Yes, but as a human, how many times can you resist those trades that are 'just outside acceptable parameters'?
It's only a matter of time before you get the urge and commit.
Every discretionary trader is going to break their rules.
However, the winning trader will break their rules less often than the losing trader.
I'm a discretionary trader btw although I've written a lot of code for high probability trades.
I've never programmed a bot to make trades for me.
I don't mind missing out on trades that fit my rules. I don't enter sometimes because
the market isn't black and white to me. If the trade I don't enter ends up being a nice winner, I'm okay with that. There's always more.
Discretionary and auto traders develop their strategies in the same manner.
Every strategy that has a perceived edge is based upon historical performance/experiences. You're simply basing your supposition that you can make money by predicting market actions based upon previous experience.
You observe a certain event, and after numerous occurrences, you decide that it's a relevant pattern. From that point, it's simply a matter of calculating the expectation with a set of input parameters (i.e. slippage, commission, profit/loss, etc).
even if you were to simply execute trial/error while trading live, you're still developing an edge that you'll apply to future trades and that is historical or past performance based.
I think there's some misconceptions or misapplications of the word "backtesting."
Obviously there's a spectrum of effectiveness.
When I backtest, I compare the backtest to forward simulation and live forward.
That way I can guage over given time periods, how my backtest "translates" to sim, how my sim translates to live and how my backtest translates to live.
You can do this with just about every performance metric (P/L factor, ratios, win%, drawdown, etc).
If I see a 47% win rate in backtest, and observe a 49% rate for forward sim and then observe a 46% win rate (concurrent with sim and retroactive with backtest), then I can start drawing comparisons and developing confidence intervals about the expectation.
I can then take that "translation" and apply it (with a degree of error and confidence) to other time periods that I haven't traded live.
If you're discretionary trading and you're altering your rules AT ALL, then this type of analysis is useless. Changing ANYTHING nullifies the whole concept.
I think some of you guys have this perception that auto guys write a program, backtest it once, see a black number and pull the trigger on live trading. That may be how some other guys do it....but not everyone operates that way.
This post has been selected as an answer to the original posters question
Perhaps the better question is - why would you bother?
It would be a mammoth task for many traders to do this. For me, it would take possibly a year to define, code, refine this and then I could never guarantee the program would 'think' like me. All for what? To do something I can do anyway but have the chance that a shift in the markets would have me re-coding it all instead of re-thinking it?
I could probably write a program to drive a car. It would also take many years and I'd never fully trust it. One of my mentors once put it like this:
"Let's say you are in a car at a junction. There's a car coming at you and you feel you have time to pull out in front without causing that car to slow down"
Now - we can all do this. This is because repetition of an activity causes it to become second nature. When you are in the car, you don't think "That car is 600 metres away, travelling at 56kmh, my car can accelerate to 56kph withing 6 seconds and therefore I have enough time to go". That's a set of rules. That is how you would mechanically drive a care but nobody drives that way, except Rain Man.
So to trading, especially day trading. Are we Rain Men who need to go through the same mechanical process as when we are beginners, or will some of this become second nature? It seems most people think that trading isn't a skill you can become adept at just like driving, tennis, golf or any other skill we learn.
I find this most odd.
The following 3 users say Thank You to DionysusToast for this post:
You are from the old school and I think that people like you will never have a break, because they don't have imagination and cannot invent or even adapt to the changing world.
There are traders, maybe more old school than you, who still trade with a pencil and paper. You may laugh at them, but you're exactly like them.
To your example I just want to point that airplanes are already flying on autopilot. I bet that when this was invented you gave the same statement of how ridiculous it is.
And doesn't matter what you say the cars are going in the same direction.
Now about trading this is different for now. It is still too difficult to develop adapting bots, but it will happen.
As for now there are advantages and disadvantages in each type of trading, and they were mentioned here, so try to think more progressively. I know the saying "You can't teach old dog new tricks", but in this world when you compete with new dogs you have to.
I am in the mechanical camp, but I have to agree with you here. If you are not a programmer, and have no inclination to it, there's no reason you should try to program your methodology.
But I believe you should write the rules down in detail, and understand how your method works, why it works, when it works. During the raging bull market in the dot com era, many traders were making money hand over fist. Once the tech bubble burst, they lost everything, and could not make a dime to save their life. They mistook the nature of a raging bull market, and thought it was their natural trading ability and skill.
I do believe there are successful traders who have learned the skill of trading, or have an innate ability for the art of trading. But, I have to say that those guys are few and far in between. The problem is, every newby trader who opens an account thinks they will be one of those traders with this natural ability. Guess what, 95% of them don't, and 95% of them fail!
The problem with trading from that intangible, unidentified, ability, or skill, is that if you haven't broken down your method down to it's core, and don't know exactly the reason why it works, then when it stops working for some reason or another, then you are screwed, because you have no idea how to fix it.
The following 2 users say Thank You to monpere for this post: