5 years and still no strategy | Page 65 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

WRB/MC strategy:Coder needed to semi-automate my strategy 11 replies

strategy builder/tester for new scalping strategy 1 reply

Is opposite of losing strategy a winning strategy? 21 replies

After years in trading, do you still believe in get-rich-fast BS? 98 replies

Fozzy - still using after all these years 2 replies

Attachments: 5 years and still no strategy
Exit Attachments

5 years and still no strategy

  • #1,281
  • May 11, 2017 6:47pm May 11, 2017 6:47pm
Quoting VEEFX
{quote} so two question about your journey that got you here at this stage in your trading career.

1) are your exits unique to each strategy?
My exits reflect the nature of the broad class of strategy the system fits into. Momentum systems apply the 'let profits run' principle and exits use a trailing mechanism. Countertrend (mean reversion systems) apply a profit target principle at statistically significant points during the trade duration (eg. at or around the mean).

What I am focusing on now is specifically looking for un-correlated pairs of systems. For example with a momentum strategy offerring a 30% win rate and say a 2:1 reward to risk, there tends to be a counter-trend solution (as an asymmetric option of that strategy) that can be found with say a 60% win rate and a 1:1 reward to risk. If both systems over the long term offer positive expectancy then as a blended pair, they produce stunning risk-reward profiles. Effectively a hedged pair of systems that are extremely robust with positive expectancy. If I find these pairs I can leverage them up with position sizing. Given the fact that more and more instruments are correlated these days, system design is a way around this obstacle.


Quoting VEEFX
{quote} 2) if yes, how did you gain confidence in come up with a combination of strategies? via spreadsheet / offline analysis and then forward testing in demo with a lot of trail and error, permutation and combination until you found the right spread of strategies that work together?
I used to do all that....but now I simply logically define conditions or constraints to a strategy and then test them out. When I happen across a system that is profitable, I then start budding off similar systems that must also be applicable to a range of timeframes. If not....then into the system graveyard. It is the class of system I am looking for across a variety of timeframes as opposed to a single system in itself if that makes sense.


Quoting VEEFX
{quote3) if #2 is a no, what specific tool do you use for speed up your backtesting to come up with those great charts?
I had to go down the 'design my own system tester route' as there wasn't any offerings in the market that do what I want them to do. All my testing is in the Excel/MS Access environment with a bit of VBA to speed up the process.
  • #1,282
  • Edited 10:01pm May 11, 2017 7:05pm | Edited 10:01pm
Quoting VEEFX
{quote} Thanks for your time C. One more question: Have you explored machine learning capabilities? Example: if a strategy allocation is proving to be profitable across whatever sampling of live trading, then alter the allocations more towards those strategies in real time. I do this but only offline on weekends/monthend/quarterend and then tweak the parameters in realtime based on what my stats reveal. i.e. pick the winners, hardcode the winning combinations in a list and update this list in real time based on historical stats/trading results.
Nice principles at work there V. It makes sense to me....but I am out of my depth in regards to machine learning.

In my opinion there are two broad ways to tackle the markets.
1. Adaptive Specialization where you use a select number of strategies to capitalise on a specific market's characteristics - This is where I would classify your machine learning system. I would also class FA and other specialist strategies such as pattern recognition and TA in this class.
2. Generic application - Apply a broad market principle with diversification. This is the land I inhabit. Bust through the markets using a small edge ....eg. the casino principle.
  • #1,283
  • Edited 7:56pm May 11, 2017 7:29pm | Edited 7:56pm
Quoting VEEFX
{quote} Speaking of Casino, I have always wondered why FX has not arrived in Vegas Casinos? The only reason I can imply is... there is NO EDGE that they can leverage in FX ! If they cannot, then there is no strategy on this planet to consistently extract pips lol
Hehehe.

The name of the game is different but the broad principle is the same.

Casino
a) I have a roulette wheel that offers a very small house edge. One class of system with a characteristic return distribution.
b) I have a craps table that offers a very small house edge. Next class of system with a characteristic return distribution which is uncorrelated with system 1.
c)i have a black jack table that offers a very small house edge. Next class of system with a characteristic return distribution which is uncorrelated with system 1 and 2.
d) I have a one armed bandit......
e) etc...etc etc for as far as you can go.

