What works, and what doesn't

indeed agree

let me save you some time EAs don't work

Agree 100%.

For those disagreeing with these, we need to make a clear distinction between mechanical application of these "methods" and discretionary.
Let's first address chart patterns...technical analysis as they're most popular.
Many of the top schools of economics, both in the US and Europe have run extensive tests and analysis of the core elements of technical analysis principles. Chart patterns, candlestick formations, indicator systems etc.
The results categorically proved that they do not work for long term profitable trading, across multiple asset classes.

But, many claim they use them and are profitable...how can this be? First, a good portion of them are simply lying/deluding themselves..but for the few that aren't..
Well, as Hanover correctly pointed out, that a trader will make decisions based on more than a single (or multiple) set of data. So, whatever aspect of TA is being used, the trader will apply his/her bias to that information, to dictate where and when he/she trades.
So...is that trader then really applying a "system" based solely on trendlines, indicators etc?
No...they are at best using the indicator for confirmation bias in order to trade what they want...discretionary.

That's why these allegedly profitable systems that Hanover took the time to code did not, and will never work. Mechanical application; that is trading all signals based solely on head and shoulders, shooting star, or whatever bulls*#t term we see in all the trusted resources will not work long term.

Trendlines: Well these are totally subjective and decided upon the by the trader, and the price feed you have. Cannot be applied uniformly for profits.

Multiple indicators: Taking one flawed concept, and adding a bunch more flawed concepts....does this logically sound like it will make something work?

Over complicated systems: Same as above.

Trading news: Well, this is the equivalent of a little fish swimming in the middle of a shark feeding frenzy hoping to get a morsel of floating food. The most likely outcome is getting eaten. Retail Fx traders don't have the tools nor platforms to effectively trade news. It's purely a liquidity drive for the institutionals.

Instead of reading the "trusted" books and resources, it would be better to study up on efficient market hypothesis, random walk and levy flight.
The only edge to be found in retail trading is to exploit temporary inefficiencies in the market pricing.
As hanover alluded to, the market runs by having willing buyers for each seller...sometimes there is an event; news, geopolitical turmoil, natural disaster...whatever...but it creates an extreme situation.
The only thing we can say for sure in the Fx market is that the price just can't keep going. Sooner or later, there will always be a pullback...a point where the market acknowledges the price has moved far enough to factor in the new information, and opportunities exist for the institutions to take new positions.

The trick is picking those moments, and applying a set of parameters.

Bloodpoodle- For what it's worth I think your charts, your methods are good. Pretty much how I approach the market now and how I found success.
After years for looking for magic formulas, I realized none existed, and all we can do is trade what we see.

All the best

I sense that "discretionary" would be the same as "mechanical", just with a few more parameters.

My assessment is both the O/P and T 34 have right potentials to become a accomplished forex trader due to the fact that they are both at a stage whereby they consciously are aware and know what more they need to know in order to be successful in this business. The next level of competency of knowing what they need to know is the most challenging and crucial phase of this long and difficult journey. I wish to chip in here and hope to help them focus and stay on the right path becos this is where the majority of the potentially accomplished trader fail and most either give up or switch to teaching others how to trade instead of staying on the harder path of becoming a successful trader themselves. The only 2 additional things they need to know is the breached price levels and the holding levels. Now the O/P have mentioned that those indicators that dont work all the times is precisely due to the fact that even price may have breached the resistant or support levels it fails to stop at the hold levels which resulted in fakeout 75% of the time. Remember that all breached levels will be breached one time or another and only if it held on the holds levels can the PA either carry thru the momentum or fall back to its original path if the hold levels fail. In all the majors pairs the rate of change is normally limited to 3-5 sigmas of breached and hold levels in normal peak to trough cycles and 1-2 sigma for ranging market condition but it will need more than 7 sigmas to force a M/T trend change. All crosses price levels are contingent upon the ROC of all the majors vis a vis the reserve currency USD. The only remaining 5% of unexpected downside is CB aggressive interventions or other geopolitical issues.

