So where do you make YOUR money?

[DionysusToast]

Then I would submit that you did not have an effective money management strategy.

To what % did you limit your risk for each entry?
How did you set your stop loss? Not to what value, but by what analysis?
How did you set your trailing stop, if you had one. Same qualification as above.
​

When you answer these questions, then we can try to judge if you truly conducted a "rigorous" test.

Howard - trailing stop + random entry is a trend following system and it needs the underlying to trend a lot more than it doesn't.

It's shimple stuff.

I think the onus is on those claiming you can trade randomly to prove it works.

Otherwise you may as well claim trading on moon-phases is the way to go.

 

[tommog]

HowardCohodas, forgive me but I am not here to claim I have a definitive answer one way or another, anyone dealing with an experimental concept would be foolish to do so. This has just been my findings.

However in answer to your questions. Firstly I did not assign a percentage risk, we may be talking cross purposes here but I wanted to keep data as raw as possible and see if random entry can take ticks from the market. If you have a negative expectancy no amount of percentage risked on a trade will make a difference. EMploying something like fixed fractional sizing would smooth the equity curve but wont turn a negative edge into a positive one.

The only stop loss was a trailing stop. Again, the more factors you add to it the more you are starting to build the basis of a system and I would argue that would not really be random. However the stop loss was based as a function of the Average True Range so it adapts to the market and avoids curve fitting i.e if you were to use a fixed number such as 10 ticks ALL the time there will be times when this is more effective than other thus open to curve fitted results (not random).

 

[HowardCohodas]

It's shimple stuff.

Wait a minute. You are a Yank. What does that word mean?

 

[DionysusToast]

Wait a minute. You are a Yank. What does that word mean?

Are you sherious ?

 

[HowardCohodas]

Howard - trailing stop + random entry is a trend following system and it needs the underlying to trend a lot more than it doesn't.

It's shimple stuff.

I think the onus is on those claiming you can trade randomly to prove it works.

Otherwise you may as well claim trading on moon-phases is the way to go.

No disagreement there.

Besides Tharp's published work, I did some research of my own. I've mentioned before that I am moving into a new office (my son's old bedroom) and am slowly going through my old office to discard stuff I don't need (that's what I tell my wife, that's my story and I'm sticking to it) and moving stuff of current interest. One of the things I am keen to find is the report on the study I did on random entry.

My old office also contains three computers I haven't fired up in years. I'm hoping the report and the code (Visual Basic IIRC) is still around. It would be worth the trouble of resurrecting it to add facts to what is, so far, only a battle of opinions if you disregard Tharp's work.

However, my friend, you state some things without factual basis as well which I am also keen to refute since it is not representative of my experiments, namely your first sentence. Were you to state that as your opinion or support it otherwise, I would not take umbrage to it.

 

[meanreversion]

Howard, please give up this argument straight away. This is DT's absolute pet subject and he will pursue you forever (or 20 pages, whichever is longer).

I too have done research into this and random systems with a trailing stop can make money, but then it's not really random, it's just a trend system.

A truly random system would be this -

take 100 markets, choose one randomly, then flip a coin to go long or short, choose your position size randomly (0-100 pct of equity) then flip a coin every single day to decide whether to stay in or get out.

No-one has ever tested this kind of system (that I know of), and EVEN if it did make money, who on earth would trade this way?

 

[DionysusToast]

Actually MR - this guy has his prop traders do it...

http://www.futurestrader71.com/

A few people on one of his web seminars tried too. The results weren't pretty.

I agree though - random entry + trailing stop = trend system. It's a bit like doing quadratic equations - you can move 'edge' from one part of your system to another. It just takes a little introspection into what it is you actually have once you have built the system.

In the Van Tharp experiment, his mistake was that he set out to prove random entry worked and when he got the results he wanted, he didn't really stop to think about what he'd produced and why his system worked on the chosed data set.

 

[HowardCohodas]

HowardCohodas, forgive me but I am not here to claim I have a definitive answer one way or another, anyone dealing with an experimental concept would be foolish to do so. This has just been my findings.

However in answer to your questions. Firstly I did not assign a percentage risk, we may be talking cross purposes here but I wanted to keep data as raw as possible and see if random entry can take ticks from the market. If you have a negative expectancy no amount of percentage risked on a trade will make a difference. EMploying something like fixed fractional sizing would smooth the equity curve but wont turn a negative edge into a positive one.

The only stop loss was a trailing stop. Again, the more factors you add to it the more you are starting to build the basis of a system and I would argue that would not really be random. However the stop loss was based as a function of the Average True Range so it adapts to the market and avoids curve fitting i.e if you were to use a fixed number such as 10 ticks ALL the time there will be times when this is more effective than other thus open to curve fitted results (not random).

I have done lots of testing into random entry. The results are quite interesting IMO.
Can you make a profit? NO, under no rigorous test have I ever seen a random entry system make a profit

Random entry without money management is a nonsense discussion. Trading without money management is a outside my experience and practice.

Why do you assume that random entry has negative expectation?

Setting the trailing stop outside the noise level is sensible.

Not having a stop loss violates good money management practices.

 

[meanreversion]

I'm stepping out of this discussion. Howard, you have been warned.

DT ... all yours ...

 

[DionysusToast]

Random entry without money management is a nonsense discussion. Trading without money management is a outside my experience and practice.

Why do you assume that random entry has negative expectation?

Setting the trailing stop outside the noise level is sensible.

Not having a stop loss violates good money management practices.

This is simple.

Present the system Howard and we can discuss the specifics of what the whole system actually is and when it would work/fail and how it's not random.

As a silly example - let's say I had a magic indicator (and I mean real magic) that told me which direction the ES would move over the next hour. I then flip a coin and go long or short. Immediately after entry I consult my magic indicator. If the magic indicator tells me I am in the wrong direction, I close the trade. If the magic indicator tells me I am in the right direction I stay in. Now - I am sure you will agree that this is not in the spirit of a random entry system, but that a pedant would claim it's still random.

What's your opinion - is the magic system random ? Letter of the law or spirit of the law ? :whistling

 

[HowardCohodas]

Howard, please give up this argument straight away. This is DT's absolute pet subject and he will pursue you forever (or 20 pages, whichever is longer).

I good Socratic dialog is designed to distill truth. Truth is never worth abandoning. Besides DT makes me think hard and at my age that's a good thing. I don't think I'm good enough to make him think hard yet.

I too have done research into this and random systems with a trailing stop can make money, but then it's not really random, it's just a trend system.

Duh!

A truly random system would be this -

take 100 markets, choose one randomly, then flip a coin to go long or short, choose your position size randomly (0-100 pct of equity) then flip a coin every single day to decide whether to stay in or get out.

No-one has ever tested this kind of system (that I know of), and EVEN if it did make money, who on earth would trade this way?

You and I have different concepts of a random entry systems. Let me state my experiment again, but with more detail.

1. Randomly choose a trading instrument.
2. Randomly choose a starting date at least 3 years back.
3. Randomly choose a direction (long or short)
4. Set the combination of transaction size and stop loss based on sound money management principles, e.g. 1%-2% of account size.
5. Set a trailing stop. I can no longer recall, but I believe it was based on True Range.
6. Once stopped out, go back to step 3. Continue for 3 years. The 3 years was arbitrary, but meant to cover more than one market cycle.
7. Go back to step 1 for 1,000 runs.

 

[HowardCohodas]

What's your opinion - is the magic system random ? Letter of the law or spirit of the law ? :whistling

The best I can recall of my experiment is shown here.

I really, really want to find my experiment. I want to humbly apologize if I made an error in my experiment and my results are not valid. Or I want to gloat if no-one finds any flaw with them.

Meanwhile, you are making my brain hurt.

 

[DionysusToast]

Your system (2 posts back) is EXACTLY what Van Tharp did. The stop was indeed ATR based.

This is a trend following system. The trailing stop enables you to benefit from the trend. 50% of your trades will be in the wrong direction and will be stopped, which will incur slippage and fees. Trades in the right direction will run in a trending market and need to make up for losers + slippage + fees. You also need volatility to be relatively stable otherwise any average of ATR will be meaningless.

So - this system can be summarised as:

Trend following system that requires reasonably consistent volatility.

