S/R and TL on Randomized Price Movement | Page 2 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Price movement as a function of Supply and Demand 43 replies

Spot Fx Trades based on futures volume and price movement 22 replies

How to differentiate pro and amateur price movement? 6 replies

Random Price Movement 177 replies

GBP/USD daily pip movement vs. EUR/USD daily pip movement 3 replies

Attachments: S/R and TL on Randomized Price Movement
Exit Attachments

S/R and TL on Randomized Price Movement

  • #21
  • Mar 25, 2010 4:15pm Mar 25, 2010 4:15pm
While they look similar, a trend line break on many (not just one) random charts should end up being a loser over the long run, factoring in spreads etc. Now a trend line break on many real charts... over time should be a winner if we believe that the market participants will act upon it, generating higher/lower prices. Unlike randomly generated prices at every point in time which are random.

In hindsight, the random chart looks tradeable, but let me ask you... if you know ahead of time that prices will be random... would you trade a random generated chart breakout without seeing the whole chart?
Of course not, because you know the randomness is "hard-wired". Over time with spread you will lose.

In real trading, we hope to capitalize on repeated patterns that we hopefully perceive early enough when they are somewhat self-fulfilling prophecies because of current beliefs of support, resistance, and trends.
  • #22
  • Edited 8:19pm Mar 25, 2010 7:49pm | Edited 8:19pm
Here's my view, for whatever it's worth - take it or leave it.

IF patterns (S/R etc) on a real price chart reflect orderflow that creates self-fulfilling prophecy, then they have predictive value. Hence they can be potentially exploited for profit. Likewise, IF prices trend far enough, often enough (i.e. beyond what would be 'randomly'), then an exit strategy that cuts losers, and rides winners, has a potential edge.

A randomly generated 'chart' will exhibit trends, patterns, etc, but these are nothing more than the result of 'random coincidence', and offer no more than a 50/50 probability of future direction. Hence there is NO system (entries or exits) capable of attaining long term profit, under these conditions; at 50/50 probability, all wins and losses are necessarily fortuitous. Nobody can gain a statistical edge from a random game, much less one where costs make it negative expectancy. Roulette is a valid analogy.

[EDIT] It is easy to be fooled by randomness, imagining some kind of pattern exists when in reality there isn't one. Taleb's book Fooled by Randomness is a worthwhile read, IMO.
  • #23
  • Mar 25, 2010 9:08pm Mar 25, 2010 9:08pm
Quoting Sauron
Sorry, I didn't download the excel sheet and I don't have time to read the code. But to display real bars you must start from a value, generate 1000 offsets, calculate OHLC and then start from the last close. This is how the market works, if you want to simulate it you still have something to do.
I haven't simulated the OHLC, close price is also analyticable for many people.

Quoting Sauron
You could generate another random value to calculate the offset in pips. Instead of -5/+5 you'd have -1/+2/+1/-3/-1 etc. It models the market better.
You haven't understood what I said a random value between +5 and -5. Read the xls when you have time.

Quoting Sauron
That's not true, you have 1000 random values which I must repeat, are NOISE. You need 1000x1000 at least.
Not sure how you work out your standard.
  • #24
  • Mar 25, 2010 9:09pm Mar 25, 2010 9:09pm
Quoting dr_zeus
OHLC chart for Sauron.

Don't get mad Sauron, disagreement is healthy. Sorry if I came on too strong. This chart has 10 tics per period. formula for price is RANDBETWEEN (RANDBETWEEN(-12,-5), RANDBETWEEN(5,12)) to make it extra randomy :-)

Once again, TA patterns become clear. MAs indicate entries, trendlines form and break, pin bars and inside bars give good signals.

If you do this yourself you'll see all kinds of charts form - ranging periods, large trends. I've tried to pick charts that have a mix.

Next up I'll try to write a trading...
Thanks for working out the OHLC. Can you give me your xls. file?

Cheers,
  • #25
  • Mar 25, 2010 9:10pm Mar 25, 2010 9:10pm
Hanover,

That sounds like a good book, at least I like the title. I'll see if I can pick it up. I've long known there is a belief called "random walk" that says the prices are random and any predictive patterns are illusory.


If the random chart looks like a real chart, with S/R and other patterns, what makes you think the real chart is any more or less tradeable?


The fact is that forex is a losing gaming, it's not even breakeven. So over time the sum number of players will lose. The whole game at forexfactory and other sites is an attempt to gain a statistical edge, and I think the group as a whole as not been successful.

The fact is, every system will work now and then, because the random nature of the trades will work in its favor. Once the system hits a down patch, people start to lose and abandon the system.

Money Management is the only statistical edge.