List the vehicle on an exchange.....pile in investment funds....scale up with 20-30 of each in each category.........relax and bring home the bacon.

If you step into the land of offering FX....then different compliance costs and regulation would wipe the contributions from your other systems. No need to compromise a good model.

Prop Trading firm
a) I have a class of traders that delivers a very small house edge. one class of system with a characteristic return distribution.
b) I have a 2nd class of traders that also offers a very small house edge. Next class of system with a characteristic return distribution which is uncorrelated with system 1.
c)i have a 3rd class of traders that offers a very small house edge.......Next class of system with a characteristic return distribution which is uncorrelated with system 1 and 2.
d) I have a 4th class of traders ......
e) etc...etc etc for as far as you can go.

List the vehicle on an exchange.....pile in investment funds....scale up with 20-30 of each in each category.........relax and bring home the bacon.

If you step into the land of offering casino games....then different compliance costs and regulation would wipe the contributions from your other systems. No need to compromise a good model.
  • #1,284
  • May 11, 2017 7:32pm May 11, 2017 7:32pm
Quoting MaxenshteinD
What if our trading platforms only showed the following ? Could we still trade it? How? {image}
I want to say of course you can and some may even benefit from not having charts visible. If I was challenged to just use open/close and know previous days data, I'd start buy only trading Long if the price is > then open and Short if price is < than opening.
  • #1,285
  • May 11, 2017 8:08pm May 11, 2017 8:08pm
Quoting RondaRousey
The posts that gets the most likes are the ones you should ignore if you want to be a profitable trader. So much garbage written on this thread. It makes me wonder if FF is really for losing traders setup for losing traders to be losing traders.
Thats exactly what it is. NOT one person has shown a consistent high probability method. Not ONE person on FF has ever shown any ounce of the real game of market makers. A lot of very intelligent people but they are too smart in a way. Markets a re a simple game of buying and seling in line with news and world environmwnt. Thats it. To think 61.8 line or stochs or gann or macds can help you trade IN ANY MARKET is insane. EVERYDAY money is moved. WE retailers (with limited resources) can make money with banks but not everyday. Why? Big news/world environments for which they have access to we dont. But thats fine plenty of days where we can trade with him where its so simple.
  • #1,286
  • May 11, 2017 8:13pm May 11, 2017 8:13pm
With my B.Sc and MBA, I'm humbled by some of the posters here and feel like I haven't graduated Kindergarten yet, lol. As my own method serves me well, you guys are putting doubts in my mind about the efficacy of what i've been doing for years,lol. I don't think i could ever conceive or process these market analytics you speak of, as i'm just not wired this way. My hats are off to you, and if you've found something solid, i hope it serves you well. Although i very much respect the work of Mzvega, my only comment is it's been stated that unless the data is analyzed or looked at in 1 particular way, it's impossible to be successful over time. I will never claim to be the master of all traders but I've carved out a method over time that performs consistently, without this sort of number dissection. All I'm saying is it is possible to find an edge that isn't math based, as I use it everyday.... cheers!
  • #1,287
  • May 11, 2017 8:16pm May 11, 2017 8:16pm
Quoting nealpek
With my B.Sc and MBA, I'm humbled by some of the posters here and feel like I haven't graduated Kindergarten yet, lol. As my own method serves me well, you guys are putting doubts in my mind about the efficacy of what i've been doing for years,lol. I don't think i could ever conceive or process these market analytics you speak of, as i'm just not wired this way. My hats are off to you, and if you've found something solid, i hope it serves you well. Although i very much respect the work of Mzvega, my only comment is it's been stated that unless...
Retail forex trading is NOT mathematical. This is where people go wrong. Pips and Price Point? Anyone who mentions these things have no idea and you should run. "I MADE 55 pips today!!" WTF what does that mean?? It means nothing. Probably risked 150 pips to do so.
  • #1,288
  • Edited 8:59pm May 11, 2017 8:39pm | Edited 8:59pm
Quoting nealpek
my only comment is it's been stated that unless the data is analyzed or looked at in 1 particular way, it's impossible to be successful over time.
That's the key requirement IMO. Systematic application of the same rules is the way to gauge how well your strategy is performing over time. Unfortunately if you go down the discretionary route, it is so hard to tell when your system no longer stacks up against current market conditions.

