Do you think there is any trading system suitable for all circumstances? - page 6

Scott Allen  

There are some things that are universally true, regardless of market regime. For example, on most timeframes, a Yo Sen or In Sen candle (aka elephant candle) is followed like 80% of the time by another candle in the same direction. I've tested it, and had like 80%+ win rates just exiting on the next candle.

The problem, though, is that even with that high win rate, setting the stop loss at, say, the opening of that Sen candle, is such a big stop compared to the average profit that it doesn't necessarily end up being a profitable system. The high probability signal part is easy. It's all about the trade management after that. I've talked to so many algorithmic traders who, in hindsight, wished they had spent more time and effort early on in regard to position management vs. entry conditions. I'm pretty much finding the same thing myself. All the common trailing stop methods generally work better than nothing, better than fixed points, but really, your trade management needs to have unique logic for when to tighten your initial stop loss, when to move to break even, when to lock in profit, how to allow for a pullback or two, etc.

IMHO, focusing on your position management, more so than optimal entries and exits, is going to be your best track to any kind of universal system. Because once you have that, you can use it with whatever entry and exit signals you want.

Arpit T  

I have been following up some systems and If you want to get a system which works in all conditions of market.

You have to follow up this:

What is not followed up by market all time? Its the Volume. Volume is unpredictable as every trader is allowed to trade a maximum qty defined by broker's max cap. A broker put max cap on trade qty by a single trader to avoid huge candles, But it does not limit a trader to only trade that qty. allowed by broker, a trader may use friend or family account or any organization's account to put any order he wish to trade. If he do that in any particular time, none of the strategy will work in this case 100% but most of the purest and proven strategy where people are trading from years and making proiit will try to act as a resistance or support for short term but will be overtaken by trader's huge order. Many people do not know this so they keep thinking there is something which works always. 


To solve this problem there are only two ways. First adding a stop loss and second is use average or pyramid method or martingale so you win in most of cases but its never explorable as an all time working method on chart by any technical analyst.

Busingye Tusasirwe  
Sze Shing Fong:
Do you think there is any trading system suitable in all circumstance.

No Please. It is like asking whether there is a vehicle that can never have an accident.

Lorentzos Roussos  
Amir Yacoby #:

There can be one or more at any given time. For that to happen, it should be versatile and sophisticated enough and maybe even analyze the market history before it trades.

Trading is an outsmarting game, at a given moment there can be someone that outsmarts the others, that might change when the others get smarter as well.

If you know the chess rating world, and especially the chess programs rating world, then it happens in chess software as well. AI programs are competing each other, and get an ELO rating for each new version. The world champion 5 years ago is not necesserily the world champion today.
Today's best software has en ELO of about 3700, while Carlsen the human world champion has an ELO of 2850 when each year the software gets stronger in about 50 ELO points.

So in equivalence, there is something in trading that is the ELO of a software, which points to it's versatility of market conditions and it's ability to make profit - given the current other market participants which might evolve as well (which will require evolving of that software to stay in even ELO).

I searched for ELO rating and found this . 

It's relative however , i mean if a coder evaluates his "algorithms" with ELO it is only relative to this coder's solutions . 

A universal rating system for trading systems would be difficult to implement because of the "non participants" (good or worse systems) that can provide a wider comparison between systems , and because it would be restricted to a community (i.e. mql5 , ff , tv etc) , and biases etc on top of that.
(Not to mention there would be no incentive for the owner of a good trading system to rank it among others as by having a good system they have "won" already in essence.So a "universal" ELO Rating system for trading bots would essentially discover the best of the worst)

So with that in mind the Elo Rating could be applied to patterns . A coder could collect a wide enough set (you could argue universal) of patterns that will more or less be similar across platforms. 

[Deleted]  
Sze Shing Fong: Do you think there is any trading system suitable in all circumstance.

I don't think so , every symbol has its own market factors, the "fit for all" indicator just a lagging indicator, it's just price originated indicator. 

In fact you just need to make clear one symbol's connections , you can earn good profit. 

Reason: