Menu
Sister Sites

Options

Bookmark Thread

First Page First Unread Last Page Last Post

Print Thread

Similar Threads

VennD - Any Pairs - Any Style - Any Time 740 replies

Is there any one who could recommend me any LongTerm strategy 24 replies

Does anyone using MB Trading broker? What is the minumum trade size you can trade? 2 replies

Anyone experienced better results with a true ECN broker? 3 replies

Can anyone share Buy-Sell-Close scripts to use as a hotkey without using F9 option? 5 replies

Is it true that anyone can succeed using any strategy?

  • Post #41
  • Quote
  • Edited 11:26pm Dec 17, 2011 11:11pm | Edited 11:26pm
Quoting jeanlouie
I think the issue is rather a lack of clarification on what an 'edge' means and then assumptions around it, because that's what I first addressed, someone said you need an edge plus mm, then mentioned that a 20-30% winning system can win over the long run, which means that the system which is worse at picking entries than 50/50 needs some adjustment of sl/tp scaling/adding to make it win out. So here's a conflict of definitions, an edge as in win rate over 50% with equidistant tp sl, and then adjusting tp sl for a system lower than 50% win rate....
Indeed.

And here is the explanation: http://www.forexfactory.com/showthre...17#post5231417

What you earlier referred to is keeping TP as a moving target and changing the trades as market moves, just like what many long-term traders do, especially the Market Funds. The best example of what you said can be illlustrated how Mongolians used to hunt animals for food in the ancient times. They used to start with a 500 mile circle, gradually narrowing down to few miles hushing animals of their choice into this narrow circle and whoever wanted to break the circle became prey if that was an animal of their choice. A Longterm trader may trail the market from behind and when it gives first sign of reversal should get out of the positions.
The Thief of Wall Street
Quoting jeanlouie
Exactly, win rates lower than 50% require some adjustment of tp sl or mm to win out.
Of course, but there is no adjustment. I think your assumption that TP:SL starts off as 1:1, and then needs to be "adjusted" is arbitrary. TP & SL are whatever the system says they are. A system with an edge simply means that it has a positive expectancy, using whatever TP/SL are prescribed by its rules.
Quoting jeanlouie
I must say I don't quite understand your position on the topic.

You say to win in the long run you need an 'edge' and money management, in terms of winning signals, what 'edge' is there if a system spouts less than 50% winning signals? The 'edge' of the system doesn't lie in the signals but in the maths around it, the tp, the sl, scaling in/out, moving stops, trailing, adding and so on.

My position on the topic is, whatever makes money over the long run, works.
man, you still got a lot to learn.
edge = positive expectancy (profits in the long-run)
win-loss ratio is just one of the factors in the equation. Win-loss and Risk-reward ratio need to work together. e.g. an average 50% win-loss and average risk-reward of 2:1 will make you lose in the long-run, while an average win-loss of 30% and average risk-reward of 1:10 will make you win in the long-run.
and no, the edge doesn't lie just in the maths around it. Random entries will make you lose in the long-run, no matter what the tp, sl, scaling in/out,... you use.

Yeah, question is what you define as long-run. Most people here think that long-run already means something like 100 to 500 trades. Of course if you just use such a small sample, then even random entries "seem" to work in the long-run.
But if you use a large sample like 10,000+ trades, then you will find, that in the long-run you don't have an edge / no positive expectancy by using random entries.
Quoting MaryJane
A closed trade is a line segment, nothing more - there is no other data than price A and price B (..yes, there is time, MAE, MFE.. but they are irrelevant for determining the entries vs exits importance). ......
Mary Jane,

Very fine post.

Entries and exits are equally important. Both are merely vehicles for adjusting net position, and the more accurately one adjusts one's position relative to market reversals, the more pips one banks.

The coin toss entry idea is bandied around frequently, but I doubt that anybody seriously uses them.

If profitable trading were as simple as an EA that entered randomly, cut losses short, and let profits run, then surely everybody would run one, and effortlessly become a forex millionaire. To folk who believe that MM or psychology is the biggest barrier to success, such an EA would seem the obvious solution.

However, it is not........
Quoting MaryJane
In the real world, any strategy should aim to come as close as possible to this ideal (anything less is hypocrisy). However, most of them will pick both A and B points no better than a coin toss, if applied mechanically with total disrespect to market structure and behavior.
The vast majority of EAs eventually fail not because of MM, but because their entries and exits are likewise no more effective than a coin toss, hence transaction costs defeat them at a rate proportional to the frequency at which they trade. In other words, by ignoring market structure and behavior, they have no real methodological edge; just as it is with the vast majority of aspiring traders.

David
Quoting Custos
edge = positive expectancy (profits in the long-run)
win-loss ratio is just one of the factors in the equation. ........
Custos, I believe you're correct. Expectancy (or profit factor) is the product of both win% and average net RR. Hence, for example, both system 'A' (40% win rate @ 3:1 RR) and system 'B' (80% win rate @ 1:2 RR) deliver a profit factor of 2.0

Everything else being equal, increasing the RR (e.g. TP pips : SL pips ratio of a given entry strategy) will reduce the win rate in approximate proportion. The trick is to find an inefficiency, based in market structure and behavior, that breaks this inherent equilibrium, and - voila! - a methodological edge.