We have a market of participants buying and selling at different times with different volumes. Price reflects that participant behaviour. We then apply a system with pre-defined rules that succeeds when those rules relate to market conditions and fails when they don't. By applying a system we constrain the way we harvest participant behaviour. To gauge how well your system stacks up against market behaviour, you need to consistently apply those rules until you have a sufficient sample size of trades to distinguish between a random result and a non-random result in terms of a reduced confidence interval.

Exactly like flipping a biased coin. You will never detect the bias until a sufficient sample size is reached where you can conclude with a higher degree of certainty only that the coin is in fact biased.

Nothing is certain in a complex system, but there are broad features arising from collective mechanics that you extract an edge from. We cannot predict (trade) the individual motions of a water molecule in a cloud with any degree of certainty, but we can predict (trade) the gross features of the cloud itself with more precision such as it's elevation, it's overall mass, its temperature and directional motion. Simply apply the broad features of thermodynamics which deals with system complexity, namely statistical mechanics. This is the path a quant takes.....but there are other paths as well. That's the great feature of complex systems.

Of course you don't need to be a mathematician or physicist to do this as they only apply models that approximate the behaviour of a population and many of those models get surpassed by improved models that predict the behaviour of the system better as more data comes to fruition (the domain of the sample expands). Common sense with an understanding of participant behaviour is enough to harvest opportunity. We actually all apply this principle of sample projecting when we predict a future of a complex system as we apply a sample set of experiences (our thoughts) to the way we deal with uncertainty of the future.....but just may not be aware of how we actually reached our conclusions.
  • #1,289
  • May 11, 2017 8:58pm May 11, 2017 8:58pm
Quoting VEEFX
{quote} "risk must be smaller than reward over a large number of trades" - I think we need to dissect this statement further.The traditional risk x % or pips to yield y in reward in flawed imo. It is this mantra that causes the delusion of $100 to 10M in 2 weeks type compounding fantasy and the associated greed. I will admit... I have FAILED a 100% of the time in every trade every time in past 10 yrs when I try to go razor sharp based on R:R. I also feel risk based single position sizing is fundamentally flawed. I have always felt idiotic to define...
I failed to mention the compounding aspect. No. Do not compound. Risk the same amount over every trade. My belief is that compounding skews the traders equation. If you compound prematurely, your trade risks more than it rewards relative to your last trade. You only move up if the next trade you take is in line with the traders equation relative to your last trade.
If you are good at something, never do it for free--Joker
  • #1,290
  • Edited 11:10pm May 11, 2017 10:49pm | Edited 11:10pm
1
  • #1,291
  • May 11, 2017 11:02pm May 11, 2017 11:02pm
1
  • #1,292
  • May 11, 2017 11:09pm May 11, 2017 11:09pm
66 pages and the comedy still continues.
  • #1,293
  • Edited May 12, 2017 12:30am May 11, 2017 11:19pm | Edited May 12, 2017 12:30am
Quoting nealpek
Quoting nealpek my only comment is it's been stated that unless the data is analyzed or looked at in 1 particular way, it's impossible to be successful over time. That's the key requirement IMO. Copernicus, Every trade I take is the same setup. I understand your numbers in a theoretical way, but i believe this applies to a natural situation with no influence or bias built in, like movement of the planets. Left to themselves with no outside interference, mathematical models...
Hi N

I come from the point of view that the inbuilt biases in humans make them see things that might not convey actual meaning. This inbuilt bias lies with the way we have coped as a species on this planet and is an expression of our evolution on this planet. One of the symptoms for example is our ability to react to illusions. It is a survival mechanism we have acquired through dealing with an uncertain environment. It is useful for survival on this planet but not necessarily useful for trading the markets.

For example when I see the chart below, my immediate instinct is that I can project the future state of this price series as being bullish. That's my heuristic brain kicking in telling me how to respond to this pattern in projecting the future.

