Random entry systems

meanreversion

Senior member
Messages
3,398
Likes
537
Has anyone spent time looking at random systems.. e.g. let's say your designed system works well for a given timeframe, how do you know this is because it's a "good" system and not entirely due to chance?

I've started messing around with backtesting where my usual entry rules are replaced by random entry at the start of each month, but the exit rules are kept the same. So far the results have been fairly poor (which is good I suppose!) but I'm not sure if the testing is being done in a mathematically rigorous way.. e.g. the random system may have fewer trades than the 'proper' system.

Any thoughts on this one?
 
Why would the random system have fewer trades? I would think that it would tend to have more given that it doesn't need to wait for a given set of entry criteria to be met.
 
Yeah, more trades is more likely. I've tested using random system with trend filter and without trend filter and the one using filter performs significantly better (the trend IS your friend I guess). The annual return for the random system is usually only 1-3 pct below the "proper" system (although it sometimes can be more). However, random system usually has higher max drawdown, so the return/MDD ratio is better for non-random.

Having said all that, random system returns between 14-20 pct per annum with max 50 pct drawdown (over last 10 years), which is probably better than most equity funds.
 
PS just went looking at some of your posts, looks like we trade in pretty much the same way, I also identify trend on a higher time frame, and enter on overbought / oversold pullbacks on a lower time frame.
 
Thanks for the link, her conclusions are the same that I've reached. It's only recently that I've started looking at random entry and the results are not far off the proper system. Which in itself raises some questions -

1. If entry rules are not that important, why do people seem to focus on it 90 pct of the time?

2. It's possible for a random system to perform as well as fairly well regarded fund managers, except it doesn't charge fees. Are fund managers adding any alpha to their clients, or just themselves?

3. How many people trade random systems? Do the banks/hedge funds have any money allocated to this (the mind boggles -- imagine a hedgie testifying to Congress saying "Yeah, we flip a coin to choose the direction")

My conclusion at this stage is that, amongst all the hundreds of trading expressions, the two which are most pertinent are

- let your winners run, and cut your losses
- the trend is your friend [this one is quite important]
 
My conclusion at this stage is that, amongst all the hundreds of trading expressions, the two which are most pertinent are

- let your winners run, and cut your losses
- the trend is your friend [this one is quite important]

I totally agree with that !!!

This is not an exact science because markets simply aren't more or less than the sum of their participants actions, but get those 2 points down pat, forget the endless quest for a future peering crystal ball, and you've got it cracked.
 
but then what makes trading so hard and why the ridiculously high failure rate?
Are we really saying that all you need to do is have the discipline to hold on to winners and cut losers?
Enter either at random or with any old system and have 15pip stops and 55 pip targets and you'll make money?
Its interesting wathcing my friend who just strated trading. He's had a horrific start. Every single trade stops him out and then goes in the initial direction. He sent me hid statements. Only small money so no big deal, although i think he thought it'd be easier! He was about to abandon stops all together. I want him to register and start a thread. Maybe his trading will turn around with simply letting winners (when he eventually gets one) run and having tight stops on trades....
 
Well, there are countless sayings and truisms, some of which probably have merit, and some which don't. With software now, it's possible to test them all out; some of the more questionable ideas are -

1. a good win/loss ratio is essential
2. move stop to breakeven when possible (my testing only ever disproves this idea)
3. buy low/sell high - buy high, sell higher/sell low, buy lower is better
4. you never went broke taking profits - don't agree with this one
5. reduce your bet size when losing - statistically this isn't right, as it takes much longer to recover

The random system testing is interesting. Maybe the human need to feel right and vindicated about making correct market calls means it's a minority pursuit..!
 
all good stuff. I wont rep BSD again, as we clearly trade in similar fashion.

one extra piece of advice is that once you have a viable, robust system, dont start pfaffing about with other systems, and miss perfectly good trades because you were distracted by the latest shiny new system!!

I speak as someone who this morning missed a couple of perfectly identified trades on my primary system because I was entranced by DeMark fractals on some other chart, messing about on some new fangled PA gizmo. :mad::mad::mad:

EDIT: I need someone to sit next to me, and smack me in the head if I open up anything other than primary trading charts.
 
  • Like
Reactions: BSD
Maybe develop the system, get your buddy to trade your capital and go off and learn to surf/play guitar/lower golf handicap.