No amount of MM can ever rescue a negative expectancy system. Expectancy determines whether the equity gradient is upward or downward; MM merely magnifies (or diminishes) its steepness.

I also agree with your comment about statistical validity. An edge must ideally prove itself over several hundred trades and many years of changing markets. Otherwise it's all too easy to be 'fooled by randomness'.

David
The most important part to any strategy is DISCIPLINE, without it, you will not follow your plan, thus leading to losses.
Quoting Custos
Of course there are, my best systems have an average win/loss ratio of 20-30%. But just taking random entries never turned out profitable in the long-run, at least not in my backtests, no matter which money management rules used.
I think you can probably exploit market kurtois with a random entry. Its perhaps not the best strategy in the world by a long shot, but its a very sound basis from which to start building a strategy.

There's no one size fits all answer to these kind of questions, for some strategies, the entry is extremely important, and a random entry is suicide, for other strategies a random entry is perfectly fine (and maybe even impossible to improve on)

The main problem with random entries of course is that they tend to result in rather random equity curves. I think it would be almost impossible to achieve consistant long term profitability year on year from a single system.

Even having a sound fundemental edge exploiting kurtosis, by random chance (bad luck), you'll get the odd bad year. Trading multiple instances of these kinds of systems could of course be a different matter and average the long term performance out a bit, so using that approach I think its probably achievable, and I know of at lest one trader running a small fund whose been doing something very similar for quite a few years with outstanding results.

The fundemental principle is that he's compounding profits on the good equity curves at a faster rate than he acrues losses on the poorly performing equity curves. Even simple fixed fractional MM would probably work, but I suspect he's probably using much smarter MM.

There's also probably 1001 other tricks you could use to improve the edge slightly over random chance, that ultimately would make a huge difference.

The key point that most inexperienced traders fail to grasp is how small an edge you need to become exceptionally wealthy, and conversely, how small a negative edge is required to destroy an account. A 1 pip per trade spread that leads to death by a thousand cuts, if reversed in the traders favour could double an account over the same time period. How many traders do you know who would express any sort of interest in something that returns on average 1 pip per trade ?. The answer of course is almost none, because they dont understand the nature of the puzzle that they are trying to solve.

To give an example, my automated system took aproximately 1000 trades this year and returned just less than 2000 pips. Thats a return of almost 100% uncompounded at 5:1 leverage (I'm not quite so stupid to use leverage quite that high these days, but if I had less capital, I would do, and I'd probably still remain within quite acceptable drawdown levels even at 20:1). I dont need to make 100% a year, managing drawdown is far more important.

Even as a discretionary trader, on my best ever year, I was only averaging slightly more than 7 pips a trade, and I made over 450% return that year.

I'd probably argure that I could achieve similar results with any strategy, but only because I understand the true nature of the game that's being played. Before I understood the game, I was one of those guys chasing 2000 pips a month.

If someone understands the nature of the game, then I think they can probably succeed using any strategy (assuming the strategy isnt specifically designed with a large negative edge). The reality of course is that most non profitable participants are completely clueless about what exactly they are trying to achieve, and they wouldnt recognise successful strategy even when someone hands it to them on a plate.
Quoting Custos
Random entries will make you lose in the long-run, no matter what the tp, sl, scaling in/out,... you use.
Random entries by itself will not make you lose. Its what you do afterwards that will decide your fate. Random entry with good money management and sound trade management principles can have an edge.

Let me give an empirical argument. At every time in any market there are always reasons to enter both long and short. That is why the market exists. And there are also profitable traders. So one must conclude that there are traders on both the buy and sell side that are profitable. Therefore a coin toss entry, which is a random entry, can be profitable.

Having said this, a completely random system will always lose money, no matter who the trader is. A completely random system will by definition be 50-50 to win but once the spread is factored, it will become a losing system.

If you don't believe try out a completely random system on a 1 min timeframe, on a pair with a large spread. You will know the result in a week.
Quoting Eklavya
Random entries by itself will not make you lose. Its what you do afterwards that will decide your fate. Random entry with good money management and sound trade management principles can have an edge.

Let me give an empirical argument. At every time in any market there are always reasons to enter both long and short. That is why the market exists. And there are also profitable traders. So one must conclude that there are traders on both the buy and sell side that are profitable. Therefore a coin toss entry, which is a random entry, can be profitable.

Having...
While there might be reasons to enter long or short at any given time, the reasons are not usually equal. You need to trade with the heavier side. If a hedge fund just sold, you should sell instead of buying with the Audi Corporation. Just saying.
Quoting Eklavya
Random entries by itself will not make you lose. Its what you do afterwards that will decide your fate. Random entry with good money management and sound trade management principles can have an edge.
then why don't we program an ea that will do just that and everybody using it will become a forex millionair.
Money management & trade management alone does not make a system profitable. The system itself needs to have an edge, which can then be magnified by good money management.