This will be my first and last post in this thread and I wish all the best to all....

GS.

1. Trendlines do not work if you do not know how to draw them, you are right about that.

2. There is no single system that works in all markets.

3. Good point.

4. Google, good news, bad action. That might be interesting reading for you.

Good point about spikes.

I must disagree with your statements about trend-lines and patterns. Together with supply demand these are the tools I found most helpful to my trading. You need to understand them and their rules first and like everything use them in context.

As recently as 5 years ago I probably would have taken exception with this statement. I didn't realize it at the time but after creating and testing thousands of trading strategies and evaluating them for profitability and consistency with high standards, I was Don Quixote tilting at windmills. Then around 2010 I ran into a thread that altered my trajectory that discussed Systematic Trading. One of the basic points made in that thread was that without diversification, individual strategies had little/no chance to work in today's market environment. This thread was a real eye-opener for me!

So I went about creating lots of strategies and attempting to combine them together in an effort to improve both profitability and consistency. After learning a lot about the wrong way to create models I started learning more statistics and the correct way to create models that didn't lead to overfitting. It took a long time because model creation the right way takes time and isn't particularly easy, plus I had to do everything myself and had a lot to learn. In retrospect I probably focused too much time on the strategy development part of things, which isn't nearly as important as how you combine your strategies.

After finishing all these efforts I still found that the strategies developed didn't generalize to new data particularly well, either individually or in aggregate as I had hoped. I was sure I had not overfit the strategies so I knew I was missing something important. After pondering on the process (a lot) it was clear that the way forward was to include an adaptive component in each system, that allowed the static set of rules to change over time according to how well those rules were doing. Instead of a sample size of 1 (static system rules) a dynamic/adaptive component altered the rules thousands of times, and validated the adaptive process over time. This idea was inspired by the Machine Learning With algoTraderJo thread where day by day reoptimization is discussed. So in summary, the 3 concepts that I've learned (so far) for successful* automated system development are:

1) Trading systems developed and selected using sound statistical model building principles (not overfit)
2) Diversification to increase reward and reduce risk by combining/trading many low/uncorrelated strategies together.
3) An adaptive component in each system to allow the fixed system rules to change over time and to thus generalize better.

(*note: I said successful development, not successful trading as I won't know for probably over a year whether there needs to be a 4th concept).

Everything works, and everything doesn't...you just need to know how to use them

Yes and no I guess.
My definition of discretionary is something that can be quantified by a set of check boxes.
So, conditions met=trade.
Basically, the type of things people go looking for on forums such as this, and then wonder why it doesn't work for them.

The reason is, they can only follow the rules...any changes/adaptations the designer takes can and usually do make the difference between it winning and losing (if in fact it ever was a winning method).

The way I trade can't be quantified completely with check boxes. I have a "signal" I look for, but it's only the guide...previous PA and upcoming events mean more than the signal itself.
Sometimes I skip signals, other times I will take weak or invisible signals. I trade a condition, not a single formation.
But that's me..

Genghistar- Thanks for your input, but I have a feeling we are on different wavelengths. Talk of Sigma levels and failing to hold of this that and the other tells me you are trying to apply statistical modelling and probability theory to trading. I respectfully disagree with this concept.

Support/resistance, fibo, sigma etc etc levels only exist in the minds of retail forex traders. When Deutsche Bank decides to sell 5 yards of Euros for a US client, then the Euro will fall...regardless of what level it's sitting at. Talk of fake outs implies that someone is trying to deliberately mislead us...why would they need to?

When Deutsche sells off, the inevitable consequences is the euro falling, because selling volume exceeds buying at that moment. That movement, price dropping, will slow and stop as the volume of sells decreases compared to the volume of buys. In the process, one of our "levels" may have been breached, but then comes back. Is that fake? No, it's a simple transaction. There need be no subversion, no faking anything.
Banks can and will use news to provide increased liquidity, so that their fills do not force the market into a move which potentially adversely affects the price they get. They can do the same in zones where price is not moving directionally: a range. Of course this way they need to gently feed orders into the market, so as not to deviate too far from the price they have promised for exchange for whichever customer they are servicing with that transaction. If they promised their customer a rate of US$1.12 per Euro, they will make more money if they can sell at or higher than that level, and less if the price falls. If they sell at $1.13, great, they can pocket the difference!
Their goal when doing this, is to get the closest possible price(or better) to the current one, to get their fill completed so they profit more from their spread. Nothing more!