This is no more random than my magic system. Applying this to a market would be done in real life only if you were confident that the market would trend.

 

[DionysusToast]

I really, really want to find my experiment.

You checked under the sofa cushions ?

 

[HowardCohodas]

Your system (2 posts back) is EXACTLY what Van Tharp did. The stop was indeed ATR based.

This is a trend following system. The trailing stop enables you to benefit from the trend. 50% of your trades will be in the wrong direction and will be stopped, which will incur slippage and fees. Trades in the right direction will run in a trending market and need to make up for losers + slippage + fees. You also need volatility to be relatively stable otherwise any average of ATR will be meaningless.

So - this system can be summarised as:

Trend following system that requires reasonably consistent volatility.

This is no more random than my magic system. Applying this to a market would be done in real life only if you were confident that the market would trend.

You recall Van Tharp better than I.

We are tantalizingly close to agreement. Permit me some tiny quibbles.

My conjecture is that It requires dynamically identifiable volatility rather than consistent volatility. Which is why ATR is so attractive.

You appear to be jumping between the idea of a random system and a random entry system.

The only conjecture I was testing was, "Will proper money management and a market that is purported to trend 20% of the time yield a profit in excess of commissions and slippage without consideration of choice of trading instrument or entry direction?" I arbitrarily added re-entry into the same instrument with a randomly selected direction until at least 3 years of trading was complete. My recollection of Van Tharp is admittedly weak, but I thought at exit he went back to randomly picking a new trading instrument.

 

[Shakone]

This is simple.
As a silly example - let's say I had a magic indicator (and I mean real magic) that told me which direction the ES would move over the next hour. I then flip a coin and go long or short. Immediately after entry I consult my magic indicator. If the magic indicator tells me I am in the wrong direction, I close the trade. If the magic indicator tells me I am in the right direction I stay in. Now - I am sure you will agree that this is not in the spirit of a random entry system, but that a pedant would claim it's still random.

What's your opinion - is the magic system random ? Letter of the law or spirit of the law ? :whistling

Wouldn't you also agree, that the more you impinge upon anything to do with the exit, the more you are not testing a random entry system, but actually testing a random entry plus...(insert your allowable strategies for exit) system. Where is the middleground on exits? You have to allow some strategy on the exit, but you won't allow instant exits...what about soon after entry exits...5 mins after entry exits? Until there is some middleground, there is no sensible way to answer the question.

Is there a video of those prop traders (futurestrader71) flipping the coin trading?

 

[TheBramble]

You want random?

At 09:23 (CET) tomorrow Thursday December 9th. Go short EURCAD.

Use a stoploss equivalent to 50% of the max-min of prior 11 bars.

Exit at 14:34 (CET).

I leave it to others equally fuggin amazed at the capcity of some to endlessly debate hypotheticals while seemingly unable to realise there is nothing stopping them rolling their sleeves up and getting stuck into the empiricals.

Load an excel spreadhseet with randoms for FX pair or instrumernt to trade. Time to enter. Direction. Prior number of bars to calc 50% of stoploss. Time to exit. All random. Do it for a few weeks.

Let us know how you get on.

 

[meanreversion]

This whole discussion is absolute horse poo.

Random or random entry, call it what you want, it's a silly endless discussion with little to no point as no-one will ever trade this way.

Random b0ll0cks more like.

 

[wackypete2]

You want random?

At 09:23 (CET) tomorrow Thursday December 9th. Go short EURCAD.

Use a stoploss equivalent to 50% of the max-min of prior 11 bars.

Exit at 14:34 (CET).

Have these signals been backtested?
I think I throw about $4,152.89 into this to see how good this works.

Peter

 

[HowardCohodas]

This whole discussion is absolute horse poo.

Random or random entry, call it what you want, it's a silly endless discussion with little to no point as no-one will ever trade this way.

Random b0ll0cks more like.

The purpose of the experiment and the discussions is not that anyone would trade this way. The key point is to emphasize money management as a core component of trading success. That, in and of itself, is an extraordinary result given the number of traders who ignore this valuable concept.

It is unfortunate that you have no appreciation for this worthy concept.