Random entries don't have an edge. Period. It's like talking against a wall here on ff.
Quoting Eklavya
Let me give an empirical argument. At every time in any market there are always reasons to enter both long and short. That is why the market exists. And there are also profitable traders. So one must conclude that there are traders on both the buy and sell side that are profitable. Therefore a coin toss entry, which is a random entry, can be profitable.
You forgot the vast majority of market participants such as banks and companies just hedging their currency exposure and the vast world-wide exchanges of money to be able to operate across the world.
Yes, there are always reasons to be long and short, but most of the market participants are not after profits, but just hedge exposure. So your argument is flawed.
And just concluding that two traders, who individually have great system, one a contrarian and one a trend following, thus being on opposite sides can then make random entries profitable is flawed as well. The profitable traders have an edge in their systems. Random entries don't.
Quoting Custos
Random entries don't have an edge. Period. It's like talking against a wall here on ff.
I agree that its not the random entry that has the edge, but an edge can be derived even from a random entry (and there are actually advantages to doing so in some circumstances)

A random entry has every chance of getting on the right or wrong size of a large move in a trending market. When that happens, you can cut the winner short, or let it run to end. You can cut the loser short, or allow it to grow into a very large loss. Trade management is the important variable rather than the entry.

Its also important to distinguish between entries that are random in both timing and market direction from those that are random in just timing, but with a directional bias (or random bias, but with non random triggers).
Quoting Custos
then why don't we program an ea that will do just that and everybody using it will become a forex millionair.
Money management & trade management alone does not make a system profitable. The system itself needs to have an edge, which can then be magnified by good money management.

Random entries don't have an edge. Period. It's like talking against a wall here on ff.

You forgot the vast majority of market participants such as banks and companies just hedging their currency exposure and the vast world-wide exchanges of money to be able to...
So what you are saying is at any point in time, systems having an edge will always be either all long or all short? In other words it is not possible for one profitable system to trigger long and another profitable system to trigger short at the same time?
Quoting Eklavya
So what you are saying is at any point in time, systems having an edge will always be either all long or all short? In other words it is not possible for one profitable system to trigger long and another profitable system to trigger short at the same time?
when did I say that? unbelievable.

okay one time again: you said random entries are good, cause profitable traders are at all times short and long.

I said that just because two profitable traders are long and short, doesn't make random entries a viable edge.

Please show me again where I said that all profitable systems only trigger long or short at the same time.
Quoting Razzle
I agree that its not the random entry that has the edge, but an edge can be derived even from a random entry (and there are actually advantages to doing so in some circumstances)

A random entry has every chance of getting on the right or wrong size of a large move in a trending market. When that happens, you can cut the winner short, or let it run to end. You can cut the loser short, or allow it to grow into a very large loss. Trade management is the important variable rather than the entry.

Its also important to distinguish between entries...
yeah and that is where I disagree. Trade management alone won't make a system profitable. The entries are equally important.
Quoting Scotty B
While there might be reasons to enter long or short at any given time, the reasons are not usually equal. You need to trade with the heavier side. If a hedge fund just sold, you should sell instead of buying with the Audi Corporation. Just saying.
I think the question here is that is a trader employing random entries is doomed to failure or can he turn a profit.

I am trying to argue that such a trader can be profitable.

Will someone do it? Probably not as one would want to maximize profit.
Quoting Custos
when did I say that? unbelievable.

okay one time again: you said random entries are good, cause profitable traders are at all times short and long.

I said that just because two profitable traders are long and short, doesn't make random entries a viable edge.

Please show me again where I said that all profitable systems only trigger long or short at the same time.
Okay. So we both agree that there can be two profitable systems, one of which can trigger long and another can trigger short.

Now, what I do is I toss a coin. If it is heads, I go long and follow the rules of the profitable system triggering long, and if it is tails, i go short and follow the rules of the profitable system triggering short.

Now I have a RANDOM ENTRY that can turn a profit.
Quoting Eklavya
Okay. So we both agree that there can be two profitable systems, one of which can trigger long and another can trigger short.

Now, what I do is I toss a coin. If it is heads, I go long and follow the rules of the profitable system triggering long, and if it is tails, i go short and follow the rules of the profitable system triggering short.

Now I have a RANDOM ENTRY that can turn a profit.
okay man, then please do that.
  • Post #58
  • Quote
  • Edited 12:18pm Dec 18, 2011 12:07pm | Edited 12:18pm
Quoting Custos
okay man, then please do that.
No. Coin toss never helps. Never. The right way is to watch all vehicles stopping at a traffic light. If the last digit of the registration number of first vehicle that stops at the signal is an Odd number, enter short. If its an Even number enter long. For every ploice or fire brigade vehicle.................. well, they normally don't stop at a light.
The Thief of Wall Street
Quoting Chicky
No. Coin toss never helps. Never. The right way is to watch all vehicles stopping at a traffic light. If the last digit of the registration number of first vehicle that stops at the signal is an Odd number, enter short. If its an Even number enter long. For every ploice or fire brigade vehicle.................. well, they normally don't stop at a light.
How about...all systems work providing the trader knows when and when not to use it.

All wins and losses have a common denominator. Knowing these will keep you out when it isnt favourable.

Nice theory isnt it?
0 traders viewing now