Attached Image (click to enlarge)
Click to Enlarge Name: Random trend.JPG Size: 28 KB


The bottom line however is that this is a purely random generated price series which I made myself. By definition you simply have no predictive power when it comes to a random price series. If I attempt to trade a random price series using a system configured to respond to this illusion, I may be profitable or not in the short term (even exceedingly profitable or unprofitable), but over the longer term as the sample size of trades increase and include transaction costs your returns will deteriorate with a high confidence interval.

Here is another chart. I assume that this also has a bullish future.

Attached Image (click to enlarge)
Click to Enlarge Name: Random trend with drift.JPG Size: 32 KB


In this case I would be right as this was a random series including a very small 2% serial correlation which acts as a bullish drift on the otherwise random price series. It is the drift (deviation from randomness) that's important for projections and not the surrounding randomness which has no future value. If I trade this price series, I still face a significant degree of randomness, but it is the slight edge provided by the non-random component of price data (provided it exceeds the transaction costs) that allows me to achieve a long term profitable outcome with a higher degree of certainty than the random series. No guarantees...just probabilities.

Now putting both charts together...this is the difference.

Attached Image (click to enlarge)
Click to Enlarge Name: Capture.JPG Size: 37 KB


Now we are getting somewhere. If I had this info available to me I can take advantage of it and project where I expect price to be in the future (within acceptable limits) and obtain a profitable edge by adopting a system that capitalizes on this bias. Unfortunately we simply cannot use our brains to do this.

The problem in using our brains to assess price is that those inbuilt biases are actually deciding how we play out these future opportunities.

The only way I can attempt to discriminate between a random series and random series with minimal bias is to statistically interrogate the data. My brain is simply not equipped with the tools to do this. Our brains have been very effective in dealing with the surviving in this complex system (namely our planet earth), but are not equipped very well to project a future outcome in another complex system (being these financial markets). There are rare instances when human behaviour kicks in to market behaviour (eg. in times of crisis) but to allow for coordinated participant behaviour (eg. fight or flight) these are the exceptions as opposed to the rule.

In general we have so many different participants and reasons for trading that to faithfully predict this complex morass of activity is simply intractable. This is what we mean in terms of statistical randomness. The overlay of all participant behaviour in terms of the impact on short term price action is impossible to fathom. We simply do not have access to perfect information. We can gain info that other general participants don't have (eg. insider info etc.) that allows us to project a future outcome....but unless you have access to this info.....the chances of successfully navigating the next price move over a specified duration is virtually impossible.

We therefore need to look at other ways of skinning the cat and that's where statistics come in...or you could go to the other camp of FA and look at the drivers of longer term underlying price behaviour. You need to look at price data over an extended price series to be able to separate what is random and what may be a serial correlation.

At the level of a tick where granularity is as fine as you can go you will not be able to deduce whether a price movement has any significant impact on longer term price direction. Yes...there will be a reason for why price moved in terms of tick activity but unless you are trading at that level of granularity it will do you no good as a trader.
For longer term traders who enter and exit over longer time durations, we need an appreciation of the consolidated picture of that price data and where any exploitable opportunity lies to project future direction.

It is only when you accumulate tick data and consolidate it over an extended duration that a bias in participant behaviour can be detected. With this info you can take an action now, and with a certain confidence interval expect future price at a future date to be in the ball park of x assuming of course that the serial correlation in the price history translates to a bias in the price series.

