure amusement, gambling, or 'vanity trading' - it doesn't matter if the amount that is lost is small compared to the amount that is made by the profitable systems. This puts you in no different a position to, say, a multimillionaire who makes his money selling crisps, but likes to dabble in financial spread-betting at 50p per point 'for a bit of fun'.

 

If, at the end of it, you do uncover some hidden intuitive knack for discretionary trading, then great; otherwise it's no big loss.

 

Having said all this, Swansjr has obviously taken an even more sensible approach by confining the development of his/her discretionary trading to simulated accounts.

 

Hope that's helpful to some of you with bucket loads of capital and thirst for constant action!

[mickro 10]

If you can automate 90% of the strategy (which is obviously better than 0%, what is stopping you from automating the other 10%? if this "discretion" that you speak of has specific rules, why can't it be automated? Perhaps you are afraid of what pure objectivity will reveal about your strategy.

 

I'm fully automated and profitable. So, looking back through this thread, it is possible to do this on a live system.

 

100% is a milestone... Actually, the first automated trade for me was a milestone. You start slow and work up.

 

But I think the real trick is the level of dedication that it takes. You cannot take something off the shelf and make it work because there are just too many pieces moving. I believe you have to roll your own system if you want any real scalability.

[mickro 10]

this is the trading process:

 

you monitor the price/volume movements,

(HH, HL? LH, LL? 50 Week High? at pivot point? reached support/resistence?, etc.,)

 

you analyze the price/volume behaviors,

(bounce off double top? break out of formed channels? stuck in congestion? volume increase with price? etc.,)

 

based on the above...

you can come to a decision on where you are in the bigger scheme of things

and what opportunities are available

(are you in the midst of a major up trend? in the midst of a minor downtrend retrace?)

 

you act on your conclusion -- how best to capitalize on the opportunity

(e.g. buy call? sell put? buy stock on margin? short ETF? enter a credit spread? long on buy stop? etc.,)

 

 

if you can quantify the above, you can mechanize your trading.

 

100% requires automation. Means initate relationship with hardware, kernel, risk controls, position sizing, order routing. Market microstructure and that stuff you never cared about just become important.

 

Knowing how to trade is < 10% how to actually make money.