Here's another question - is there ANYONE out there who thinks they've maxed their P/L? For 99 pct of people, it's usually "if I had just done this, or not done that ... " etc., similar to when Sergio shoots 65 and moans about leaving "3 shots out there". Is this another human trait, we always think we could have done better?
 
Random trades will work over time as lonmg as you have deep enough pockets for the draw-down until you come positive. This tends to make it impractical, but what I mean is - open 1000 positions comprising half long and half short. Set a stop-loss of X pips and profit target of X+1 pips, where X is a very modest figure. In such a large statistical sample, its likely that half your trades will be right and half wrong. The 500 winning trades will therefore make 500 pips profit in total.

But who would do this?
 
but then what makes trading so hard and why the ridiculously high failure rate?
Are we really saying that all you need to do is have the discipline to hold on to winners and cut losers?
Enter either at random or with any old system and have 15pip stops and 55 pip targets and you'll make money?
Its interesting wathcing my friend who just strated trading. He's had a horrific start. Every single trade stops him out and then goes in the initial direction. He sent me hid statements. Only small money so no big deal, although i think he thought it'd be easier! He was about to abandon stops all together. I want him to register and start a thread. Maybe his trading will turn around with simply letting winners (when he eventually gets one) run and having tight stops on trades....

Good points.

I think this goes quite a way to explaining the failure rate:

"An Analysis of the Profiles and Motivations of Habitual Commodity Speculators

The focus of this study is the habitual speculator in commodity futures markets.
Responses to a 73 question survey were collected directly from retail commodity brokers with offices in Alabama. Each questionnaire recorded information on an individual commodity client who had traded for an extended period of time.

The typical trader studied is a married, white male, age 52. He is affluent and well educated. He is a self-employed business owner who can recover from financial setbacks. He is a politically right wing conservative involved in the political process. He assumes a good deal of risk in most phases of his life. He is both an aggressive investor and an active gambler. This trader does not consider preservation of his commodity capital to be a very high trading priority. As a result, he rarely uses stop loss orders. He wins more frequently than he loses (over 51% of the time) but is an overall net loser in dollar terms. In spite of recurring trading losses, he has never made any substantial change in his basic trading style. To this trader, whether he won or lost on a particular trade is more important than the size of the win or loss. Thus he consistently cuts his profits short while letting his losses run. He also worries more about missing a move in the market by being on the sidelines than about losing by being on the wrong side of a market move; i.e., being in the action is more important than the financial consequences."

An Analysis of the Profiles and Motivations of Habitual Commodity Speculators

I honestly believe that given a choice most people would prefer being "right", and being seen as being right and accordingly clever, over making money.

Who was it said that it's amazing how much money one can make by not being perfect, think Larry Hite eh.
 
chris capre says in a recent vid when he worked for a broker he studied the top 100 winning accounts. only 1 was a scalper all the rest were long term ie 4h and above.
 
There's more to it than that, and I don't think it's possible to prove/disprove if random entry works purely by using a simple arithmetic example. My random system enters using a coin flip, but the scaling in and the exits are pre-determined, and this is where the money is made. A system which randomly enters and then exits based on a simple profit target would lose money over time, after commissions and bid/ask are accounted for.

Entry and exit rules are both important, but the entry rules are not nearly as important is believed, is my conclusion.
 
all good stuff. I wont rep BSD again, as we clearly trade in similar fashion.

So at a minimum that's you and meanreversion and me then already.

Shall we go start a hedge fund or what !!!

:D
 
That's the route I want to go down at some point. The advantage of pure rules trading is that the size should matter less, i.e. a discretionary trader will act differently if you double his capital, a systematic trader won't.

What I find amusing about rules trading is that even if the rules SEEM simple, the market always throws you curveballs which test your belief or approach. E.g. the trend filter has been falling hard and LOOKS like it's about to go negative (even though it's still tiny positive) and they you get a buy signal. Buy or not buy? Hahaha

Something else which is simple - making a cup of tea. But milk first? Tea first? How long to let the bag brew etc.

Nothing is TRULY simple if you start to dissect it piece by piece.
 
Some info here:

http://www.trade2win.com/boards/gen...setting-up-incubator-fund-uk.html#post1085600

Re track record, amazingly Larry Hite only had his back and forward whatever tests when he got MAN to come onboard.

But for a discretionary trader, I dunno, whatever your investors deem statistically relevant I guess.

That said if one wanted to start a hedge fund I'd say a mix of systems and discretionary trading would be best.

That's what I'd want to go for just to balance things out a bit.
 
Top