.....I will stop there. Hopefully you get the gist of where my head is at with this world of trading :-)
  • #1,294
  • May 11, 2017 11:40pm May 11, 2017 11:40pm
Quoting jmn5611
{quote} You are not wrong in what you say. I would add however that MA's do not tell the complete picture. The point being made is that there is no real certainty in the markets. Trends are essentially illusions, they exist because we think they do. This is why markets oscillate. The amplitude and wave peaks and troughs are random. The drivers behind this behavior is not. We have all learned the same TA, FA, and PA. In effect, these things become self fulfilling prophesies. You can flip a coin to take a position, but one must still adhere to the...
We both understand that I wasn't wrong. We also know I wasn't exactly 'right'. The logic is simple; the market exists in a superposition at all times between trending and ranging. It can neither trend forever nor can it range forever(It is inevitable that it will do both until the end of human civilization). The longer that it exists in one state the more likely that it will soon deviate from that state. There are market forces/participants ensuring that it happens because they, unlike the market, will not live as long as the rest of human civilization. They will not be alive to witness the end of all things. The market is inherently unstable and our objective as traders is to attempt to stabilize it (either profitably or unprofitably). That is what we get rewarded for doing. That is our purpose.
"The fun is in the hunt. Not the kill."
  • #1,295
  • May 11, 2017 11:44pm May 11, 2017 11:44pm
Quoting Redeflect
{quote} The logic is simple; the market exists in a superposition at all times between trending and ranging. It can neither trend forever nor can it range forever
That's a good way of looking at it IMO. Live cat trader and dead cat trader at the same time. The way you interrogate the market (interact with it) decides if you survive or not. :-)
  • #1,296
  • May 12, 2017 12:15am May 12, 2017 12:15am
Great posts! Thought provoking.
FF journal: Peaks and Troffs
  • #1,297
  • May 12, 2017 12:48am May 12, 2017 12:48am
Quoting nealpek
With my B.Sc and MBA, I'm humbled by some of the posters here and feel like I haven't graduated Kindergarten yet, lol. As my own method serves me well, you guys are putting doubts in my mind about the efficacy of what i've been doing for years,lol. I don't think i could ever conceive or process these market analytics you speak of, as i'm just not wired this way. My hats are off to you, and if you've found something solid, i hope it serves you well. Although i very much respect the work of Mzvega, my only comment is it's been stated that unless the...
There is no math based edge. The only edge is time based.
"The fun is in the hunt. Not the kill."
  • #1,298
  • Edited 1:05am May 12, 2017 12:51am | Edited 1:05am
Quoting VEEFX
{quote} It took a while to change my mindset from the highlight bold (for me). Market are ALWAYS trending imo. The only thing that changes is the scale and frequency inbetween those trends. I realize it's an irrational view with no objectivity but this shift has changed my perception lol
How do you know the market isn't ALWAYS ranging and the only thing that changes is the scale and frequency between oscillations?

The only thing you can be certain of is uncertainty. If you feel the need to be certain then you'll certainly be a losing trader.
"The fun is in the hunt. Not the kill."
  • #1,299
  • Edited 1:33am May 12, 2017 12:59am | Edited 1:33am
Quoting RondaRousey
{quote} Why do you want to win, to be profitable at trading? you will be despised, envied, harassed and blamed for making money from misery. It's not like sports where you are adored by fans for your talent and skill. Profitable traders are not adored by people, you will have few if any fans. It is most likely that you will go in hiding for your skills and talent in trading. You will spend most of the time worrying if someone has stolen or figured out how you are profitable. So quit trying to win at trading. It is one of the loneliest places to...
You jest. I look forward to such things. If they figure it out then they deserve it. They earned it... and we all benefit from their progress. Maybe not financially but is that what matters? We created the system as a tool. We all benefit from it and none of us can take credit for it's inception. The only thing that makes us worthy of it was our capacity for realizing it's uses.

If they succeed... good for them. We should all remember what it was like before understanding success. It's within their will and not ours to dictate what they will or will not become. I welcome them whether or not they join us.
"The fun is in the hunt. Not the kill."
  • #1,300
  • Edited 2:02am May 12, 2017 1:34am | Edited 2:02am
Quoting nealpek
With my B.Sc and MBA, I'm humbled by some of the posters here and feel like I haven't graduated Kindergarten yet, lol. As my own method serves me well, you guys are putting doubts in my mind about the efficacy of what i've been doing for years,lol. I don't think i could ever conceive or process these market analytics you speak of, as i'm just not wired this way. My hats are off to you, and if you've found something solid, i hope it serves you well. Although i very much respect the work of Mzvega, my only comment is it's been stated that unless the...
thats simple fundamental edges. both, fundamental and statistical edges can be in market for many years or decades. even 12 of my statistical strategies work since much more then 10 years ago near every years about same. so its not uncommon that fundamental edges work very long (often law changes can change something) or good statistical "3-4 step calculations" work many decades too. if you implement, that strategies is changing by changed pattern, then it will work "forever" of course, until the strategy is correct.
1 trader viewing now