As for the rest of the comments regarding self fulfilling prophecy and group movement, I am no longer convinced that this occurs. Large options on 1.xx00 and 1.xx50 numbers may attract price to them. Fundamental news certainly causes long term trends. These may be the only truly predictive events in forex.

On the other hand...

It could be the S/R, trendlines, and M/A's provide real information about a random map and can be tradeable regardless of if the instrument is real or fake.

Finally....

We know that forex has busy periods and quiet periods, despite being 24/6. We can use the quiet squeeze to our advantage and trade breakouts since even though they are random, they are moving in larger amounts so getting in early gives us a better chance of the large move going our direction.
  • #26
  • Mar 25, 2010 9:11pm Mar 25, 2010 9:11pm
Quoting hanover
Here's my view, for whatever it's worth - take it or leave it.

IF patterns (S/R etc) on a real price chart reflect orderflow that creates self-fulfilling prophecy, then they have predictive value. Hence they can be potentially exploited for profit. Likewise, IF prices trend far enough, often enough (i.e. beyond what would be 'randomly'), then an exit strategy that cuts losers, and rides winners, has a potential edge.

A randomly generated 'chart' will exhibit trends, patterns, etc, but these are nothing...
Thanks Hanover,

The strangest thing here is it seems that "patterns" can be seen from a random chart.

cheers,
  • #27
  • Mar 25, 2010 9:55pm Mar 25, 2010 9:55pm
Quoting dr_zeus
As for the rest of the comments regarding self fulfilling prophecy and group movement, I am no longer convinced that this occurs. Large options on 1.xx00 and 1.xx50 numbers may attract price to them. Fundamental news certainly causes long term trends. These may be the only truly predictive events in forex.
All I can say is that I know where my edge lies. I've explained the basics of my trading strategy elsewhere on the forum, and thus far I've found that it's more than enough to beat the negative expectancy imposed by the spread.

Quoting dr_zeus
The fact is, every system will work now and then, because the random nature of the trades will work in its favor. Once the system hits a down patch, people start to lose and abandon the system.
Money Management is the only statistical edge.
The randomness topic has been done to death many times before. I will always contend that only entries and exits based around non-random aspects of market behavior can provide a real edge, and that the most MM can ever achieve is to redistribute wins and losses.

Sorry if this sounds condescending, but I'm too lazy to debate the same old issues by retyping material that I've posted many times before. If people want to try to prove that a coin toss method can be beaten by using MM, well, that's their prerogative.
  • #28
  • Mar 25, 2010 10:35pm Mar 25, 2010 10:35pm
I made my OHLC with a simpler randomizing method. Yes, I can mark the chart with technical tools too.

Seems a lot of profitable Pinbars too. Psyho level works too. PPZ confluence works too!

Quoting dr_zeus
OHLC chart for Sauron.

Don't get mad Sauron, disagreement is healthy. Sorry if I came on too strong. This chart has 10 tics per period. formula for price is RANDBETWEEN (RANDBETWEEN(-12,-5), RANDBETWEEN(5,12)) to make it extra randomy :-)

Once again, TA patterns become clear. MAs indicate entries, trendlines form and break, pin bars and inside bars give good signals.

If you do this yourself you'll see all kinds of charts form - ranging periods, large trends. I've tried to pick charts that have a mix.

Next up I'll try to write a trading...
Attached Image (click to enlarge)
Click to Enlarge Name: RandSim04a.GIF Size: 12 KB
  • #29
  • Mar 25, 2010 10:50pm Mar 25, 2010 10:50pm
Seems that it is possible to use technical methods to win in these randomly generated charts. Sound crazy?? Indeed it is.
Attached Image (click to enlarge)
Click to Enlarge Name: RandSim04b.GIF Size: 14 KB
  • #30
  • Mar 25, 2010 11:36pm Mar 25, 2010 11:36pm
This is the OHLC version randomized chart in Excel. For each bar, I have randomized 10 movements between +5 and -5 to get the High and Low (Sheet3). Interested parties can download it, randomize and examine themselves.
Attached File(s)
File Type: xls RandSim.xls   879 KB | 450 downloads
  • #31
  • Mar 25, 2010 11:37pm Mar 25, 2010 11:37pm
Quoting kk007
This is the OHLC version randomized chart in Excel. For each bar, I have randomized 10 movements between +5 and -5 to get the High and Low. Interested parties can download it, randomize and examine themselves.
You can see that on your randomly generated charts, those price action such as Pinbar, Outside bar and Inside Bar can help you to be profitable.
  • #32
  • Mar 25, 2010 11:39pm Mar 25, 2010 11:39pm
Quoting kk007
Seems that it is possible to use technical methods to win in these randomly generated charts. Sound crazy?? Indeed it is.
Here's the implication of what you're saying: if you could find a casino that allowed you to use a palmtop PC, you could use the roulette wheel to generate random candles based on reds (uptick) and blacks (downtick) spun. Then you could apply your pinbars, TLs, PPZs and psych levels, and make a handsome living milking money from the roulette tables.