This...market making, providing genuine exchange for their commercial clients is what really drives the forex market. Retail forex brokers allow us to watch this interbank market, but never participate truly. We place off track bets on what we see, thinking there is some grand scheme. There isn't. There is only the day to day business of commerce and global transactions that gives rise to the movement of a currency's value with respect to another.

For this reason, with ever changing variable sets, it stands to reason that no edge can be found using levels, or mathematics in the conventional sense. The market is always adapting, and HFT has changed the game, exploiting any perceived inefficiencies very very quickly, to the point where edge is extremely hard to find.
Again..the only thing we can definitely say, is that the direction of any currency cannot be continued forever.
How we take advantage of that, will likely dictate our success, or failure in this industry.

Thanks, and all the best to you also.

I certainly agree with this post.

There are subtle variations to this dependent on available capital constraints such as the extent of instument diversification and methods of diversification but the principles highlighted here are invaluable for those seeking to truly quantify those trading programmes with an 'edge'. Without re-inventing the wheel, there is an excellent source of literature available today that assists in expounding on these principles. You can make it so much easier on yourself if you invest some time in researching the principles surrounding the points highlighted by FXEZ above and then once armed with the knowledge, determining ways to capitalise on them.

For those interested in looking into these ideas, sources of valuable info can be found in the following:

1. Building Winning Algorithmic Trading Systems - Kevin Davey;
2. Trading Systems - A New Approach to System Development and Portfolio Optimisation - Jaekle & Tomasini;
3. Quantitative Investing - Strategies to Exploit Stock market Anomalies for All Investors - Piard;
4. Beyond Technical Analysis: How to Develop and Implement a Winning Trading System, 2nd Edition - Chande; and
5. Following the Trend - Diversified Managed Futures Trading - Clenow.

Yes!

Market Structure, Supply/Demand, Support/Resistance, and many Trendlines.

Trendlines really depends on the timeframe and how it is drawn.

Trendlines are tradeable, just need to re-draw many lines.

You many only see like 3-4 trendlines on your NZD/JPY Daily chart, but your chart looks like is over a year or a few years.
Meaning need more trendlines drawn to make it tradeable, rather than waiting for a month or a year for a setup.

Below are your charts, redrawn with my RED trendlines.

I think this is true.

I would say something even more dangerous: that high levels of (mathematical) intelligence may be actually counterproductive in this business. Smart people are used to success, and feel that they should achieve their goals easy in the market, as it consists of data/numbers. They are usually in for a rude awakening. The true enemy of the profitability is not the market, it is oneself, and it takes a while to focus on the real problem, particularly when (our) autistic mind is focused on the numbers...

k

Slow grind only to be shorted if they first show you weakness and make a lower low.... never be the first to short...

The odds are much higher in your favor Shorting a fast move up... or buying a fast move down... then getting caught in grinding market that may continue against you.

That is one of the most bizarre things I've read in a while. Trendlines are an abstract concept, how can they not work?

It's like saying this picture of a pigeon doesn't work.

Todays trade.. Thin run up = weak.. Hit area of liquidty which was cap from NFP and came right back down.

odds were in my favor.. quick run up was weak.. and it was overbought.. Sell high.. don't buy high.

Interaction of bulls and bears is the essential concept based on which price moves up/down; you are using this market concept to analyze a price chart in order to find entry/exit points.
How did you get 75% success rate for this analysis?

Market ranges approximately 75% of the time on average... and trends 25%.. So I'm betting on a ranging market because odds are so in favor of a range.. Take a look at your charts.. Most of the time market pulls back into range after a spike out of the range.