Hmmm..........
  • #33
  • Mar 25, 2010 11:48pm Mar 25, 2010 11:48pm
Quoting hanover
Here's the implication of what you're saying: if you could find a casino that allowed you to use a palmtop PC, you could use the roulette wheel to generate random candles based on reds (uptick) and blacks (downtick) spun. Then you could apply your pinbars, TLs, PPZs and psych levels, and make a handsome living milking money from the roulette tables.

Hmmm..........
Thanks hanover,

I don't know. What I said seems crazy. But the technical methods seem work on the randomized charts. I am not sure if it is a tea-leafs self-fulfilling prophecy.

I think the only way to make a confirmation is to backtest-trade a huge number of these randomized chart and verify a statistical edge on applying those method on these charts.

kk
  • #34
  • Mar 26, 2010 2:30am Mar 26, 2010 2:30am
Quoting kk007
The strangest thing here is it seems that "patterns" can be seen from a random chart.
It's your brain, not the chart.
  • #35
  • Mar 26, 2010 2:53am Mar 26, 2010 2:53am
kk007, if you would really understand that the next random number is completely independent from the last N which you generated, this topic wouldn't exist.

In mathematics we don't prove our statements with examples but we find an abstract way of representing one idea and we prove that it is valid for all imaginable interpretations of it. The Pythagora's theorem is valid for all the triangles and we'd laugh hearing that someone is searching for a triangle which doesn't exhibit this property.

Because your assumption here is so easily refuted I don't even bother generaring charts to prove that pinbars are more than one of the many formations which will appear from time to time. Knowing how many ticks you use for a bar I can say exactly how many pinbars will appear every 1000 bars.

There's more on the table than this. If you would generate a random chart with 100 million bars and you would show me that you could profit trading pinbars or any other formation, I would surely say that your number generator is flawed or biased. I use mathematics to know when I am wrong because it is infinitely more probable for me to make mistakes than to disprove a theorem.

These are facts. You choose to accept them or you ignore them. By accepting you will know what you can do or not, ignoring them will cost you hopefully only money.
  • #36
  • Mar 26, 2010 3:10am Mar 26, 2010 3:10am
Yes. I understand that when I said I can be profitable with randomly generated chart I was illogical. That is why I said it sounded crazy. Nevertheless, those technical methods really seemed to work on the chart. Perhaps, it was an illusion, but it was so real to me.

Quoting Sauron
kk007, if you would really understand that the next random number is completely independent from the last N which you generated, this topic wouldn't exist.

In mathematics we don't prove our statements with examples but we find an abstract way of representing one idea and we prove that it is valid for all imaginable interpretations of it. The Pythagora's theorem is valid for all the triangles and we'd laugh hearing that someone is searching for a triangle which doesn't exhibit this property.

Because your assumption here is so easily refuted I...
  • #37
  • Mar 26, 2010 3:12am Mar 26, 2010 3:12am
Perhaps you are right. But the randomed charts is so real compared to the real one.

Quoting mikkom
It's your brain, not the chart.
  • #38
  • Mar 26, 2010 3:57am Mar 26, 2010 3:57am
Quoting kk007
Perhaps you are right. But the randomed charts is so real compared to the real one.
It's well known fact that human brains are prone to see patterns where there are none, that is exactly why I tend to overtest everything I think is correct - usually it isn't

As I mentioned, I tested one random candle system (that was even based on the correct distribution) that looked like your charts and yes, I saw all the things you see too, trendlines, horizontal lines etc. There are however differences and finding those differences is really what trading is all about.

Here is one relevant post I posted some time ago, you won't find similar stats on random data and this is exactly why the market is tradeable.

Market may look like random but on many occasions it really is not. Another example is the volume distribution for each day.
  • #39
  • Mar 26, 2010 4:10am Mar 26, 2010 4:10am
1-http://en.wikipedia.org/wiki/Random_walk_hypothesis
2-It is impossible to have an edge along a random walk with 50/50 chance of up/down a pip. (best expectancy you can get is 1). Of course, if it's 51/49 it's easy to create an edge, even without knowing whether it's 51% up or 51% down.

This is where the concept of drift comes up, and then we can move onto stochastic calculus (because that's a bundle of joy).
  • #40
  • Mar 26, 2010 4:38am Mar 26, 2010 4:38am
Quoting TraderHotch
1- http://en.wikipedia.org/wiki/Random_walk_hypothesis
Every time someone points to random walk hypothesis, I feel need to point to this http://press.princeton.edu/titles/6558.html (full text available for free)
0 traders viewing now